Look up Table Research Papers

Direct Torque and Indirect Flux Control of Brushless DC Motor

2000, IEEE/ASME Transactions on Mechatronics

In this paper, the position-sensorless direct torque and indirect flux control of brushless dc (BLDC) motor with nonsinusoidal back electromotive force (EMF) has been extensively investigated. In the literature, several methods have been... more

Fig. 1. Rotor and stator flux linkages of a BLDC motor in the stationary a 3-plane and synchronous dgq-plane [31].

Fig. 2. Dodecagon trajectory of stator flux linkage in the stationary a3-plane.

shown in Fig. 1, where La, = Lys = Ls and Lags = Lgds = 0. Lags and Lqs are the mutual inductances between d- and q-axis. Last and L,;f are the mutual inductances between dq-axes and permanent magnet (PM), respectively, and 7; is the equivalent current generated by PM. In Fig. 1, unlike PMSM with sinu- soidal back EMF synchronous reference frame flux linkages Y7s and yj, vary with time, therefore, stator flux amplitude \y.| is not constant anymore. 7, p, and 6 in Fig. 1 can be ob- tained, respectively, as Because of the rotor position dependant terms in the dg frame stator flux linkages in (5) and (6) and inductances, conventional torque estimation in stator reference frame used for DTC of sinusoidal ac motors is no longer valid for BLDC motor, there- fore, a new torque estimation algorithm is derived in dq frame consisting of actual dg-axes back EMF constants and currents. Instead of the actual back EMF waveforms, Fourier approxi- mation of the back EMFs could have been adopted in torque estimation, but the results would not truly represent the reality and more complex computations are required.

Fig.5. Actual qg- and d-axis rotor reference frame back EMF constants versus electrical rotor position (ka (9r¢) and kg (A,¢)). Fig. 4. Simulated indirectly controlled stator flux linkage trajectory under the sensorless three-phase conduction DTC of a BLDC motor drive when i//* is changed from 0 to —5 A under 0.5 N-m load torque.

‘ig. 3. Overall block diagram of the position-sensorless direct torque and indirect flux control (DTIFC) of BLDC motor drive using three-phase conduction node.

Fig. 6. Experimental test-bed. (a) Inverter and DSP control unit. (b) BLDC motor (Tom_rated = 1.28352 N-m) coupled to dynamometer and position en- coder (2048 pulse/revolution) is not used in the control system.

Fig. 10. (Top) Steady-state behavior of the experimental electrical rotor posi- tion error and (bottom) estimated electromechanical torque when i’) = 0 under 0.5 N-m load torque.

Fig. 8. Experimental indirectly controlled stator flux linkage trajectory under the sensorless three-phase conduction DTC of a BLDC motor drive when 77," = 0 at 0.5 N-m load torque. Fig. 7. Steady state and transient behavior of the experimental. (a) (Top) esti- mated electromagnetic torque, (bottom) error between reference and estimated electromagnetic torque. (b) (Top) q-axis stator current and d-axis stator current and (bottom) ba—ca frame currents when 7//* = 0 under 0.5 N-m load torque.

Fig. 9. (a) Steady state and transient behavior of the actual and estimatec electrical rotor positions from top to bottom, respectively and (b) error betweer actual and estimated electrical rotor positions under 0.5 N-m load torque.

Fig. 11. Experimental estimated electromechanical torque under time-varying reference when 2/;° = 0 under 0.5 N-m load torque.

Fig. 12. Steady-state flux-weakening behavior of the experimental actual speed and estimated electromechanical torque, respectively, (a) when 7!" = 0 and (b) when 7!" = —4.51 A under 1.1926 N-m load torque at 540 electrical rad/s desired speed (Vac jink = 115 V).

Electrical Rotor Position [pi/5 rad/div]

SWITCHING TABLE FOR DTC OF BLDC Moror USING THREE-PHASE CONDUCTION

SPECIFICATIONS AND PARAMETERS OF THE BLDC MoToR

descriptionView Paper arrow_downwardDownload

A core approach to practice-based evidence: A brief history of the origins and applications of the CORE-OM and CORE System

by Jane cahill

2006, Counselling and Psychotherapy Research

This article reviews the development of the CORE-OM and CORE System from 1995 to 2005 in the context of the need to measure, monitor, and manage the delivery of counselling and the psychological therapies in service of providing best... more

descriptionView Paper arrow_downwardDownload

Content-addressable memory (CAM) circuits and architectures: A tutorial and survey

by Ali çam

2006, Solid-State Circuits, IEEE …

We survey recent developments in the design of large-capacity content-addressable memory (CAM). A CAM is a memory that implements the lookup-table function in a single clock cycle using dedicated comparison circuitry. CAMs are especially... more

Fig. 1. Conceptual view of a content-addressable memory containing w words. In this example, the search word matches location (w — 2) as indicated by the shaded box. The matchlines provide the row match results. The encoder outputs an encoded version of the match location using logs w bits. Kostas Pagiamtzis, Student Member, IEEE, and Ali Sheikholeslami, Senior Member, IEEE

Fig. 2. Plot of CAM capacity (log scale) versus year of publication [21]-[27].

Fig. 3. CAM-based implementation of the routing table of Table I.

Fig. 5. CAM core cells for (a) 10-T NOR-type CAM and (b) 9-T NAND-type CAM [26]. The cells are shown using SRAM-based data-storage cells. For simplicity, the figure omits the usual SRAM access transistors and associated bitlines. The SRAM storage and access transistors account for six of the cell transistors.

Fig. 4. Simple schematic of a model CAM with 4 words having 3 bits each. The schematic shows individual core cells, differential searchlines, and matchline sense amplifiers (MLSAs).

Fig. 6. CAM core cells variations for (a) 9-T NOR-type CAM and (b) 10-T NAND-type CAM. The cells are shown using SRAM-based data-storage cells. For simplicity, the figure omits the usual SRAM access transistors and associated bitlines. The SRAM storage and access transistors account for six of the cell transistors.

Fig. 7. Ternary core cells for (a) NOR-type CAM and (b) NAND-type CAM [32], [33]. TERNARY ENCODING FOR NOR CELL

Fig. 8. Structure of a NOR matchline with n cells. Transistor M,,- precharges the matchline and the MLSA evaluates the state of the matchline, generating the match result.

Fig. 10. This figure shows: (a) the schematic with precharge circuitry for matchline sensing using the precharge-high scheme, and (b) the corresponding timing diagram showing relative signal transitions.

‘ig. 9. NAND matchline structure with precharge and evaluate transistors. final matching transistor in the serial nMOS chain. Typically, only one matchline is in the match state, consequently most matchlines have only a small number of transistors in the chain that are ON and thus only a small amount of power is consumed. Two drawbacks of the NAND matchline are a quadratic delay de- pendence on the number of cells, and a low noise margin. The quadratic delay-dependence comes from the fact that adding a NAND cell to a NAND matchline adds both a series resistance due to the series nMOS transistor and a capacitance to ground due to the nMOS diffusion capacitance. These elements form an RC ladder structure whose overall time constant has a quadratic de- pendence on the number of NAND cells. Most implementations limit the number of cells on a NAND matchline to 8 to 16 in order to limit the quadratic degradation in speed [38]. The low noise margin is caused by the use of nMOS pass transistors for the comparison circuitry. Since the gate voltage of the NAND match- line transistors (M, through M,,) when conducting, in Fig. 9, is (Vpp — Vin), the highest voltage that is passed on the match- line is (Vpp — 2Vin), (where V;,, is the threshold voltage of the nMOS transistor, augmented by the body effect). NOR cells avoid this problem by applying maximum gate voltage to all CAM cell transistors when conducting. One implementation of a NAND-based CAM reclaims some noise margin by employing the bootstrap effect by reversing the polarity of the matchline precharge and evaluate [26], [39].

Fig. 12. Two possible configurations for the NOR cell: (a) the stored bit is connected to the bottom transistors of the pulldown pair, and (b) the stored bit is connected to the top transistors of the pulldown pair.

Fig. 11. Matchline circuit model for (a) the match state and (b) the miss state.

Fig. 14. Low-swing matchline sensing scheme of [21].

Fig. 13. Replica matchline for timing control.

Fig. 15. For current-race matchline sensing [42]: (a) a schematic of the circuit implementation including precharge circuitry and (b) a timing diagram for a single search cycle.

Fig. 16. Sample implementation of the selective-precharge matchline technique [43]. The first cell on the matchline is a NAND cell, while the other cells are NOR cells. Precharge occurs only in the case where there is a match in the first cell. If there is no match in the first cell, the precharge transistor is disconnected from the matchline, thus saving power.

Fig. 17. Pipelined matchlines reduce power by shutting down after a miss in a stage.

Fig. 18. Current-saving matchline-sensing scheme [52], [53].

Fig. 19. Hierarchical searchline structure [49], [50]. 'ML swing of 0.45V. Single bit used for first segment. 3 ML divided into three ML segments. the LSL, saving power. Thus, the overall power consumption on the searchlines is

Fig. 20. Simplified diagram of a bank-selection scheme. The bank-selection scheme partitions the CAM and shuts off unneeded banks.

Fig. 21. Conceptual view of pre-computation-based CAM.

Fig. 23. Two ternary CAM cells are viewed as four independent half cells in the dense encoding scheme of [66] and [67]. Fig. 22. Schematic view of matchline structure of pre-computation-based CAM.

Fig. 24. Each local matchline (LML) is encoded as in Table VII using active low logic. The logic level on the global matchline operates with the typical matchline logic levels with a logic high indicating a match and a logic low indication a miss [66], [67]. ?OSSIBLE ENCODINGS FOR TWO TERNARY CELLS IN DENSE ENCODING We also expect the promise of emerging nonvolatile memories to spur innovation in CAM design. These memories, such as ferroelectric RAM (FeRAM), magnetoresistive RAM (MRAM), and phase-change RAM (PRAM) [71]-[73], present several new challenges. The primary one comes from the fact that unlike SRAM cells (or DRAM cells), these memories do not simply store data in the form of low and high voltage levels. Instead, for example, MRAM and PRAM store data as low and high resistance values This means that the comparison operation and the matchline-sensing scheme in nonvolatile CAM will likely be implemented in a different fashion from that now used.

