Classifying means of transportation using mobile sensor data
ABSTRACT Mobile phone sensors have the ability to provide significant information about environme... more ABSTRACT Mobile phone sensors have the ability to provide significant information about environmental conditions as well as about different activities of persons. This paper deals with the acceleration sensor of a commercial mobile phone that makes it possible to identify and classify the user's means of transportation. The needed pre-processing of the sensor data is described as well as the used classification algorithms, i.e. Naive Bayes classifier and Support Vector Machine (SVM). Their performance for solving the classification problem in combination with the different pre-processings is analyzed. It is shown that both classifiers are able to solve the classification task with high accuracy given a proper preprocessing. Support Vector Machines outperform Naive Bayes classifiers and achieve a classification accuracy of over 97% on an unknown test data set.
An extended version of low-complexity IP Core for image/video transformations based on the CORDIC... more An extended version of low-complexity IP Core for image/video transformations based on the CORDIC architecture is presented. This IP core is able to perform quantized 8×8 IDCT and quantized 8×8/4×4 H.264-inverse integer transforms on a configurable architecture by using only shift and add operations. Furthermore, the number for CORDIC iterations and compensation steps can be adjusted, which enables a trade-off between video compression quality in PSNR and computational complexity.
ABSTRACT The Jacobi method for singular value decomposition is well-suited for parallel architect... more ABSTRACT The Jacobi method for singular value decomposition is well-suited for parallel architectures. Its application to signal subspace computations is well known. Basically the subspace spanned by singular vectors of large singular values are separated from subspace spanned by those of small singular values. The Jacobi algorithm computes the singular values and the corresponding vectors in random order. This requires sorting the result after convergence of the algorithm to select the signal subspace. A modification of the Jacobi method based on a linear objective function merges the sorting into the SVD-algorithm at little extra cost. In fact, the complexity of the diagonal processor cells in a triangular array get slightly larger. In this paper we present these extensions, in particular the modified algorithm for computing the rotation angles and give an example of its usefulness for subspace separation.
Improving the performance of UHF RFID communication in a small scale fading channel
ABSTRACT Radio Frequency Identification (RFID) is mostly used in harsh environments were deep fad... more ABSTRACT Radio Frequency Identification (RFID) is mostly used in harsh environments were deep fades occur. Taking into account a small scale fading channel and the UHF RFID transmission specified by the the EPCglobal protocol a poor communication performance is observed. As the decoding process of the received information is not specified in this protocol, there is a significant potential of increasing the transmission performance with an appropriate decoding. This work will introduce a novel possibility to decode the received information. To match the decoding process with the transmission behavior, it will be shown how the system with a basic decoding scheme performs. Based on these results several modifications, like amplitude threshold, differential decoding and the modification of the used codes will be presented. These modifications take the channel behavior and its characteristic into consideration and lead to a performance gain of 26.2 dB using a certain encoding scheme. These results will be supported by the analysis of the bit error characteristic to demonstrate how the transmission is influenced by the modifications. Note that most modifications presented in this work can be implemented on the RFID reader side. Only the modification of the tag encoding sequence requires a change in the transponder structure.
<title>Parallel SVD updating using approximate rotations</title>
Proceedings of SPIE, Jun 7, 1995
ABSTRACT In this paper a parallel implementation of the SVD--updating algorithm using approximate... more ABSTRACT In this paper a parallel implementation of the SVD--updating algorithm using approximate rotations is presented. In its original form the SVD--updating algorithm had numerical problems if no reorthogonalization steps were applied. Representing the orthogonal matrix V (right singular vectors) using its parameterization in terms of the rotation angles of n(n Gamma 1)=2 plane rotations these reorthogonalization steps can be avoided during the SVD-updating algorithm [18]. This results in a SVD--updating algorithm where all computations (matrix vector multiplication, QRD--updating, Kogbetliantz&#39;s algorithm) are entirely based on the evaluation and application of orthogonal plane rotations. Therefore, in this form the SVD--updating algorithm is amenable to an implementation using CORDIC--based approximate rotations. Using CORDIC--based approximate rotations the n(nGamma1)=2 rotations representing V (as well as all other rotations) are only computed to a certain approximation accuracy (in t...
Due to the increased use of Radio Frequency Identification (RFID) in different fields of applicat... more Due to the increased use of Radio Frequency Identification (RFID) in different fields of application it is reasonable to explore the benefit that can be obtained by the simultaneous localization of RFID tags. This paper describes the localization of a passive UHF RFID tag via Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) using the Received Signal Strength Indicator (RSSI) values. Simulation results based on measurements show that UKF achieves higher localization accuracies than EKF.
