Academia.eduAcademia.edu

Figure 111 – uploaded by wendy moody

See full PDFdownloadDownload figure
Experimental Prototype

Figure 111 Experimental Prototype

Related Figures (136)

Fig. 1. Early US newspaper extract featuring the DataGlove concept and hinting at future roboti: applications and haptic feedback variants. Two of relevance to the emergence of haptic feedback developments are teleoperation — the remote location and extension of a person’s sensing and manipulation capability to a telepresence — the ideal of sensing sufficient information about the teleoperator and task environment, and communicating this to the human operator in a sufficiently natural way, that the operator feels physically present at the remote site. The “Holy Grail” of te human-system inter VIEW (eg. Fisher epresence also provided the motivation behind some of the early face efforts underpinning NASA’s Virtual Environment Workstation, et al. [8], which included investigations of basic glove-mounted vibrotactile feedbac k transducers; see Fig. 1), and the commercial VR aspirations of the late VPL Inc with its flagship product, the DataGlove.
Fig. 1. Early US newspaper extract featuring the DataGlove concept and hinting at future roboti: applications and haptic feedback variants. Two of relevance to the emergence of haptic feedback developments are teleoperation — the remote location and extension of a person’s sensing and manipulation capability to a telepresence — the ideal of sensing sufficient information about the teleoperator and task environment, and communicating this to the human operator in a sufficiently natural way, that the operator feels physically present at the remote site. The “Holy Grail” of te human-system inter VIEW (eg. Fisher epresence also provided the motivation behind some of the early face efforts underpinning NASA’s Virtual Environment Workstation, et al. [8], which included investigations of basic glove-mounted vibrotactile feedbac k transducers; see Fig. 1), and the commercial VR aspirations of the late VPL Inc with its flagship product, the DataGlove.
a scissor, pistol, or similar control grip device), the two components being linked by means of chains, cables or some other electromechanical motion system. “Mini- masters”, such as that proposed in the 1980s for the original NASA Flight Telerobotic Servicer and other remotely controlled space, subsea and land vehicles are, as the name suggests, small master controllers. These may or may not be kinematically similar to the remote slave device and have met with mixed levels of success when applied to laboratory or field demonstrators. By far the most publicised use of master control arms for Virtual Reality applications has been for molecular modelling (the well-known GROPE IIIb Project) and haptic interaction with electrostatic molecule-substrate force simulations and nano-level surfaces (generated from Scanning Tunnelling Microscope data) at the University of North Carolina at Chapel Hill (eg. Brooks [9]; Brooks et al. [10]; see Fig. 2). Early work at UNC utilised an Argonne Remote Manipulator (ARM) system, one of two donated from the Argonne National Laboratory, and a field sequential computer screen (based on liquid crystal shutter glasses). Later, the screen was replaced with a projection display, with users of the ARM interacting with 3D images produced using polarised projection display lenses and spectacles. Compared with mechanical MSMs. seryvomagninulators have the advantasces of
a scissor, pistol, or similar control grip device), the two components being linked by means of chains, cables or some other electromechanical motion system. “Mini- masters”, such as that proposed in the 1980s for the original NASA Flight Telerobotic Servicer and other remotely controlled space, subsea and land vehicles are, as the name suggests, small master controllers. These may or may not be kinematically similar to the remote slave device and have met with mixed levels of success when applied to laboratory or field demonstrators. By far the most publicised use of master control arms for Virtual Reality applications has been for molecular modelling (the well-known GROPE IIIb Project) and haptic interaction with electrostatic molecule-substrate force simulations and nano-level surfaces (generated from Scanning Tunnelling Microscope data) at the University of North Carolina at Chapel Hill (eg. Brooks [9]; Brooks et al. [10]; see Fig. 2). Early work at UNC utilised an Argonne Remote Manipulator (ARM) system, one of two donated from the Argonne National Laboratory, and a field sequential computer screen (based on liquid crystal shutter glasses). Later, the screen was replaced with a projection display, with users of the ARM interacting with 3D images produced using polarised projection display lenses and spectacles. Compared with mechanical MSMs. seryvomagninulators have the advantasces of