SUMMARY OF SEARCHLINE-DRIVING APPROACHES | Simulated for a 1024 x 144 CAM.

ground indicates a match, while no path to ground indicates a miss. To see how this works, we assume that an LML stores the numbers 0, 2, and 3, which according to Table VII, means that ABCD = 1101. In the case where the search word is 0, 2, or 3 (corresponding to searchlines set to 0001, 0100, and 1000, re- spectively), there is a path to ground indicating a match. In the case of search for a | (corresponding to searchlines set to 0010) there is no path to ground indicating a miss. The GMLs are used to construct longer words that are made up of multiple LMLs.

descriptionView Paper arrow_downwardDownload

Robot Pengikut Garis Berbasis Mikrokontroler AT89C51 Menggunakan Sensor Infra Merah

by Bambang Sutopo

2004

Line Follower Robot is one of autonomous mobile robot form that has mission following autonomously the guided line. In its design and implementation, problems that should be solved are system of robot vision, architecture of hardware... more

descriptionView Paper arrow_downwardDownload

A fast digital image correlation method for deformation measurement

by Bing Pan

2011, Optics and Lasers in Engineering

Fast and high-accuracy deformation analysis using digital image correlation (DIC) has been increasingly important and highly demanded in recent years. In literature, the DIC method using the Newton-Rapshon (NR) algorithm has been... more

descriptionView Paper arrow_downwardDownload

Forest fire detection based on MODIS satellite imagery, and Comparison of NOAA satellite imagery with fire fighters' Information

by Hiroshi Tani

2006, IARC/JAXA Terrestrial …

Impacts of boreal forest fires have absorbed intensive attention because of huge fires in these years in Alaska as well as Siberia. To reduce impacts of forest fire in boreal forest area, the early fire detection is one of essential... more

Table 1: Numbers of reports of boreal forest fires in Siberia region, and duration of collecting correspond to reports from JAL passenger flights projected on Miller Cylindrical.)

Figure 2 An example of comparison of a time series of satellite imagery with a ground

Figure 8: A series of imagery with detected hotspots of forest fires by MODIS. (Red plots

Table 2: Forest fire detection rate for each observation data sources. 4. Conclusions

descriptionView Paper arrow_downwardDownload

A Modified Approach Towards Testing of LUTs in a FPGA

by Sumit Raj

Testing an FPGA chip poses a challenging problem for test engineers. It requires implementing various configurations of the FPGA, but changing configurations incurs reprogramming costs. The fundamental question for FPGA testing is how to... more

The detail structure can be found in various Xilinx FPGA families [7]. Let k denote the number of input lines of an LUT; then an LUT can implement 2” logic functions, n = 2". The configuration memory cells (CMCs) are used to configure its logic functions. While programming an FPGA we load the bit patterns corresponding to the truth table of the function to be implemented into the CMCs and the interconnection networks are also configured. Such a programming process is called a configuration. Several FPGA architectures have been developed for different applications. The most widely used type is the Look- Up-Table (LUT) FPGA (eg. Xilinx [5]-[7] focus on the testing of SRAM-based FPGAs. 1 in which the functional unit consists of several LUTs. This type of FPGA can be reprogrammed any number of times. In this paper we The architecture of a SRAM-based FPGA is a two dimensional array in which m x m logic cells can be configured to implement user defined ogic functions. These configurable logic cel by a configurable interconnection network. F this m x m two-dimensional array with m = 3. Ss communicate ig. 1 illustrates It is well known that an unprogrammed FPGA can be configured to implement different logic functions. Therefore, testing this category of FPGAs implies testing of all possible

Fig. 4 Address decoder logic in 3-input SRAM based LUT

General approach of testing FPGA is configuring LUTs ina row as ILA (Iterative Logic Array) such LUT is given to the output to the next LUT results in a problem that in a k-LUT there the output is 1 bit. Hence, it is not beneficia that output of one in that array. This are k bit input but to view an FPGA as a N X N array of CLBs. Thus, it is impossible to construct an iterative logic array. The remaining k- provided from the helper LUTs on the inputs are either side or they are common input as in the previous LUT as shown in Fig. 5. Thus, in both the cases we don’t have full controllability as we cannot give our own input to the LUT exhaustively. This results in controllability problem in testing FPGA. Thus, FPGA testing suffers from both controllability and observability problem. The output of a given module is connected to an input of the following one. Each module is also connected to a set of primary inputs that are common to every module in the FPGA. The input of the first left most module is a primary input and the output of the last right most module is a primary output, as shown in Fig. 6. This practical solution does not require many /O pads and has been used by many authors [14]-[19]. It must be clear that this solution requires less I/O pads but the fundamental problem is the controllability of the ‘embedded’ module through the preceding ones and the observability through the following ones. Usually the full controllability and observability of every module cannot be guaranteed by a unique test configuration.

Fig. 6 Configuration of FPGA as ILA structure

Fig 9: Hierarchical approach towards testing of LUTs in an FPGA Fig 7: Testing LUT using common input and giving output to compressor

Fig 8: Routing resources of Xilinx Virtex II FPGA [23]

descriptionView Paper arrow_downwardDownload

Design and experimental verification of a constrained finite time optimal control scheme for the attitude control of a quadrotor helicopter subject to wind gusts

by Kostas Alexis

2010, Robotics and Automation ( …

In this paper the design and the experimental verification of a Constrained Finite Time Optimal (CFTO) control scheme for the attitude control of an Unmanned Quadrotor Helicopter (UqH) subject to wind gusts is being presented. In the... more

descriptionView Paper arrow_downwardDownload

Lookup table based multiplication technique for GF(2m) with cryptographic significance

by Athar Mahboob

2005, Iee Proceedings-communications

Finite field multiplication is one of the most important arithmetic operations in the binary finite field GF(2 m ). Multiplication is required for many cryptographic techniques based on the discrete logarithm problem (DLP) in the... more

5.3 Fast GF(2™) multiplication algorithm using mullut- F-LUT Based on our experience with our B-LUT multiplier we then implemented a Fast GF(2”) LUT multiplier (F-LUT). Our F-LUT multiplier removes the performance bottle- necks in the B-LUT multiplier. We recognise that the performance of the B-LUT multiplier degrades owing to a large number of word-level operations, i.e., SHIFTs, XORs and ANDs arising out of the fact that we process only d bits of the input at a time. With a view to minimising the number of word-level operations the first thing to observe is that a product of values up to 8 bits can be simply done by a single table lookup into mullut[256][256] with the 2 values used as indices. Now to multiply any two 16 bit values A and B we can think of them as consisting of two 8-bit halves A,||Ag and B,||Bo respectively. Their product then is (A: By) «16) ® ((Ay-By) <8) © (Ao: Bi) «8) @ Ag: Bo. The idea is shown in Fig. 3.

5.2 Basic GF(2™) multiplication algorithm using mullut — B-LUT

Fig. 2 Multiplication of 32 bit values in GF(2”) using mullut the product using 16 bit multiplications as outlined above combined with appropriate shifts and XORs. Using this scheme our F-LUT multiplier matches the unit of processing to w, the word size of the machine as soon as possible. We calculate the product of w = 32 bit values with the help of mullut by using the scheme just described but then we create another level of processing by using the 32

"ig. 4 Processing 8 bits at a time against processing w = 32 bits a left shift by one position of the vector representing a(x). The addition specified in Step 6 of Algorithm 4 may be performed using GF(2”) addition.

6 Other polynomial basis GF(2™) multiplication techniques Several algorithms have been reported for multiplication in GF(2”) in the polynomial basis. These include:

Table 1: Standard irreducible polynomials

Table 3: Number of word-level operations required for different GA\2™) multiplication techniques

Table 4: Number of word-level operations required for various binary finite fields multiplication techniques for m= 163, 283, 571 (w= 32 bits)

Table 2: Number of operations for base case of GF(2°”) multiply using mullut number of shifts are Sgs*, the number of ANDs are Ags and the number of table lookups are Ts”.

Table 5: Timings (in ys) for different finite field multiplication methods (Pentium IV 2.0 GHz with 512 kB L2 cache amount to several finite field arithmetic operations such as addition, multiplication, squaring and inversion in the underlying finite field which in our case is GF(2”). The finite field arithmetic operations have then to be implemented as word-level operations for a software implementation on a general purpose processor or they have to be implemented as register-level operations in a hardware implementation. For example, the simple elliptic curve Diffie-Hellman (EC- DH) analogue performs the elliptic curve trapdoor opera- tion four times to establish a common secret key between two parties. Each trapdoor operation requires a number of group operations which in turn require a number of finite field operations. As a matter of fact, the efficiency and performance of the ECC implementation depends, for the most part, on the performance of the lowest level which is concerned with the finite field operations and arithmetic [15, 16]. In Fig. 6 we show one possible abstraction of the hierarchy of implementation of operations for ECC. here on performance improvements with various window sizes. From the results for window sizes of 2, 4 and 8 bits, as given in Table 5, it can be seen that performance improves most with window size u=2 and modestly with a window size of u=4, when m is smaller but significantly when m is large. However, for a window size of u= 8 the performance degrades significantly in all cases. This is because the time spent in precomputation increases exponentially with the increase in window size (wu) and starts to dominate, thereby denying any improvement in performance. However, for multiplication with a constant, where the precomputation can be done once, for all, the larger window sizes would result in better performance than smaller window sizes. Lastly, our result for u=2 is also better because we have used the trick multable[0]=0 and multable[1|=a to cut down 50% of the precomputation values for u=2 but for larger wu this does not help significantly. Our timings for F- LUT based GF(2”) multiplication are comparable to or better than other reported timings [10-14]. We believe that there is room to optimise further our algorithm in addition to our implementation to meet the fastest published results in [11, 12]. We are working on improving the performance of our Comb and Karatsuba based implementations. NUT. 4.2. .....* 4. st. yt... | th Ul lhClU TQ CQgte Cli Tt

Table 6: Time required to calculate the 128kbyte lookup table mullut[256][256] ) Impact of our method on ECC performance We also provide the time required to calculate the table mullut{256][256] in Table 6. It may be noted that our LUT is calculated once, for all, and can even be loaded from a file at system initialisation.