A Low-Power and High-Quality Cordic Based Loeffler DCT
In this paper, a low-power and high-quality preserving DCT architecture is presented. It is obtai... more In this paper, a low-power and high-quality preserving DCT architecture is presented. It is obtained by optimizing the Loeffler DCT based on the Cordic algorithm. The experimental results show that the proposed DCT architecture occupies 19% of the area and consumes about 16% of the power compared to the original Loeffler DCT. Additionally the good transformation quality is retained. In this regard, the proposed Cordic based Loeffler DCT is very suitable for low-power and high-quality CODECs, especially for battery-based systems
Radio Frequency Identification (RFID) is a wireless communication method mostly used in rough ind... more Radio Frequency Identification (RFID) is a wireless communication method mostly used in rough indoor environments. In such environments Forward Error Correction (FEC) is a popular method to improve the transmission quality. However, the common RFID protocol (EPCglobal) only provides error detection based on Cyclic Redundancy Check (CRC) codes. The replacement of this code with an arithmetically similar FEC code improves the transmission and requires only minor changes to the protocol structure. A BCH code fulfills this requirement. Nevertheless, this code has the disadvantage that it uses only hard-coded bits for the decoding process. This work presents a BCH decoding using the Chase algorithm which also takes the soft information of the received sequence into account. It is shown that a coding gain of 1 dB is achievable in an RFID transmission application, compared to a transmission with a common BCH code. Also an estimation of the computational complexity when using the Chase algorithm is given as this complexity generally increases with this modification.
The localization of objects equipped with passive UHF RFID labels using the Received Signal Stren... more The localization of objects equipped with passive UHF RFID labels using the Received Signal Strength Indicator (RSSI) values is investigated. An Unscented Kalman Filter (UKF) is used for the localization of the object. Adding (N-1) additional tags with known relative position to the object (relative to its main tag), this information can be incorporated into the UKF resulting in an Unscented Kalman Filter with relative position information (UKF_rp). Simulation results based on measured RSSI values show that the usage of relative positioning leads to an increase in localization accuracy especially when few antennae are used.
Performance of BCH Codes and Transmit Diversity for MISO UHF RFID Communication
ABSTRACT Radio Frequency Identification (RFID) is used in environments with significant fading ch... more ABSTRACT Radio Frequency Identification (RFID) is used in environments with significant fading characteristic. In this paper the Electronic Product Code (EPC) Class-1 Gen-2 protocol is extended by spatial diversity (SD) and Forward Error Correction (FEC) in order to improve the transmission reliability. A dyadic backscatter channel is used. The respective channel model gives a good approximation for the RFID small scale fading characteristic. Based on simulations of the bit error-rate (BER) and the packet error-rate (PER) the improvement using FEC (a BCH code) and SD (with two tag antennas) are presented and the trade-off between FEC and SD is discussed. It is shown that a BCH encoded transmission can lead to 8 dB coding gain which is relatively constant for varying channel parameters. Higher gains (up to 27 dB) can be achieved by using two tag antennas, but varying the channel parameters results in a higher variance of the gain due to different correlation scenarios.
European Signal Processing Conference, Aug 1, 2011
A new design concept for accelerating Sparse Matrix-Vector Multiplication (SMVM) in FPGA by using... more A new design concept for accelerating Sparse Matrix-Vector Multiplication (SMVM) in FPGA by using Network-on-Chip (NoC) is presented. In traditional circuit design on-chip communications have been designed with dedicated pointto-point interconnections or shared buses. Therefore, regular data transfer is the major concern of many parallel implementations. However, when dealing with the SMVM operation, which is the main step of most iterative algorithms for solving systems of linear equations, the required data transfers are usually dependent on the sparsity structure of the matrix and can be extremely irregular. Using a NoC architecture makes it possible to deal with arbitrary structure of the data transfers, i.e. with arbitrary structured sparse matrices. In this paper, a configurable interface is presented which can generate the pipelined SMVM calculator based on NoC architecture with size of 2×2, 4×4,. . ., p× p (p ∈ N). The implementation is done in IEEE-754 single floating-point precision on the Xilinx Virtex-6 FPGA.
European Signal Processing Conference, Sep 4, 2006
In this paper a computationally efficient and high-quality preserving DCT architecture is present... more In this paper a computationally efficient and high-quality preserving DCT architecture is presented. It is obtained by optimizing the Loeffler DCT based on the Cordic algorithm. The computational complexity is reduced from 11 multiply and 29 add operations (Loeffler DCT) to 38 add and 16 shift operations (which is similar to the complexity of the binDCT). The experimental results show that the proposed DCT algorithm not only reduces the computational complexity significantly, but also retains the good transformation quality of the Loeffler DCT. Therefore, the proposed Cordic based Loeffler DCT can be used in low-power and high-quality CODECs, especially in battery-based systems.
Power and signal transmission system for radio communication equipment involves transmitter sending first signal to first circuit and second transmitter sending second signal to second circuit
Uploads
Papers by Jürgen Götze