Until quite recently, the exoskeleton concept had been unpopular, due to limitation in the functional anatomy of the human arm. Also, force-reflecting actuators had to b mounted on the outside of the exoskeletal framework to accommodate the users’ arm Furthermore, there were concerns with such devices’ small operating volume, possibl safety hazards (associated with toppling and locking) and electro-mechanical inefficienc (see also Wilson [12]; Salisbury [13]). Nevertheless, thanks in part to the emergence of range of lightweight, low-cost body systems developed under the VR banner, th exoskeleton received renewed interest in the early 1990s as a means of registering bod movement in a virtual environment and, importantly, as a technique for feeding hapti data back to the immersed user (eg. Bergamasco [14]; see Fig. 4). However, even to this day, exoskeletons have been confined to academic research labs or industrial design organisations (eg. Zechner and Zechner in Austria; see Fig. 5) and noticeably absent from commercial catalogues. Witness the fate of the pioneering
Until quite recently, the exoskeleton concept had been unpopular, due to limitation in the functional anatomy of the human arm. Also, force-reflecting actuators had to b mounted on the outside of the exoskeletal framework to accommodate the users’ arm Furthermore, there were concerns with such devices’ small operating volume, possibl safety hazards (associated with toppling and locking) and electro-mechanical inefficienc (see also Wilson [12]; Salisbury [13]). Nevertheless, thanks in part to the emergence of range of lightweight, low-cost body systems developed under the VR banner, th exoskeleton received renewed interest in the early 1990s as a means of registering bod movement in a virtual environment and, importantly, as a technique for feeding hapti data back to the immersed user (eg. Bergamasco [14]; see Fig. 4). However, even to this day, exoskeletons have been confined to academic research labs or industrial design organisations (eg. Zechner and Zechner in Austria; see Fig. 5) and noticeably absent from commercial catalogues. Witness the fate of the pioneering
Fig. 7. The Teletact Commander hand controller, showing the simple layout of some of the ai pockets and VR interactive control buttons. virtual object. These pockets were controlled either by compressor or by a single solenoid-actuated piston. Other haptic developments at the ARRC included a prototype minimally invasive surgery haptic feedback system, funded by the Department of Health and Wolfson Foundation (Fig. 8). This device actually pre-dated the Immersion Corporation Impulse Engine and used basic strain gauge, potentiometer and servomotor devices to provide position sensing and feedback to a laparoscopic instrument in 3 translational degrees of freedom, with grip/forceps actuation. A simple wire frame cube provided the test environment, hosted on a 486 PC
Fig. 7. The Teletact Commander hand controller, showing the simple layout of some of the ai pockets and VR interactive control buttons. virtual object. These pockets were controlled either by compressor or by a single solenoid-actuated piston. Other haptic developments at the ARRC included a prototype minimally invasive surgery haptic feedback system, funded by the Department of Health and Wolfson Foundation (Fig. 8). This device actually pre-dated the Immersion Corporation Impulse Engine and used basic strain gauge, potentiometer and servomotor devices to provide position sensing and feedback to a laparoscopic instrument in 3 translational degrees of freedom, with grip/forceps actuation. A simple wire frame cube provided the test environment, hosted on a 486 PC
Fig. 13. Early experiments with visual cueing of haptic events, in this case showing “ghosted” pipe images used to convey collisions with adjacent aerospace engine components.
Fig. 13. Early experiments with visual cueing of haptic events, in this case showing “ghosted” pipe images used to convey collisions with adjacent aerospace engine components.
Fig. 1. An example of one of the PVC overlays used within the Weasel music notation system.
Fig. 1. An example of one of the PVC overlays used within the Weasel music notation system.
Fig. 2. An example of one of the new Weasel overlays.