Table 7: Timings (in ms) for elliptic curve scalar multiplication using affine coordinates except as noted (Pentium IV 2.0 GHz with 512 kB L2 cache)

descriptionView Paper arrow_downwardDownload

PSYCHOPHYSICAL CALIBRATION FOR CONTROLLING THE RANGE OF A VIRTUAL SOUND SOURCE: MULTIDIMENSIONAL COMPLEXITY IN SPATIAL AUDITORY DISPLAY

by William Martens

2001, International Conference on Auditory Display

Just as control over perceived azimuth and elevation of a virtual sound source should be psychophysically calibrated in spatial au- ditory display, so should perceived range; however, in contrast to azimuth and elevation display, precise... more

descriptionView Paper arrow_downwardDownload

Objective Performance Testing and Quality Assurance of Medical Ultrasound Equipment

by Johan Thijssen

2007, Ultrasound in Medicine & Biology

There is an urgent need for a measurement protocol and software analysis for objective testing of the imaging performance of medical ultrasound equipment from a user's point of view. Methods for testing of imaging performance were... more

There is an urgent need for a measurement protocol and software analysis for objective testing of the imaging performance of medical ultrasound equipment from a user's point of view. Methods for testing of imaging performance were developed. Simple test objects were used, which have a long life expectancy. First, the elevational focus (slice thickness) of the transducer was estimated and the in-plane transmit focus was positioned at the same depth. Next, the postprocessing look-up-table (LUT) was measured and linearized. The tests performed were echo level dynamic range (dB), contrast resolution (i.e., gamma of display, number of gray levels/dB) and sensitivity, overall system sensitivity, lateral sensitivity profile, dead zone, spatial resolution and geometric conformity of display. The concept of a computational observer was used to define the lesion signal-to-noise ratio, SNR L (or Mahalanobis distance), as a measure for contrast sensitivity. All the measurements were made using digitized images and quantified by objective means, i.e., by image analysis. The whole performance measurement protocol, as well as the quantitative measurements, have been implemented in software. An extensive data-base browser was implemented from which analysis of the images can be started and reports generated. These reports contain all the information about the measurements, such as graphs, images and numbers. The approach of calibrating the gamma by using a linearized LUT was validated by processing simultaneously acquired rf data. The contrast resolution and echo level of the rf data had to be compressed by a factor of two and amplified by a gain factor corresponding to 12 dB. This resulted in contrast curves that were practically identical to those obtained from DICOM image data. The effects of changing the transducer center frequency on the spatial resolution and contrast sensitivity were estimated to illustrate the practical usefulness of the developed approach of quality assurance by measuring objective performance characteristics. The developed methods might be considered as a minimum set of objective quality assurance measures. This set might be used to predict clinical performance of medical ultrasound equipment, taking into account the performance at a unique point in space i.e., the coinciding depths of the elevation and in-plane (azimuth) foci. Furthermore, it should be investigated whether the approach might be used to compare objectively various brands of equipment and to evaluate the performance specifications given by the manufacturer. Last but not least, the developed approach can be used to monitor, in a hospital environment, the medical ultrasound equipment during its life cycle. The software package may be viewed and downloaded at the website http://www.qa4us.eu. (E-mail: j.thijssen@cukz.umcn.nl) Objective performance of medical ultrasound equipment • J. M. THIJSSEN et al. 461

Fig. 1. Scheme of tissue-mimicking test phantom containing various objects, base material urethane rubber (model 539; ATS Inc., Bridgeport, CT, USA).

Fig. 2. Image of phantom showing (a) gray wedge (upper right) with interactively placed gray level measurement box; (b) Step curve: example of gray level wedge curve corresponding to post- processing curve (or gray level map or LUT); dotted line: nineth order curve fitted to step curve; dot-dash line: ideal linear LUT.

Fig. 3. Scheme of slice thickness phantom. The —6 dB beam width, t, in elevation direction is obtained from measured —6 dB width, x, by using known angle (45°) of scattering layer: t = x/tan(45). In general, array transducers have a fixed focus cylindrical lens at their surface to narrow the transmitted sound field. This focusing action is in the direction perpendicular to the scanning plane, hence the term “elevation focusing”, or “slice thickness reduction”. A special phantom is used to measure the depth where the elevation focus is located (ATS, model 538N). The scheme of the phantom is shown in Fig. 3. It contains a plane, thin scattering sheet at a 45° angle with the hori- zontal plane. A manual shifting of the transducer over the phantom surface yields a shifting, horizontal bar image with varying widths being displayed (Fig. 4a) that can be used to obtain the slice thickness versus depth. An image was stored at five or more positions of the transducer. The software checks during the measurements whether overflow of the stored gray levels occurs, e.g., whether the gray level exceeds the value of 250 (maximum in 8 bits storage is 255). Each image was analyzed after interactive placement of a rectangular box over the bar. The gray levels within the box were laterally averaged and then plotted versus depth. By using a gamma of 4 (it is assumed here that the image contains four gray levels per dB echo level), the gray level versus depth plot was transformed into a dB versus depth plot and, for each stored image, the —6 dB width was estimated by the software (taking into account also the 45° angle). Finally, by using the data of all the stored images, a plot of the

Fig. 4. (a) Image obtained from slice thickness phantom, show- ing the measurement box. Width of bar is defined as —6 dB width averaged over box. (b) Example of slice thickness vs. depth (asterixis), parabolic curve fitted to the points and the estimated focal depth and thickness shown (filled circle, at minimum of parabola).

Fig. 5. (a) Echographic image obtained from scattering lesions and measurement disc inserted in one of them. (b) Plot of mean gray level of lesions (average of five measurements and stan- dard deviation; left abscissa) vs. nominal contrast values of scattering lesions. Linear regression yields contrast resolution (gamma = number of gray levels/dB) and echo level dynamic range (# dB for full range (O—255)). Plot of lesion signal-to- noise ratio (SNR,, or Mahalanobis distance, MD) vs. nominal contrast values of scattering lesions. Value at 3 dB is taken as the measure of contrast sensitivity.

Fig. 6. Measurement of system sensitivity. (a) Image obtained from homogeneous part of the phantom for the estimation of penetration depth (overall sensitivity). TGC is adjusted to ob- tain homogeneous gray level over maximum attainable depth range. Preferably, the automatic-gain-control facility of the equipment should be used when available. Measurement box interactively overlaid by using developed software. (b) Mean, gray level vs. depth curve with horizontal line fit and estimated penetration depth at —6 dB (gray level 67) as indicated at 28 mm.

Fig. 8. Measurement of spatial resolution by assessment of the in-plane point-spread-function (IPSF). (a) Echographic image of the wires around the focal zone. Box (interactively placed around wire image at depth of the transmit focus) indicates area selected for measurement of the spatial resolution, i.e., size of point-spread function. (b) Exmaple of gray level profiles in dB through center of wire image (stars, dotted lines) and parabolic fit (drawn curves) with horizontal lines indicating —6 dB and —20 dB levels from top. Full-width-at-half-maximum (FWHM, i.e, at —6 dB) and full-width-at-tenth-maximum (FWTM, i.e., at —20 dB) widths are estimated by using the estimated gamma, i.e., recalculating gray levels into decibels.

Fig. 9. Measurement of geometrical conformity of the echo- graphic image. The software reads depth ruler. The horizontal and vertical extreme wire spots are located and used to calcu- late the corresponding horizontal and vertical distances. These distances are compared with the nominal distances indicated at the phantom and the relative deviation (%) is calculated and mentioned in the report. The results of the contrast measurements made with simultaneous video and rf acquisitions are discussed in Fig. 10. In this figure, the lowest data points and fitted straight line are the results from the rf data after the log (Hilbert) processing. It can be concluded that the straight line fits the data points closely. The slope of the line corresponds to a dynamic range of 117 dB. The com- pression chosen for the video processing within the

Fig. 10. Comparison of gray level encoding and contrast resolution for video image data vs. rf data. Ensemble mean gray level of contrast discs vs. nominal echo levels (dB). Lower data and linear regression line: log (AM demodulated rf) data; lower-but-one data and regression: data from lower curve after compression by a factor of 2; upper-but one data and regression: data from lower-but one curve after additional gain factor of 12 dB; upper curve: data and regression from simultaneously measured DICOM data.