Fig. 2. An example of one of the new Weasel overlays.
presentation of and interaction with virtual obj ects. Consequently, one cannot assume that the findings from experiments into the haptic perception of real objects will apply to the haptic perception of virtual objects. To conducted several experiments investigating t investigate this issue, Colwell et al [1] he perception of virtual objects and the virtual representation of texture using a three-degree of freedom device (i.e. one which allows three-dimensional interaction with virtual objects), the Impulse Engine 3000 (IE 3000) from Immersion Corporation www.mmersion.com) (Fig. 1). Texture was chosen as it is a highly salient object attribute and therefore could be a useful means of representing information in virtual environments incorporating haptics.
presentation of and interaction with virtual obj ects. Consequently, one cannot assume that the findings from experiments into the haptic perception of real objects will apply to the haptic perception of virtual objects. To conducted several experiments investigating t investigate this issue, Colwell et al [1] he perception of virtual objects and the virtual representation of texture using a three-degree of freedom device (i.e. one which allows three-dimensional interaction with virtual objects), the Impulse Engine 3000 (IE 3000) from Immersion Corporation www.mmersion.com) (Fig. 1). Texture was chosen as it is a highly salient object attribute and therefore could be a useful means of representing information in virtual environments incorporating haptics.
Fig. 2: The Phantom 1.0A haptic device.
Fig. 2: The Phantom 1.0A haptic device.
Table 1. Keyboard Commands
Table 1. Keyboard Commands
Fig. 4. Flat Rectangular Polygon with Hole for Function.
Fig. 4. Flat Rectangular Polygon with Hole for Function.
Fig. 5. Deformation of Hole to Contain X-axis Shape.
Fig. 5. Deformation of Hole to Contain X-axis Shape.
Several line graphs were created for this experiment (Figure 2). In common, two lines were displayed on a graph and they were either located separately or cross- intersecting each other. Two different friction properties were applied to the lines and they were classified as sticky and slippery. The lines were made up by cylinders, which are one of the primitive shapes supported by the GHOST programming SDK. Therefore, straight-line approximation was used to construct simple line graphs. All the lines were half sub-merged into a flat surface and thus a semi-circle cross-section was formed on all the line models. Due to the circular cross-section of cylinders, users can feel the jagged corner at the joints of the graph lines. Both sighted and blind people were recruited as participants for this experiment. Ten sighted participants were used and nine were male. Most of them were from a postgraduate course in I.T. Their ages range from 20 to 35. Five blind participants were invited from a local organization for blind people. Their ages were between 30 and 71 and they had different education backgrounds. It was hoped that ten blind participants could have been obtained but only five participants volunteered to take part in the event. Such a small number is insufficient to provide any statistically meaningful results but some implications can still be obtained from the experiment results.
Several line graphs were created for this experiment (Figure 2). In common, two lines were displayed on a graph and they were either located separately or cross- intersecting each other. Two different friction properties were applied to the lines and they were classified as sticky and slippery. The lines were made up by cylinders, which are one of the primitive shapes supported by the GHOST programming SDK. Therefore, straight-line approximation was used to construct simple line graphs. All the lines were half sub-merged into a flat surface and thus a semi-circle cross-section was formed on all the line models. Due to the circular cross-section of cylinders, users can feel the jagged corner at the joints of the graph lines. Both sighted and blind people were recruited as participants for this experiment. Ten sighted participants were used and nine were male. Most of them were from a postgraduate course in I.T. Their ages range from 20 to 35. Five blind participants were invited from a local organization for blind people. Their ages were between 30 and 71 and they had different education backgrounds. It was hoped that ten blind participants could have been obtained but only five participants volunteered to take part in the event. Such a small number is insufficient to provide any statistically meaningful results but some implications can still be obtained from the experiment results.
Fig. 3. Correct Distinction of Lines according to Friction in Part 1 (sighted participants: 1-10; blind participants 11-15).
Fig. 3. Correct Distinction of Lines according to Friction in Part 1 (sighted participants: 1-10; blind participants 11-15).
Fig. 4. Participants’ rating on the effectiveness of the friction key (sighted participants: 1-10: blind participants 11-15).
Fig. 4. Participants’ rating on the effectiveness of the friction key (sighted participants: 1-10: blind participants 11-15).
Fig. 5. Participants’ rating on the effectiveness of the gridlines (sighted participants: 1-10; blind participants 11-15).
Fig. 5. Participants’ rating on the effectiveness of the gridlines (sighted participants: 1-10; blind participants 11-15).