Table 1. Quality assessment of three different transducers. Gamma is the number of gray levels per dB nominal echo level; SNR, is the lesion signal-to-noise ratio, FWHM and FWTM are the full-width-at-half-maximum and full-width-at-tenth-maximum, i.e., the in-plane spatial resolution in axial and in lateral directions

descriptionView Paper arrow_downwardDownload

An Enhanced Approach for Arabic Sentiment Analysis

by International Journal of Artificial Intelligence (IJAIA)

In recent years, there are massive numbers of users who share their contents over wide range of social networks. Thus, a huge volume of electronic data is available on the Internet containing the users' thoughts, attitudes, views and... more

descriptionView Paper arrow_downwardDownload

A simple dynamic overmodulation strategy for fast torque control in DTC of induction machine with constant switching frequency controller

by Tole Tole

2009

This paper presents a simple dynamic overmodulation method to obtain a quick dynamic torque response in Direct Torque Control with constant switching frequency (will be referred as DTC-CSF) induction machines. A fast dynamic torque... more

descriptionView Paper arrow_downwardDownload

Optimal Exploitation of the Sentinel-2 Spectral Capabilities for Crop Leaf Area Index Mapping

by Wolfram Mauser

2012, Remote Sensing

The continuously increasing demand of accurate quantitative high quality information on land surface properties will be faced by a new generation of environmental Earth observation (EO) missions. One current example, associated with a... more

Figure 1. Distribution and variation of the resulting LAI RMSE values for each band arrangement for the LUT based approaches (a): LUTyso and (b): LUTy; as well as the ANN approach (c). The bottom and the top of the boxes describe the 25th and the 75th percentiles; the red line within the boxes corresponds to the 50th percentile (or median); lower and upper ends of the whiskers show lowest and highest values within the 1.5 interquartile range (IQR). Outliers are marked with a red cross. Whereas the absolute lowest RMSE values are on a similar leve provides lower variability of RMSE differences (i.e., highest RMSE for ANN and LUT yso, the latter C min — lowest RMS] Emin): 0.17 of LUTnso compared to 0.25 of the LUTn; and 0.33 of the ANN methods. This indicates higher robustness and a lower sensitivity of the LUTnso method regarding the optimal number of spectral bands. As demonstrated, the results of such analyses may depend on the algorithms employed. However,

Whereas the highest accuracy was found by all statistical indicators for onions and alfalfa using the UTwnso approach, the ANN method obtained better results for maize, sugar beet and garlic. For these hree crops, however, the spatial patterns (indicated by R?) were better reproduced by the LUTNso ypproach, although the correlations were still low (from R? ~ 0.1 to 0.4). Moreover, all NSE are < 0, In fact, it is notable that the retrieval performance depends on the crop type (Figure 3, Table 3)

the corresponding pictures of each crop field, taken during the campaign, are presented Figure 4. Selected simulated (PROSAIL generated database for LUT and ANN approaches) and HyMap spectra, one for each crop type; band setting of best results (RMSEnin) of LUTnso approach, i.e., first eight S2 bands used in the study (i.e. from 490 nm to 865 nm). All simulated spectra with radiometric RMSE value within 20% of the lowest RMSE are presented. Measured LAI values are indicated in the titles. On the right the corresponding pictures of each crop field, taken during the campaign, are presented.

Even when looking at the crop-specific results, it cannot be concluded that one method outperforms the other. Both approaches, LUTnso and ANN, reveal similar performances with reasonable results. However, problems are still present depending on the architecture and actual growth stages of the crops. In order to obtain an idea of the model’s ability to reproduce the HyMap reflectance data, some exemplary spectra are presented in Figure 4. For this purpose, the simulated spectra obtaining the best radiometric match within the LUTnso approach were chosen (eight bands). It is clearly visible in Figure 4 that in all cases the HyMap reflectance is appropriately reproduced by the model.

Ranges of variables were applied according to: ' [37], 7 [21], ° [20], * [50], > [56]. data base for LUT and ANN approaches. Atmospheric correction and instrumental noise can result in multiplicative and additiv uncertainties. Radiometric calibration might be inaccurate, which leads to systematic errors. Moreover the used RTM may contain errors depending on its (simplified) description of the radiation regime in vegetation canopy. Thus, to at least partly account for these uncertainties, the inclusion of noise in th simulations was decided. Fifty random initializations were generated adding and multiplying Gaussia white noise (absolute: 0.01 and relative: 4%) to the simulated reflectance. This was don band-dependently and band-independently according to [37], who demonstrated that the combinatio of all these error-terms performed best for variable retrievals: r 7 7

3. Results and Discussion In this section, outcomes of the spectral band analyses are presented and discussed. Three aspects are considered: first, the distribution of RMSE values between measured and estimated LAI (“LAI RMSE”) for all possible band arrangements is analyzed. Second, the importance of the different spectral regions for LAI estimation is addressed. Finally, crop specific differences are elaborated.

Table 3. Goodness-of-fit statistics (R’, RMSE, NRMSE and NSE) between observed and estimated LAI from SPARC’03 data. Higher performance values between the two approaches are emphasized in bold. Thus, the models prediction capabilities are doubtful. This may be due to the above-mentioned imitations of the used RT model: the erectophile canopy of garlic, for instance, may lead to a strong nfluence of the soil, being a very critical factor for model inversion [31,37]. Looking at Figure 4, the sarlic spectra resemble bare soil signatures (both measured and simulated) due to absent chlorophyll ibsorption even though the field measurements indicated a LAI of 0.8. According to the picture taken luring the campaign, garlic exhibits already a senescent growth stage. The measurement from the _AI-2000 instrument (PATef) is therefore strongly influenced by non-green plant components. This uggests that the LAI estimated by the LUTnso approach is closer to the green LAI value than stimated by the ANN approach (see also Figure 3) or measured in the field. Despite the fact that the neasured and ANN estimated LAI-values of garlic are very close, the LUTnso approach seems to give he most accurate interpretation of the spectral signature. NA... ..... abt. 2... CUD 2. Onde kk. 8k 2 ge LY LLL Od Oe a lg i. ) CL. . Le. the. 1tCO OT NT ....W 1.1 .......

descriptionView Paper arrow_downwardDownload

Memory Optimization Technique

by Gopika Krishnan

A lookup table is an array containing Pre-calculated values that can be retrieved from memory whenever needed. The project optimizes the memory space needed to store data in a look up table by combining the methods of Anti symmetric... more

Figure 2.1 : conventional LUT-based multiplier

Figure 3.2: Proposed APC—OMS combined LUT design for the multiplication Of W-bit fixed coefficient A with 5-bit input X. line decoder is a simple modification of 3-to- 8-line decoder, as shown in Figure 3.3

3.2 and 3.3.The control circuit to generate the control word and RESET is shown in Figure 3.4.

The address-generator circuit receives the 5-bit input operand X and maps that onto the 4-bit address word (d3d2didb). 3.2.3 Optimized LUT Design for Signed and Unsigned Operands he XOR operation of sign bits of both multiplicands. When both operands are in two’:

Figure 3.6: Optimized implementation of the sign modification of the odd LUT output sign-modification circuit as in Figure 3.6.

3.3 MEMORY-BASED MULTIPLICATION BY INPUT OPERAND DECOMPOSITION

Figure 3.8: Wallace tree multiplier 3.3.2 Adder Tree

Figure 3.9: Proposed LUT -based multiplier for L =5

operations. The functionality of the design is explained below. used for tracking induced miss rate in cache memory. Any data requested by microprocessor present in cache memory then cache, as compared to RAM, chip energy consumption. Thi Jetecting miss rate in cache the term is called “cache hit“. The advantage of storing data on is that it has faster retrieval times, but it has disadvantage of on- S project deals with the design of efficient cache memory for memory and less power consumption.[2] Cache controller that communicates between microprocessor and cache memory to carry out memory related \naratonne Tha finchonnalitz on F tha dacinn ic ayvniainad halonwz communicates between microprocessor and cache memory to carry out memory related

The 10-bit counter, shown in figure is used to generate 10 bit address for writing and are stored in tag memory. The cache tag memory shown in Figure 3.11 stores address of data in reading operation of cache tag memory. A Ten bit counter shown in Figure 3.12.

For each clock pulse counter is incremented by one. 10- bit counter provides 10-bit

Figure 3.14: Implementation of cache memory

In combined method the number of slice Flip Flop is 16 and number of input LUT is ‘his is much reduced than APC alone.

Table 3.1: APC words for different input values for L =5 these two input values on the same row is 32A.[1]

The product values and encoded words for input words X = (00000) and (10000) are t left shifts of 2A. For X= (10000), the APC word “0” is derived by resetting the LUT output,

Table 3.3: Products and encoded words for X = (00000) and (10000) by an active-high RESET signal given by RESET =(x0 +x1 +x2 +x3) -°x4.

descriptionView Paper arrow_downwardDownload

Real-time machine vision weed sensing

by B. Steward

1998

Much work has been done to employ machine vision technology to sense weeds in crop fields. However, the use of machine vision weed-sensing with real-time objectives under variable outdoor lighting conditions is a relatively new area. This... more

descriptionView Paper arrow_downwardDownload

Efficient Implementation of Marching Cubes' Cases With Topological Guarantees

by Helio Lopes

2003, Journal of Graphics Tools

descriptionView Paper arrow_downwardDownload

HYDROPT: A fast and flexible method to retrieve chlorophyll-a from multispectral satellite observations of optically complex coastal waters

by Reinold Pasterkamp

2008, Remote Sensing of Environment

We present a generic innovative algorithm for remote sensing of coastal waters that can deal with a large range of concentrations of chlorophyll-a, SPM and CDOM and their inherent optical properties. The algorithm is based on the exact... more

Provided for non-commercial research and education use. Not for reproduction, distribution or commercial use. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited.

Fig. 1. Results of the comparison between the remote sensing reflectance as computed by HYDROLIGHT and approximated by a table of polynomial coefficients. The results are based on 100 randomized sets of absorption, scattering, solar zenith angle and fraction of diffuse sky irradiance, for a nadir-viewing angle. The ¢ is the root- mean square error in R,,(A). HJ. Van Der Woerd, R. Pasterkamp / Remote Sensing of Environment 112 (2008) 1795-1807

Fig. 2. Results of the inverse modeling test for concentrations of (a) chlorophyll-a, (b) suspended particulate matter and (c) the absorption of colored dissolved organic matter. Diagonal dotted lines indicate 50, 100 and 200% error bounds. The thick line follows from log—log linear regression.