Fig. 6. Participants' rating on the usefulness of the toggling gridlines (sighted participants: 1-10; blind participants 11-15). Three main issues were investigated in the experiment. Using friction key was shown to be useful to distinguish different lines on a simple haptic graph but become less effective in a complex graph. Some participants were confused by the cross intersection points between two lines. They thought the graph consisted of two separate lines combining sticky and slippery parts. This can be seen from the sketch drawn by the participants after the graph exploration (Figure 7b). Effect of the friction key was hindered by the embossed cylindrical lines. The line modeling technique, which simply joined cylinder objects together, caused this confusion. Participants found it hard to keep their pointer on the line, especially at the corners and the end points of the lines (Figure 7c). This increased the difficulty of tracing the trend of the lines and instead participants struggled to figure out the shape of the line model. Therefore, this simple line modeling technique cannot provide users with effective graph exploration but obstructs users from perceiving correct information about the graph. Different line modeling techniques which can produce curved lines and smooth joints are required. Gridlines provide an aid to find the values on a haptic graph. However, the ayneriment reailte chnwed that they were not very effective and nat every narticinant
Fig. 6. Participants' rating on the usefulness of the toggling gridlines (sighted participants: 1-10; blind participants 11-15). Three main issues were investigated in the experiment. Using friction key was shown to be useful to distinguish different lines on a simple haptic graph but become less effective in a complex graph. Some participants were confused by the cross intersection points between two lines. They thought the graph consisted of two separate lines combining sticky and slippery parts. This can be seen from the sketch drawn by the participants after the graph exploration (Figure 7b). Effect of the friction key was hindered by the embossed cylindrical lines. The line modeling technique, which simply joined cylinder objects together, caused this confusion. Participants found it hard to keep their pointer on the line, especially at the corners and the end points of the lines (Figure 7c). This increased the difficulty of tracing the trend of the lines and instead participants struggled to figure out the shape of the line model. Therefore, this simple line modeling technique cannot provide users with effective graph exploration but obstructs users from perceiving correct information about the graph. Different line modeling techniques which can produce curved lines and smooth joints are required. Gridlines provide an aid to find the values on a haptic graph. However, the ayneriment reailte chnwed that they were not very effective and nat every narticinant
There are many different parts on a graph so that various surface textures can be applied in order to tell them apart. Since preliminary results have shown that users can distinguish different frictions applied on the lines, mixtures of friction and texture can be used as a distinctive feature of an object. Using surface texture not only can solve the confusion of multiple lines but also gives an indication of different parts of the graph so that users will know where they are on the graph. Investigation is needed here to identify which type of texture can be easily perceived by users via the PHANToM because force feedback devices are generally not good at presenting cutaneous sensation [4].
There are many different parts on a graph so that various surface textures can be applied in order to tell them apart. Since preliminary results have shown that users can distinguish different frictions applied on the lines, mixtures of friction and texture can be used as a distinctive feature of an object. Using surface texture not only can solve the confusion of multiple lines but also gives an indication of different parts of the graph so that users will know where they are on the graph. Investigation is needed here to identify which type of texture can be easily perceived by users via the PHANToM because force feedback devices are generally not good at presenting cutaneous sensation [4].
Fig. 1. Logitech Wingman Force Feedback Mouse Though haptic feedback is a relatively new modality for HCI, its potential fo providing access to GUIs for blind computer users was recognized early on and ha been explored by many researchers. Broadly speaking, these efforts can be divide into two strands, projects which have concentrated on rendering the components c the ‘environment’, the GUI itself [1], [2], [3], and those which have focused o rendering ‘content’, the non-text information that might be produced by application such as mathematical packages etc. [4], [5], [6], [7]. Two further studies hav specifically addressed issues related to haptic rendering of objects and images on th World Wide Web [8], [9]. However, both differ from the present study in that the concentrate on the rendering of web page layout and address only in passing th possibility of rendering content haptically as well. Treen ntiatiinn 4thnt Lancs Jsaemmlinntinnn Frew hnth ntenndNn RE wrnnnnwenh Lbacrn £rnniianr ao
Fig. 1. Logitech Wingman Force Feedback Mouse Though haptic feedback is a relatively new modality for HCI, its potential fo providing access to GUIs for blind computer users was recognized early on and ha been explored by many researchers. Broadly speaking, these efforts can be divide into two strands, projects which have concentrated on rendering the components c the ‘environment’, the GUI itself [1], [2], [3], and those which have focused o rendering ‘content’, the non-text information that might be produced by application such as mathematical packages etc. [4], [5], [6], [7]. Two further studies hav specifically addressed issues related to haptic rendering of objects and images on th World Wide Web [8], [9]. However, both differ from the present study in that the concentrate on the rendering of web page layout and address only in passing th possibility of rendering content haptically as well. Treen ntiatiinn 4thnt Lancs Jsaemmlinntinnn Frew hnth ntenndNn RE wrnnnnwenh Lbacrn £rnniianr ao
Fig. 3. Screen shot of the Analysis Mode of the electric field laboratory.
Fig. 3. Screen shot of the Analysis Mode of the electric field laboratory.
However, we noticed that using this system, users experienced difficulties — they became lost and unable to follow the gesture. We attributed this to the fact that forces of attraction used are relatively weak and become weaker as a user approaches a target area, making it difficult to locate these areas. There were several solutions to this problem. As we had mapped larger forces to greater distances we did not want to simply increase the magnitude of the forces when users became close to a point. Not did we want to increase the size of the range at which a user is said to have reached a point as doing this would reduce the fidelity of the gesture — small perturbations would not be recorded. We also felt that it would be easier for users to detect changes in the direction of a force rather than just its magnitude. bg ee ee ae Se eS ee Ce i, ee, a ae), a
However, we noticed that using this system, users experienced difficulties — they became lost and unable to follow the gesture. We attributed this to the fact that forces of attraction used are relatively weak and become weaker as a user approaches a target area, making it difficult to locate these areas. There were several solutions to this problem. As we had mapped larger forces to greater distances we did not want to simply increase the magnitude of the forces when users became close to a point. Not did we want to increase the size of the range at which a user is said to have reached a point as doing this would reduce the fidelity of the gesture — small perturbations would not be recorded. We also felt that it would be easier for users to detect changes in the direction of a force rather than just its magnitude. bg ee ee ae Se eS ee Ce i, ee, a ae), a
Fig. 1. Subjects are doing tasks using two versions of the PHANToM, on the left a "T" model and on the right an "A" model.
Fig. 1. Subjects are doing tasks using two versions of the PHANToM, on the left a "T" model and on the right an "A" model.
The cubes were modelled to simulate simplified cubes with form, mass, damping and surface friction, but lacked e.g. the ability to rotate. The cubes were identical in dynamic behaviour, form and mass but were of four different colours (green, blue, yellow and orange, two of each) to make them easily distinguishable. The subjects could lift the cubes in two different ways. Either the users collaborated in lifting the cubes by pressing into the cube from opposite sides and lifting upwards simultaneously, or a single user lifted a cube by pressing it against the wall and pushing it upwards. The subjects were represented by spheres in the graphical environment which were distinguishable by colour (one was blue, the other green). In the version without haptic force feedback the PHANToM functioned solely as a 3D mouse, as the user could feel neither the cubes, nor the walls, nor the other user in the environment.
The cubes were modelled to simulate simplified cubes with form, mass, damping and surface friction, but lacked e.g. the ability to rotate. The cubes were identical in dynamic behaviour, form and mass but were of four different colours (green, blue, yellow and orange, two of each) to make them easily distinguishable. The subjects could lift the cubes in two different ways. Either the users collaborated in lifting the cubes by pressing into the cube from opposite sides and lifting upwards simultaneously, or a single user lifted a cube by pressing it against the wall and pushing it upwards. The subjects were represented by spheres in the graphical environment which were distinguishable by colour (one was blue, the other green). In the version without haptic force feedback the PHANToM functioned solely as a 3D mouse, as the user could feel neither the cubes, nor the walls, nor the other user in the environment.
Haptic feedback No haptic feedback
Haptic feedback No haptic feedback
front and on the tail of the body and can sense obstacle contact in six different directions. The ultrasonic sensors are mounted to detect obstacles in front of the robot without touching.
front and on the tail of the body and can sense obstacle contact in six different directions. The ultrasonic sensors are mounted to detect obstacles in front of the robot without touching.
Fig. 3. Decomposition of the intentional force exerted to the arm part (Top View). With the use of bridged strain gauges, the arm part also functions as a force sensor We adopt the Four Active Gage Method for measuring the force/torque. Each set 0: the two bridges (one on the vertical part of the arm and the other on the horizonta part) provides an independent output corresponding to the bend in a particulai direction, that is, either forward/backward or clockwise/counter-clockwise (Fig.3) Independence as well as linearity of the force sensor output is confirmed ir experiment as shown in Fig.4.