Fig. 3. Outcome of the chlorophyll retrieval by the HYDROPT algorithm on a set of 76 reflectance spectra collected in the REVAMP project. Note that the chlorophyll concentration covers almost two orders of magnitude and seven different water masses with variations in IOP and concentrations in the optical properties of SPM and CDOM. The log-log linear fit is described by y=0.022+ 1.16x, R°=0.84, RMS error is 0.23. The algorithm was run on a set of remote sensing reflectance spectra to test the inverse modeling approach. The test is a first basic step to check if the Levenberg—Marquardt procedure, combined with the polynomial fit, can invert the reflectance spectra in a satisfying manner. The HYDROLIGHT code was used to calculate the R,, spectra. During the Particulate Matter North Sea (PMNS) program in 1993 and 1994 a data set was

Fig. 4. Sample fitted remote sensing reflectance spectra. The solid spectra were extracted from a MERIS L2 image (see text), the dashed spectra were fitted by the algorithm. The spectra were located over (a) the Flemish Banks (b) the Central North Sea and (c) a bright patch (algal bloom).

Fig. 5. Concentration retrieval with the HYDROPT algorithm from a MERIS observation made on April 22nd 2003. The left panels show, from top to bottom, the concentration of CHL, SPM and the absorption by CDOM. The right panels show the estimated standard error per pixel for CHL, SPM and CDOM, based on the HYDROPT spectral fit procedure. No satisfying fit could be found off shore from the Danish Coast and near clouds. HJ. Van Der Woerd, R. Pasterkamp / Remote Sensing of Environment 112 (2008) 1795-1807

HJ. Van Der Woerd, R. Pasterkamp / Remote Sensing of Environment 112 (2008) 1795-1807 Fig. 6. The median CHL values for the month April 2003, derived from MERIS RR observations with HY DROPT. The image is one example from the REVAMP atlas; see http://www.brockmann-consult.de/revamp. Spring blooms have started along the coast of Belgium, Holland and Denmark and in the central North Sea above 54°.

The coefficients of the optical model for the North Sea used in testing and demonstrating the algorithm [he absorption and scattering of pure water, and the absorption and scattering per unit concentration of chlorophyll-a, suspended particular matter and colore Jissolved organic matter are presented for wavelengths corresponding to the MERIS bands. Table 1

Summary of the data from in-situ measurements for the validation of the HYDROPT algorithm The first four columns describe for each geographical region the number of spectra collected, the range in in-situ CHL concentration and the range in R,, (at 560 nm) as an indication of the backscatter in these waters. The last two columns give the number of valid HYDROPT output CHL values and the ratio of computed versus measured CHL in each region. Table 2

Fit results for three sample spectra, also presented in Fig. 4a—c, extracted from the HYDROPT processing of the MERIS observation at September 2nd, 2002 For each pixel the location, retrieved concentrations and estimated standard error are presented. The 7’ is calculated over 7 consecutive band differences, normalized tc a standard error of 3:10 * (sr '). The cumulative probability (P.ni2) gives the probability that the 7° exceeds the measured value by chance for 5 degrees of freedom Table 3

descriptionView Paper arrow_downwardDownload

ROBOT PENGIKUT GARIS BERBASIS MIKROKONTROLER AT89C51 MENGGUNAKAN SENSOR INFRA MERAH

by Budi Raharjo

Line Follower Robot is one of autonomous mobile robot form that has mission following autonomously the guided line. In its design and implementation, problems that should be solved are system of robot vision, architecture of hardware... more

descriptionView Paper arrow_downwardDownload

Material removal simulation of peripheral milling of thin wall low-rigidity structures using FEA

by Svetan Ratchev

2004, Advances in Engineering Software

The paper reports on a new methodology for simulation of the material removal process that takes into account the deflection of the part during machining. A finite element part model is used as an input to the simulation. The algorithm... more

descriptionView Paper arrow_downwardDownload

Detection of land cover changes using MODIS 250 m data

by Charlene DiMiceli

2002, Remote Sensing of Environment

The Vegetative Cover Conversion (VCC) product is designed to serve as a global alarm for land cover change caused by anthropogenic activities and extreme natural events. MODIS 250 m surface reflectance data availability was limited both... more

The Vegetative Cover Conversion (VCC) product is designed to serve as a global alarm for land cover change caused by anthropogenic activities and extreme natural events. MODIS 250 m surface reflectance data availability was limited both spatially and temporally in the first year after launch due to processing system constraints. To address this situation, the VCC algorithms were applied to available MODIS 250 m Level 1B radiance data to test the VCC change detection algorithms presented in this paper. Five data sets of MODIS Level 1B 250 m data were collected for the year 2000, representing: (1) Idaho -Montana wildfires; (2) the Cerro Grande prescribed fire in New Mexico; (3) flood in Cambodia; (4) Thailand -Laos flood retreat; and (5) deforestation in southern Brazil. Decision trees are developed for each of the VCC change detection methods for each of these six cases. These decision trees are to be used for updating the look-up tables required by the VCC production code. For these change detection cases, the VCC change detection methods worked reasonably well. In the Idaho -Montana wildfire case, a fire perimeter polygon data set compiled by the USDA Forest Service was used to validate the output of the VCC change detection methods. Although the VCC output identified only 32% of the burned pixels within the ground observed Idaho -Montana fire perimeter polygons, the detection accuracy of the VCC output did reach 99% when the VCC product is considered as an alarm system identifying the occurrence of the change in an area. For other cases, the detection accuracy in per-pixel terms of the VCC output ranges from 55% to 90% against reference change bitmaps that were created by image interpretation. Look-up tables created with AVHRR and Landsat Thematic Mapper data require modifications for the MODIS data due to differences in radiometric response between MODIS and the heritage instruments. The applications presented in this paper also evaluate the relative performance of each of the five change detection methods used as VCC algorithms. Conclusions reached in this paper will be used for future refinement of the VCC product. D

descriptionView Paper arrow_downwardDownload

Estimation of cloud optical thickness by processing SEVIRI images and implementing a semi analytical cloud property retrieval algorithm

by Nicole van Lipzig

Clouds play a very important role in the Earth's climate system, as they form an intermediate layer between Sun and the Earth. Satellite remote sensing systems are the only means to provide information about clouds on large scales.... more

Fig. 1. Background surface albedo (A in %) for June 2006 from minimum value compositing method. Region with high background surface albedo are shown in darker shade whereas lighter shade shows the region with low background surface albedo.

Fig. 2. The parameterised transmittance given by Eq. (6) (continuous curve) and the transmittance computed from TUV model simulations (symbols).

Fig. 3. SCOT fields at several selected instants, shown on top left of each map, corresponding to passes of the CloudSat instrument. The CloudSat passes are denoted by the dashed line.

Fig. 4. Scaled cloud optical thickness (r*) retrieved from our approach (continuous black curve) and from CloudSat (green symbols). The uncertainity in the CloudSat values are given by the shaded region.

Fig. 5. Location of the measurement sites on a Belgian map with daily noon (11:57 UTC) cloud optical thickness values as retrieved from SEVIRI for the month of June 2006.

Fig. 6. Time series of modelled (solid lines) and measured (orange symbols) UV-A (a) and UV-B (b) irradiance over Ostend and UV-A (c) and UV-B (d) irradiance over Redu for the month of June 2006.

Fig. 7. A scatter diagram of modelled cloud modification factor (CMFmoq) versus observed cloud modification factor (CMFops) over Ostend during June 2006.

Table 1. Error Statistics between estimated SCOT from SEVIRI and SCOT from CloudSat. The difference statistics we obtained for the results based on our approach versus the CloudSat are reasonably satis- factory. This leads us to conclude that the application of SACURA algorithm to SEVIRI data yields satisfactory val- ues for the scaled cloud optical thickness. negative values. The root mean square difference (RMSD) is generally of the order of 1 to 2, which is rather low com- pared to the typical SCOT values occurring in the CloudSat tracks considered here, reaching up to values of ~ 10 and beyond. The correlation between both sets of SCOT values is also reasonably good, in the range 0.51-0.78. Finally, the index of agreement also exhibits satisfactory values ranging between 0.68 and 0.86, which points to a good performance of the retrieval scheme.

Table 2. Error Statistics between estimated surface level UV irradiance and measurements. ditrerence between [OMs UV estimate and the ground mea- surements is between +10%. They also found that TOMS overestimated the ground measurements under overcast con- dition as the value of the bias reached 15-17%. Fioletov et al. (2002) found that the TOMS UV estimate versus ground measurements at Toronto, Canada, has a correlation coeffi- cient of 0.9 with a bias of 9%. McKenzie et al. (2001) found an average correlation of 0.81 between TOMS estimate and ground measurements of UV erythemal dose at four stations around the world. Peeters et al. (2000) compared the UVB estimated by GOME (Global Ozone Monitoring Experiment against measurements over Uccle, Belgium. They found that the satellite estimations were twice as high as the measure- ments, particularly in the case of cloudy conditions. Among other factors, this difference pertains to the spatial resolu- tion of GOME ~ 40 x320 km?. Whereas, the results obtained using our approach yielded a mean correlation of 0.91 be- tween modelled UV irradiance and ground based measure- ments with a low bias. The assessment of error statistics with that of previous studies reveal that the results obtained in this study matches satisfactorily with the results from pre- vious studies. There are several reasons that limit an accu- rate intercomparison between our results with previous stud- ies. Various studies employ various instruments/approaches, which does not measure exactly the same quantities and are focused over different geographical domains. These factors complicate the intercomparison of our results with previous studies. Moreover, the difference in spatial resolution of vari- ous remote sensing instruments, SEVIRI = 3 x3 km2at nadir, TOMS = 50x 200 km2and GOME ~ 40 x320km2, possess additional complications for comparisons together with the fact of using different spectral bands.

Table 3. Mean normalised absolute difference (MNAD) for UVA and UVB irradiance under varying aerosol optical thickness (AOT) values. by calculating this quantity using the mean AOT (0.25), as well as the mean AOT plus and minus one standard devia- tion (AOT values of 0.10 and 0.40, respectively). Perturbing the AOT as just explained induces mean normalised absolute difference (MNAD), which was calculated as following:

descriptionView Paper arrow_downwardDownload

A PSO based optimal switching technique for voltage harmonic reduction of multilevel inverter

by Debashis Chatterjee

2010

Selected lower order harmonics of multilevel inverter are eliminated while the overall voltage THD is optimized by computing the switching angles using particle swarm optimization (PSO) technique. The discontinuity in the solution of... more

descriptionView Paper arrow_downwardDownload

Statistics For Strontium Isotope Stratigraphy: A Robust Lowess Fit to the Marine Sr-Isotope Curve For 0 to 206 Ma, With Look-Up Table For Derivation of Numeric Age

by Richard Howarth

1997, Journal of Geology 105 441 456

We provide a best-fit curve to 1849 strontium isotope data for the period 0 to 206 Ma using the LOcally-WEighted regression Scatterplot Smoother (LOWESS) method. This is a robust, nonparametric modern regression technique. Since it does... more

Figure 1. (a) Robust LOWESS fit to 1849 *’Sr/**Sr data for the period 0 to 206 Ma; curve shows mean *’Sr/*Sr corre- sponding to a given age (Ma) of a sample; solid dots show data points. (b) Half-width of the two-sided 95% confidence interval on mean *’Sr/**Sr (X10°) as a function of age (Ma).

Figure 2. Details of robust LOWESS fit to *’Sr/*°SR. (a) 0-25 Ma; (b) 25-55 Ma; (c) 50-105 Ma; (d) 100-130 Ma; (e) 125-180 Ma, ( f) 175-206 Ma. In each case the bold line defines the mean *’Sr/**Sr, the dashed lines define the 95% confidence interval on the mean predicted *’Sr/**Sr value (these lines merge in figure 2a where uncertainties are small at the scale of the figure). The back- ground grid is drawn at intervals of 2.5 Ma and 0.0001 *’Sr/**Sr.

Figure 3. Diagnostic plots for residuals (i.e. observed—fitted) *’Sr/*°Sr 10%) from the fitted values of the robust LOWESS model: (a) Histogram. (b) Plot of the ordered residuals as a function of the expected values of the standard normal distribution N(0,1); plot would be linear if the residuals were themselves normally distributed. (c) Plot of residuals as a function of predicted *’Sr/**Sr; there is no sign of correlation between the residuals and predicted values.

Figure 4. Boxplots of *’Sr/*°Sr residuals from global robust LOWESS fit (x10~°) by principal author of data sourc« (see table 1 for author codes). Central dotted straight line represents the mean of all data. Width of the boxes is propor tional to the number of data in each subset. Height of the boxes represents the inter-quartile range (IQR), i.e. the central 50% of the data; vertical dashed lines extend from boxes to most outlying data =1.5 X (IQR); individual dot: show outliers. Vertical width of the notches in the sides of the boxes shows approximate 95% confidence interval: on the medians; three subsets show slight upward bias.

Figure 5. Log-log plot of range as a function of mean 87§r/®°Sr (X 10°) for biogenic calcite in 104 biogenic car- bonate samples for which replicate determinations are available; open circles, N, = 2; solid circles, N,= 3. Fitted linear regression of range on mean (dashed line) is virtu- ally horizontal, indicating complete absence of system- atic dependency of range on the mean.

Figure 6. Plot of standard error (6,/Vn) and 95% confi- dence intervals of *’Sr/*°Sr (x10~°) in long-term replicate analyses of NIST 987 ina single laboratory over 4.5 years. Number of replicates in each sub-period (n) varies be- tween 15 and 73. Calculated from data in table 5 of Thirlwall (1991), with addition of 95% confidence inter- vals after Hahn and Meeker (1991).

* Not all data from each author has been included; age ranges denote data used. b See Goure 4. Table 1. Data Sources for Robust LOWESS Fit and As- sociated Uncertainty Based on Quoted Long-term Stan- dard Deviation for Control Materials

Notes. (1) Where estimated mean age is blank, *’Sr/*°Sr corresponds to confidence region only; ND, not determined. (2) A copy of the full table (interpolated in steps of 0.000001 *’Sr/**Sr) may be obtained, either on a 3.5” disk or via e-mail, by contacting the second author at j.mcarthur@ucl.ac.uk. Table 2. Example of Look-up Table

Table 3. Computation of Analytical System Variance for 104 Biogenic Calcite Samples Analyzed with Repli- cate Determinations of §’Sr/**Sr

descriptionView Paper arrow_downwardDownload

Evaluation of satellite based primary production modelling in the semi-arid Sahel

by Inge Sandholt

2006, Remote sensing of …

descriptionView Paper arrow_downwardDownload

Look-Up Table Based Implementations of SHA3 Finalists: JH, Keccak and Skein

by Athar Mahboob

2012

Cryptographic hash functions are widely used in many information security applications like digital signatures, message authentication codes (MACs), and other forms of authentication. In response to recent advances in cryptanalysis of... more

Where A° is most significant and A? is least significant bit of A. Same convention is used for B, C and D. The permutation P of JH is a combination of three small permutations (g, P’ and m). In terms of hardware, permutations are achieved through simple rewiring of nets. Description of these permutations can be found in [7]. A sample two rounds of 64-bit state is shown in Fig. 2. Round constant bits are not shown in Fig. 2. Each JH round uses a 256-bi round constant. For 256 4-bit pairs of JH state, each round constant bit selects either S-box So or S, at respective positions. Round constant for each round may be generated on the fly by using identical bijective compression function of JH in parallel to JH data path and setting initial constant value to zero. Alternately, round constants may be pre-computed and stored in memory to be used later. The initial value of JH state H° is computed by setting first two bytes of H~* equal to the desired hash digest size and remaining bytes to zero, setting message M° to zero and then applying JH compression to H~? and M° to obtain H®. For an N block message final hash HN is calculated by applying JH compression function to message blocks M1, M?, M3,....., MN iteratively as follows: 3.2 Keccak G. Bertoni et al. designed and proposed the K eccak Hash Function for SHA-3 [8]. Keccak is a family of sponge functions with members Keccak [r, c] characterized by two parameters, bitrate r and capacity c. The sum r +c determine the width of the K eccak- f permutation used in the sponge construction and is restricted to values in {25, 50, 100, 200, 400, 800, 1600}. For SHA-3 proposal Keccak team proposed the Keccak [1600] with different r and c values for each desired length of hash output [8]. For 256-bit hash output r = 1088 and c = 512. The 1600-bit state of Keccak [1600] consists of 5x5 matrix of 64-bit words. Each compression step of Keccak consists of 24 rounds. Let us denote the state matrix with A. Each round then consists of following five steps:

Fig. 2. Two sample rounds of JH using 64-bit state size

The block and key sizes of Threefish are equal and can be set to 256, 512 or 1024 bits, and they are designated as: Threefish-256, Threefish-512, and Threefish-1024, respectively. The tweak value is 128 bits for all block sizes. Threefish structural design consists of round operations. Threefish-256 and Threefish-512 compression functions are made of 72 consecutive round operations while the Threefish-1024 requires 80 rounds. Each round of the Threefish-256 block cipher is made of two instances of a MIX function along with a permutation module, while a round key is added to the data before the first round and after each 4 consecutive rounds as shown in Fig. 3. Subkeys or round keys consist of three contributions: an input key word, tweak words, and a counter value. The key schedule turns the key and tweak words into a sequence of subkeys, each of which is equal to the size of the block. Tweak depends upon number of factors including position and the bit length of the message block.

nnn EE EE EE The UBI construction is a variant of the Cascade or (Merkle-Damgard) construction. It uses a tweakable block cipher in Matyas-Meyer-Oseas mode to form a compression function, and uses the bit offset of the block being hashed as the tweak [9]. An example of UBI mode is shown in Fig. 4.

and 5(b) configures the LUT to perform desired operation at its inputs. The INIT value is derived by laying down the truth table for all possible combinations of LUT inputs and outputs. We have used these primitives excessively in architectural designs of SHA-3 finalists. We also exploit the techniques presented in [11] for efficient utilization of modern FPGA resources.

5.1 Common I/O Interface Developed input/output interface is shown in Fig. 6. All I/O transactions are synchronized. Each I/O is sampled at the rising edge of clock cycle. The first_blk and last blk signals indicate the first and last block of message, respectively. The input cycle is initiated by I/O interface by putting load signal to high. Hash Module acknowledges the request if it is able to receive data by putting ack signal to high. After receiving acknowledgment, I/O interface makes available 64-bit word of data at each rising edge of clock cycle. During the transaction of data, ack signal remains at logic high. After receiving desired amount of input words Hash Module puts the ack signal to low. Accordingly I/O interface pulls the load signal to low if no more transactions are required. If message blocks are still present, load signal will remain high but Hash Module acknowledges it after one clock cycle from the previous transaction.

In the same way when Hash Module is ready with a valid hash value it signals the I/O interface by putting Hash Valid signal to high. After putting Hash_ valid signal hash module outputs 64-bit words on each rising edge of clock cycle until the desired hash length is achieved. I/O interface is designed in a way that it does not affect the ongoing processing of hash module. That is, we can make I/O transactions at the same time while hash of a message block is in progress. Fig. 7. Hash Module separated in Control and Data paths

sizes is of initial values (IV) and hash output registers. In the beginning of every hash process state_reg is initialized with IV of desired hash digest size. Then a complete JH compression is processed by setting msg and round constant RC to zero. The higher order 512 bits of resulting state of JH compression is then XORed with first message block and stored in state reg. Then contents of state reg are processed through JH compression function with respective round constant. JH compression function consists of 42 rounds of its arithmetic logic unit (ALU). A single round is processed in one clock cycle. Therefore 42 clock cycles are required to complete 42 rounds of JH compression function. After completion of 42 rounds on a message block, resulting lower order 512 bits of JH compression state is XORed with msg, to obtain next chaining hash value. The higher order 512 bits of resulting chaining hash value is then XORed with next message block and stored in state reg and same compression sequence is repeated again. This process continues till the end of all message JH hash function uses the same algorithm for all hash digest sizes. Hence, same data path is utilized for all hash digest sizes. Only the difference between data path for different hash -

pliMutive Gepicted 1) Pig. LUC). Lt le serect DIC S, seieCts Detwee#l CWO FOtated lWstalCes OF X47, according to round number, to XOR with yp. For first four rounds s is zero and upper half rows of rotation constants’ table are used for respective MIX modules. For next four rounds s will be 1 and lower half rows of rotation constants’ table are used for respective MIX modules. For example, x; « 46 will be selected and XORed with yy in first round and x, « 39 will be selected and XORed with yg in fifth round. Hardware architecture of key schedule module is shown in Fig 10(b). The extended key Kg is obtained by XORing the input 64-bit key words (Ko... K7) and constant C249. The extended tweak t, is obtained by XORing the two input 64-bit tweak word (ty and t,). The extended key and tweak words are then loaded into the circular shift registers K (576 bit) and t (192 bit). These two registers are clocked and rotated once for each subkey. Key Schedule module generates subkeys on every falling edge of clock pulse. Add Subkey module gives output on the rising edge of each clock pulse. Next subkey is available on falling edge of the same clock pulse. In this way one clock cycle is required to complete four rounds, subkey addition and subkey generation. Therefore to complete 72 rounds and 19 subkey addition of Skein, 19 clock cycles will be required. The next chaining hash value will be available after 19 clock cycles.