Fig. 3. Decomposition of the intentional force exerted to the arm part (Top View). With the use of bridged strain gauges, the arm part also functions as a force sensor We adopt the Four Active Gage Method for measuring the force/torque. Each set 0: the two bridges (one on the vertical part of the arm and the other on the horizonta part) provides an independent output corresponding to the bend in a particulai direction, that is, either forward/backward or clockwise/counter-clockwise (Fig.3) Independence as well as linearity of the force sensor output is confirmed ir experiment as shown in Fig.4.
Fig. 5. The micro-switch sensor on the bottom side of the Hand Part.
Fig. 5. The micro-switch sensor on the bottom side of the Hand Part.
Fig. 10. Fluctuation of distance between the user and the robot in human-following task.
Fig. 10. Fluctuation of distance between the user and the robot in human-following task.
Table 3. Questionnaire answers of the Human-Leading Experiment Comparative Experiment In order to evaluate the efficiency of the interface system, a comparative experiment is also held with the help of the same 10 volunteers. Two other different types of the existing interface devices along with the hand-shaped force interface are used in comparison. In the same experimental environment as shown in Fig.12, this time, the volunteer users are requested to lead the robot from the start to the goal. Two of the existing interface devices are a digital joystick and a remote controller. Each interface
Table 3. Questionnaire answers of the Human-Leading Experiment Comparative Experiment In order to evaluate the efficiency of the interface system, a comparative experiment is also held with the help of the same 10 volunteers. Two other different types of the existing interface devices along with the hand-shaped force interface are used in comparison. In the same experimental environment as shown in Fig.12, this time, the volunteer users are requested to lead the robot from the start to the goal. Two of the existing interface devices are a digital joystick and a remote controller. Each interface
Fig. 12. Environmental map of the Human-Leading Experiment (Top View) Table 2. Goal time of the tasks with and without human follower
Fig. 12. Environmental map of the Human-Leading Experiment (Top View) Table 2. Goal time of the tasks with and without human follower
Table 5. Comparative evaluation in terms of scores from the questionnaires. In this paper, the hand-shaped force interface for human-cooperative mobile robot is proposed. By utilizing hand-to-hand force interaction, profuse communication with intentional force between a human and a robot is achieved. In the human-following task, the robot not only follows the human user to the direction in which the intentional force is exerted, but also recognizes obstacles and communicates that information to the user. In the human-leading task, the robot moves as it is pre- programmed. It stops when the human follower exerts intentional force to the opposite direction of its motion. As for evaluation of the proposed robotic system, we experimented on both tasks in real human-robot cooperation. Efficiency of the system as a human interface is also testified in comparison to other interface systems. The
Table 5. Comparative evaluation in terms of scores from the questionnaires. In this paper, the hand-shaped force interface for human-cooperative mobile robot is proposed. By utilizing hand-to-hand force interaction, profuse communication with intentional force between a human and a robot is achieved. In the human-following task, the robot not only follows the human user to the direction in which the intentional force is exerted, but also recognizes obstacles and communicates that information to the user. In the human-leading task, the robot moves as it is pre- programmed. It stops when the human follower exerts intentional force to the opposite direction of its motion. As for evaluation of the proposed robotic system, we experimented on both tasks in real human-robot cooperation. Efficiency of the system as a human interface is also testified in comparison to other interface systems. The
diffuse and point light sources. The haptical properties were varied to obtain cubes of different weight and different frictional relations with the floor. These manipulations resulted in cubes of different weight that due to friction with the floor were more or less sensitive to displacement forces. The friction between the fingertips and the cube was held constant. For the experiment 3 object weights were used simulating mass (M) of 50, 100 and 200 grams. Four static coefficients of friction (! ,) were used, simulating values of 0.12, 0.30, 0.65 and 1.0. The dynamic friction coefficients (! g) corresponding to these 4 values were 0.12, 0.12, 0.48 and 0.82 respectively. The experiment consisted of a weight series of the 3 weights with the coefficient of friction held constant at 0.65 for static and 0.48 for dynamic, and a friction series of the 4 frictional values with the mass held constant at 68 grams. A summary of these values is presented in Table 1. The visual properties of the object to be grasped were held constant except for the cealnr af the ahiect which wae waried tn cerve ac an indicatar that the cihe tan he
diffuse and point light sources. The haptical properties were varied to obtain cubes of different weight and different frictional relations with the floor. These manipulations resulted in cubes of different weight that due to friction with the floor were more or less sensitive to displacement forces. The friction between the fingertips and the cube was held constant. For the experiment 3 object weights were used simulating mass (M) of 50, 100 and 200 grams. Four static coefficients of friction (! ,) were used, simulating values of 0.12, 0.30, 0.65 and 1.0. The dynamic friction coefficients (! g) corresponding to these 4 values were 0.12, 0.12, 0.48 and 0.82 respectively. The experiment consisted of a weight series of the 3 weights with the coefficient of friction held constant at 0.65 for static and 0.48 for dynamic, and a friction series of the 4 frictional values with the mass held constant at 68 grams. A summary of these values is presented in Table 1. The visual properties of the object to be grasped were held constant except for the cealnr af the ahiect which wae waried tn cerve ac an indicatar that the cihe tan he
Fig. 3. Example of the calculation of contact force. In (a) we see the timecourse of force measurements for both the thumb and the index finger. In (b) we see the difference between these two curves. The corresponding estimate of the contact force would be approximately 0.07 Newtons as found on the peak of the curve shown in (b). For both (a) and (b) time of zero corresponds to contact of the first finger with the object.
Fig. 3. Example of the calculation of contact force. In (a) we see the timecourse of force measurements for both the thumb and the index finger. In (b) we see the difference between these two curves. The corresponding estimate of the contact force would be approximately 0.07 Newtons as found on the peak of the curve shown in (b). For both (a) and (b) time of zero corresponds to contact of the first finger with the object.
Fig. 4. A plot of the measured contact force versus the quantity !. M for each of the 7 experimental conditions. A linear regression was performed on the points and the resulting line is shown on the graph as well as, r’, the percentage of variance accounted for by the linear relationship. The predicted relation is that contact force should increase for experimental conditions with higher values of ! , M.
Fig. 4. A plot of the measured contact force versus the quantity !. M for each of the 7 experimental conditions. A linear regression was performed on the points and the resulting line is shown on the graph as well as, r’, the percentage of variance accounted for by the linear relationship. The predicted relation is that contact force should increase for experimental conditions with higher values of ! , M.
Fig. 5. The four different kinematic markers of a) maximum aperture, b) maximum velocity, c) maximum acceleration and d) maximum deceleration are plotted versus the measured contact forces. Each graph includes the resultant linear regression line and the corresponding value of 2 Tr.
Fig. 5. The four different kinematic markers of a) maximum aperture, b) maximum velocity, c) maximum acceleration and d) maximum deceleration are plotted versus the measured contact forces. Each graph includes the resultant linear regression line and the corresponding value of 2 Tr.
Fig. 1. Sample visual environment. Top grating is standard, spatial period = 2mm. Lowe grating is test, with spatial period = 2.5mm.
Fig. 1. Sample visual environment. Top grating is standard, spatial period = 2mm. Lowe grating is test, with spatial period = 2.5mm.
It could perhaps be argued that texture is more suitable to production by tactile devices. Despite the early perceptual and physiological arguments for a spatial code to texture, three-dimensional force feedback interfaces are able to simulate surface texture (Weisenberger and Krier, 1997). It is the degree of fidelity and realism achievable with such devices that is of primary interest. The interaction issue then is how to overcome any limitations of using force feedback devices alone to represent texture. Table 1: Definitions (Oakley, McGee, Brewster and Gray, CHI 2000).
It could perhaps be argued that texture is more suitable to production by tactile devices. Despite the early perceptual and physiological arguments for a spatial code to texture, three-dimensional force feedback interfaces are able to simulate surface texture (Weisenberger and Krier, 1997). It is the degree of fidelity and realism achievable with such devices that is of primary interest. The interaction issue then is how to overcome any limitations of using force feedback devices alone to represent texture. Table 1: Definitions (Oakley, McGee, Brewster and Gray, CHI 2000).