Table 2. Implementation Results for 256-bit and 512-bit variants of SHA-3 finalists For both Virtex 5 and Virtex 6, designs of JH and Keccak result in higher frequencies, due to their simple compression functions. However, more number of rounds make JH algorithm’s throughput restricted to the lower end. The larger message block size and average number of rounds of Keccak make it the top performer in terms of both throughput and throughput per area. In terms of area consumption, JH is the most compact algorithm. BLAKE and Skein are computationally rigorous algorithms as compared to other candidates, hence, results in low throughput and throughput per area designs.

Table 3. Comparison with previous work. Fy, in MHz, Area in Slices and TPA in Mbps/Slice

descriptionView Paper arrow_downwardDownload

A new model of gross primary productivity for North American ecosystems based solely on the enhanced vegetation index and land surface temperature from MODIS

by Walter Oechel

2008, Remote Sensing of Environment

Many current models of ecosystem carbon exchange based on remote sensing, such as the MODIS product termed MOD17, still require considerable input from ground based meteorological measurements and look up tables based on vegetation type.... more

descriptionView Paper arrow_downwardDownload

Fast Combinatorial RNS Processors for DSP Applications

by Gianni Orlandi

1995, IEEE Transactions on Computers - TC

It is known that RNS VLSI processors can parallelize fixed-point addition and multiplication operations by the use of the Chinese remainder theorem (CRT). The required modular operations, however, must use specialized hardware whose... more

descriptionView Paper arrow_downwardDownload

Simple Dynamic Overmodulation Strategy for Fast Torque Control in DTC of Induction Machines with Constant Switching Frequency Controller

by Tole Tole

2010

This paper presents a simple dynamic overmodulation method to obtain a fast dynamic torque response in DTC of induction machines with constant switching frequency controller. A fast dynamic torque response can be obtained by switching... more

descriptionView Paper arrow_downwardDownload

SENSORLESS SPEED ESTIMATION OF INDUCTION MOTOR IN A DIRECT TORQUE CONTROL SYSTEM

by Khoa Dang Hoang

2006

Fast and robust torque control in a very wide range of speed is very needed by various industrial AC drive applications. Therefore, since 1986 [1], direct torque control (DTC) has been introduced to satisfy this desire. In order to... more

descriptionView Paper arrow_downwardDownload

Image encryption with dynamic chaotic Look-Up Table

by Med Karim Abdmouleh and

2012, 2012 6th International Conference on Sciences of Electronics, Technologies of Information and Telecommunications (SETIT)

In this paper we propose a novel image encryption scheme. The proposed method is based on the chaos theory. Our cryptosystem uses the chaos theory to define a dynamic chaotic Look-Up LUT) to compute the new value of the current pixel to... more

descriptionView Paper arrow_downwardDownload

Inversion of a coupled canopy-atmosphere model using multi-angular top-of-atmosphere radiance data: A forest case study

by Wouter Verhoef and

2011, Remote Sensing of Environment

Since the launch of sensors with angular observation capabilities, such as CHRIS and MISR, the additional potential of multi-angular observations for vegetation structural and biochemical variables has been widely recognised. Various... more

Since the launch of sensors with angular observation capabilities, such as CHRIS and MISR, the additional potential of multi-angular observations for vegetation structural and biochemical variables has been widely recognised. Various methods have been successfully implemented to estimate forest biochemical and biophysical variables from atmospherically-corrected multi-angular data, but the use of physically based radiative transfer (RT) models is still limited. Because both canopy and atmosphere have an anisotropic behaviour, it is important to understand the multi-angular signal measured by the sensor at the top of the atmosphere (TOA). Coupled canopy-atmosphere RT models allow linking surface variables directly to the TOA radiance measured by the sensor and are therefore very interesting tools to use for estimating forest variables from multi-angular data. We investigated the potential of TOA multi-angular radiance data for estimating forest variables by inverting a coupled canopy-atmosphere physical RT model. The case study focussed on three Norway spruce stands located at the Bily Kriz experimental site (Czech Republic), for which multi-angular CHRIS and field data were acquired in September 2006. The soil-leaf-canopy RT model SLC and the atmospheric model MODTRAN4 were coupled using a method allowing to make full use of the four canopy angular reflectance components provided by SLC. The TOA radiance simulations were in good agreement with the spectral and angular signatures measured by CHRIS. Singular value decompositions of the Jacobian matrices showed that the dimensionality of the variable estimation problem increased from 3 to 6 when increasing the number of observation angles from 1 to 4. The model inversion was conducted for two cases: 4 and 7 variables. The most influential parameters were chosen as free variables in the look-up tables, namely: vertical crown cover (Cv), fraction of bark material (fB), needle chlorophyll content (needleCab), needle dry matter content (needleCdm) for the 4-variable case, and additionally, tree shape factor (Zeta), dissociation factor (D), and needle brown pigments content (needleCs) in the 7-variable case. All angular combinations were tested, and the best estimates were obtained with combinations using two or three angles, depending on the number of variables and on the stand used. Overall, this case study showed that, although making use of its full potential is still a challenge, TOA multi-angular radiance data do have a higher potential for variable estimation than mono-angular data.

descriptionView Paper arrow_downwardDownload

Multipoint phase calibration for improved compensation of inherent wavefront distortion in parallel aligned liquid crystal on silicon displays

by Joaquín Otón

2007, Applied Optics

The inherent distortion of a reflective parallel aligned spatial light modulator (SLM) may need compensation not only for the backplane curvature but also for other possible nonuniformities caused by thickness variations of the liquid... more

descriptionView Paper arrow_downwardDownload

A Data-Driven Approach to A Priori SNR Estimation

by Tim Fingscheidt

2000, IEEE Transactions on Audio, Speech, and Language Processing

The a priori signal-to-noise ratio (SNR) plays an important role in many speech enhancement algorithms. In this paper we present a data-driven approach to a priori SNR estimation. It may be used with a wide range of speech enhancement... more

descriptionView Paper arrow_downwardDownload

Satellite-derived parameters for biological modelling in coastal waters: Illustration over the eastern continental shelf of the Bay of Biscay

by Francis Gohin

2005, Remote Sensing of Environment

In biological modelling of the coastal phytoplankton dynamics, the light attenuation coefficient is often expressed as a function of the concentrations of chlorophyll and mineral suspended particulate matter (SPM). In order to estimate... more

descriptionView Paper arrow_downwardDownload

Stabilization of switched affine systems: An application to the buck-boost converter

by Daniel Corona

2007

In this paper we extend a technique developed to design a feedback stabilizing control law for autonomous switched systems all modes of which are unstable. More specifically, we extend the switching table procedure to the class of affine... more

descriptionView Paper arrow_downwardDownload

Goal Evaluation of Segmentation Algorithms for Traffic Sign Recognition

by Pedro Gil Jiménez

2010, IEEE Transactions on Intelligent Transportation Systems

This paper presents a quantitative comparison of several segmentation methods (including new ones) that have successfully been used in traffic sign recognition. The methods presented can be classified into color-space thresholding, edge... more

descriptionView Paper arrow_downwardDownload

Look-Up Table Based Implementations of SHA-3 Finalists: JH, Keccak and Skein

by Athar Mahboob and

2012, KSII Transactions on Internet and Information Systems

Cryptographic hash functions are widely used in many information security applications like digital signatures, message authentication codes (MACs), and other forms of authentication. In response to recent advances in cryptanalysis of... more

descriptionView Paper arrow_downwardDownload

Fuzzy controller design for parallel hybrid vehicle analysis using forward simulation

by Mojtaba Mirsalim

2009, 2009 IEEE Vehicle Power and Propulsion Conference

In order to convert conventional vehicles to their hybrid counterparts, trustworthy simulation with driving realities are essential. Here, a seven-degree-of-freedom model is developed to simulate the dynamical behavior of vehicles. Next,... more

descriptionView Paper arrow_downwardDownload

<title>An overview of the holographic display related tasks within the European 3DTV Project</title>

by Atanas Gotchev

2006, Photon Management II

A European consortium has been working since September 2004 on all video-based technical aspects of threedimensional television. The group has structured its technical activities under five technical committees focusing on capturing 3D... more

descriptionView Paper arrow_downwardDownload

Analysis of a 12/16 switched reluctance machine using combined circuit and field computation

by Amissa Arifin

2011

Field analysis in conjunction with circuit model has been used to predict the performance of a switched reluctance machine. This paper describes the procedure to model a 3 phase 12/16 Switched Reluctance (SR) machine. Its operation as a... more

Figure 1. Flow chart on ways to model/analyze SRM performance/behavior data is required to build the neural network model. The above technique basically gives estimation for the response of the system.

Figure 2. Representation of one phase using NN or LUT The 5tb International Power Engineering and Optimization Conf. (PEOCO2011), Shah Alam, Selangor, MALAYSIA: 6-7 June 2011 Il. MATHEMATICAL MODEL OF SWITCHED RELUCTANCE GENERATOR The SR drive system can be represented as a mathematical model comprising of electrical and mechanical section. Electrical section is built based on the phase voltage equation whereas the mechanical equation is governed by the movement equation:

The calculation of phase flux linkage for several rotor positions is implemented using COMSOL Multiphysics by subdomain integrations over the coil areas and can be represented as in Fig. 5. In this study, 19 curves for a 12/16 machine at various rotor positions from aligned at 11.25deg to unaligned at Odeg in steps of 0.625deg intervals are performed. The windings are subjected to current from 0-5A at 0.25A intervals. where | is the stack length, N is the number of turns and S area of phase winding.

Figure 6. Integration method using Simpson’s 1/3 rule for co-energy calculation Using MATLAB/SIMULINK the calculation of co-energy is as shown in Fig. 7. From the co-energy, calculation of torque is obtained using equation (6) as shown in Fig. 8. The range of co-energy data used is 2,704 ranging from 3.125deg to 28.75deg with intervals of 0.625deg.

Figure 10. Phase current and voltage for (a) NSNS and (b) NNSS winding polarity In this study the calculation of ratio between generation current and excitation current is performed for one current cycle. The cycle is separated as in Fig. 11(a) excitation stage and Fig. 11(b) generation stage. Ratio between generation stage and excitation stage is calculated for various speed ranges as shown in Fig. 12. Results shows that machine with polarity NNSS have higher ratio of generation current to excitation current. Based on this study the NNSS phase winding polarity gives better performance in terms of electrical efficiency. By changing the phase winding polarity of the machine, an optimum performance in terms of efficiency can

Figure 9. Phase winding polarity of (a) NSNS and (b)NNSS The magnetization data is then used as LUT in MATLAB/SIMULINK to determine performance of the machine. Using the same machine parameters for both winding polarity, ratio between generation current and excitation

The 5th International Power Engineering and Optimization Conf. (PEOCO2011), Shah Alam, Selangor, MALAYSIA: 6-7 June 2011

Figure 11. Phase current during 1 cycle showing the excitation stage (a) and generation stage (b) of the NNSS phase winding polarity The 5th International Power Engineering and Optimization Conf. (PEOCO2011), Shah Alam, Selangor, MALAYSIA: 6-7 June 2011

Figure 12. Ratio between generated and excitation current for NNSS and NSNS winding polarity NOINO WIS PULALILY be achieved. This is mainly contributed by the flux pathway as a result of 4 poles per phase excitation. The pathway of flux is clearly shown for NSNS and NNSS phase winding polarity in Fig. 13(a) and (b). It can be seen that the flux path for NNSS winding polarity is half of the NSNS configuration, leading to a more confined magnetic flux path. Having a longer magnetic path in the stator yoke leads to a higher magnetic potential drop hence lower flux linkage. This finding ties with the lowe1 magnetization curve obtained in Fig. 5 as compared to NNSS phase winding configuration.

polarity for a four phase per pole excitation gives higher ratio of generated current to excitation current. However for higher speed its efficiency is reduced due to amount of time required for current to build up. Different phase winding strategies plays a role in determining higher performance for the machine. Future work studies the effect of performance for multiphase excitation.

descriptionView Paper arrow_downwardDownload

Composite field multiplier based on look-up table for elliptic curve cryptography implementation

by Kuspriyanto Kuspriyanto and

2011, Electrical …

Implementing a secure cryptosystem requires operations involving hundreds of bits. One of the most recommended algorithm is Elliptic Curve Cryptography (ECC). The complexity of elliptic curve algorithms and parameters with hundreds of... more

descriptionView Paper arrow_downwardDownload

Application of spheroid models to account for aerosol particle nonsphericity in remote sensing of desert dust

by Erfin Elly

2006, Journal of Geophysical Research

1] The possibility of using shape mixtures of randomly oriented spheroids for modeling desert dust aerosol light scattering is discussed. For reducing calculation time, look-up tables were simulated for quadrature coefficients employed in... more

1] The possibility of using shape mixtures of randomly oriented spheroids for modeling desert dust aerosol light scattering is discussed. For reducing calculation time, look-up tables were simulated for quadrature coefficients employed in the numerical integration of spheroid optical properties over size and shape. The calculations were done for 25 bins of the spheroid axis ratio ranging from $0.3 (flattened spheroids) to $3.0 (elongated spheroids) and for 41 narrow size bins covering the size parameter range from $0.012 to $625. The look-up tables were arranged into a software package, which allows fast, accurate, and flexible modeling of scattering by randomly oriented spheroids with different size and shape distributions. In order to evaluate spheroid model and explore the possibility of aerosol shape identification, the software tool has been integrated into inversion algorithms for retrieving detailed aerosol properties from laboratory or remote sensing polarimetric measurements of light scattering. The application of this retrieval technique to laboratory measurements by has shown that spheroids can closely reproduce mineral dust light scattering matrices. The spheroid model was utilized for retrievals of aerosol properties from atmospheric radiation measured by AERONET ground-based Sun/sky-radiometers. It is shown that mixtures of spheroids allow rather accurate fitting of measured spectral and angular dependencies of observed intensity and polarization. Moreover, it is shown that for aerosol mixtures with a significant fraction of coarse-mode particles (radii ! $1 mm), the nonsphericity of aerosol particles can be detected as part of AERONET retrievals. The retrieval results indicate that nonspherical particles with aspect ratios $1.5 and higher dominate in desert dust plumes, while in the case of background maritime aerosol spherical particles are dominant. Finally, the potential of using AERONET derived spheroid mixtures for modeling the effects of aerosol particle nonsphericity in other remote sensing techniques is discussed. For example, the variability of lidar measurements (extinction to backscattering ratio and signal depolarization ratio) is illustrated and analyzed. Also, some potentially important differences in the sensitivity of angular light scattering to parameters of nonspherical versus spherical aerosols are revealed and discussed. Citation: Dubovik, O., et al. (2006), Application of spheroid models to account for aerosol particle nonsphericity in remote sensing of desert dust,

descriptionView Paper arrow_downwardDownload

Recurrent fuzzy neural network by using feedback error learning approaches for LFC in interconnected power system

by Vijitha Karandagolle

2009, Energy Conversion and Management

In this study, a new adaptive controller based on modified feedback error learning (FEL) approaches is proposed for load frequency control (LFC) problem. The FEL strategy consists of intelligent and conventional controllers in feedforward... more

descriptionView Paper arrow_downwardDownload

MODIS-based Mosaic of Antarctica (MOA) data sets: Continent-wide surface morphology and snow grain size

by Theodore Scambos

2007, Remote Sensing of Environment

We present digital image mosaics of surface morphology and optical snow grain size for the Antarctic continent and surrounding islands, assembled from 260 Moderate-resolution Imaging Spectroradiometer (MODIS) images. The products are... more

We present digital image mosaics of surface morphology and optical snow grain size for the Antarctic continent and surrounding islands, assembled from 260 Moderate-resolution Imaging Spectroradiometer (MODIS) images. The products are derived from MODIS band 1 (red light: ∼ 647 nm) and band 2 (infrared: ∼ 857 nm) orbit swath data acquired during the 2003-2004 austral summer. Multiple images of all areas are combined in a data cumulation scheme to improve spatial resolution and increase radiometric content of the mosaics. The component images were de-striped, geo-registered, and re-sampled to the projection grid using the MODIS Swath-to-Grid Toolbox (MS2GT) software. A 125 m groundequivalent polar stereographic projection was used, identical to previous radar image mosaics and similar to several other continent-wide data sets. Geo-location accuracy of the final mosaics is better than 125 m. Swath acquisition times were limited to the range 05:00-13:30 UT to provide broad but uniform mosaic illumination. Regions of cloud cover, blowing snow, and intense forward scattering were masked prior to compositing. The MOA snow grain size data set image provides the first continent-wide semi-quantitative map of mean summer snow, firn, and blue ice grain size. Optical grain size (radius) was determined using a normalized difference index derived from MODIS band 1 and band 2 rad values. Atmospheric correction, adjustments for directional reflectance variation and snow grain size estimation were accomplished by a look-up table, generated by a combination of published models of snow spectral reflectance and atmosphere radiative transfer. Validation of the snow grain size image is provided by in situ spectra of snow-covered sea ice made in October, 2003. We use the new data sets to determine major ice shelf areas, and shelf edge, coastline, and ice grounding line vector files, and determine areas of selected blue ice regions. These data sets are available for download at http://nsidc.org/data/moa/.

descriptionView Paper arrow_downwardDownload

Influence of the architecture on ADC error modeling

by Pasquale Arpaia and

1999, IEEE Transactions on Instrumentation and Measurement

The influence of the architecture on analog-to-digital converter modeling is investigated for the three most widespread architectures: integrating, successive approximations, and flash. The effects of main error sources are analyzed in... more

descriptionView Paper arrow_downwardDownload

Portable real-time color night vision

by Alexander Toet

2008, In B V Dasarathy Multisensor Multisource Information Fusion Architectures Algorithms and Applications 2008 Spie Dss08 Ds16 2 Spie 6974 Bellingham Wa Usa the International Society For Optical Engineering 1 12

We developed a simple and fast lookup-table based method to derive and apply natural daylight colors to multi-band night-time images. The method deploys an optimal color transformation derived from a set of samples taken from a daytime... more

descriptionView Paper arrow_downwardDownload

Modeling human color categorization: color discrimination and color memory

by Egon L. van den Broek

2003, BMC Musculoskeletal Disorders

Color matching in Content-Based Image Retrieval is done using a color space and measuring distances between colors. Such an approach yields non-intuitive results for the user. We introduce color categories (or focal colors), determine... more

descriptionView Paper arrow_downwardDownload

Look up Table

Related Topics