Fig 1: The Phantom 3D force feedback device from SensAble Technologies. The PHANToM 1.0 force feedback device by SensAble Technologies will be used to create the haptic virtual surfaces (see Fig. 1). Force feedback devices have optical sensors that detect changes in the device's configuration. The device then uses mechanical actuators to apply forces back to the user calculated from the positional information and the stored algorithmic models of the objects with which the user is interacting. The interaction relies on kinesthetic information being conveyed to the user rather than cutaneous information (see table 1).
Fig 1: The Phantom 3D force feedback device from SensAble Technologies. The PHANToM 1.0 force feedback device by SensAble Technologies will be used to create the haptic virtual surfaces (see Fig. 1). Force feedback devices have optical sensors that detect changes in the device's configuration. The device then uses mechanical actuators to apply forces back to the user calculated from the positional information and the stored algorithmic models of the objects with which the user is interacting. The interaction relies on kinesthetic information being conveyed to the user rather than cutaneous information (see table 1).
Fig. 2: Diagrammatic view of interface. Hypotheses and Implications
Fig. 2: Diagrammatic view of interface. Hypotheses and Implications
The experiment was divided in three sessions, each with one of three control devices; a normal mouse, an optical mouse or an optical trackball. The order in which the devices were used was random. Each session consisted of a practice and an experimental part. In the experimental part each target was randomly presented ten times.
The experiment was divided in three sessions, each with one of three control devices; a normal mouse, an optical mouse or an optical trackball. The order in which the devices were used was random. Each session consisted of a practice and an experimental part. In the experimental part each target was randomly presented ten times.
Although there is a statistically significant curvature (as opposed to straight- ness) in some of the measured paths, the mean paths appear rather straight in comparison with standard deviations. For target estimation purposes only negligible losses are to be expected if paths are assumed to be straight lines. However, there is some reason for discussing the results on curvature. Curvature in movements to the left and right is imaginable as occurring around the wrist angle. However this angle might (a) be too small to be visible or (b) be easily compensated by the fingers.
Although there is a statistically significant curvature (as opposed to straight- ness) in some of the measured paths, the mean paths appear rather straight in comparison with standard deviations. For target estimation purposes only negligible losses are to be expected if paths are assumed to be straight lines. However, there is some reason for discussing the results on curvature. Curvature in movements to the left and right is imaginable as occurring around the wrist angle. However this angle might (a) be too small to be visible or (b) be easily compensated by the fingers.
Fig. 4. Mean path and st.dev. for three devices a)mechanical mouse, b)optica mouse and c)optical trackball in eight directions.
Fig. 4. Mean path and st.dev. for three devices a)mechanical mouse, b)optica mouse and c)optical trackball in eight directions.
Fig. 1-a. Real hand in direct interaction with Virtual object. Fig. 1-b. Indirect interaction between real hand and virtual object. It has been proven in many researches that stereoscopic images are very effective in improving interactions in virtual environment. As well, haptic sensations are known to impart users with realistic feeling about physical interactions, which improve the control over virtual objects. However, in most virtual reality systems, when these two modalities are coupled together, the real world (users) and the virtual world are separated, and do not interact directly with each other. For example, when the task of the user is to grasp a virtual object, usually we provide an imaginary graphic hand to interact directly with the virtual object (small sphere in figure 1-b), whereas the real hand is used to control the interaction without co-existence with the object, see figure 1-b. Such indirect control makes the user think in terms of manipulating objects in remote site, and do not have the conviction that he is within the virtual environment itself. Although, representing body’s parts involved in the virtual environment with similar graphics has sufficient accuracy for interactions (Butts and McAllister 1988, Spain 1984, Beaton 1987 and Reinhardt 1990), yet, it keeps the real and the virtual worlds non fused, and felt as being two different environments. What we want to achieve is direct control over the virtual object as seen in figure |-a.
Fig. 1-a. Real hand in direct interaction with Virtual object. Fig. 1-b. Indirect interaction between real hand and virtual object. It has been proven in many researches that stereoscopic images are very effective in improving interactions in virtual environment. As well, haptic sensations are known to impart users with realistic feeling about physical interactions, which improve the control over virtual objects. However, in most virtual reality systems, when these two modalities are coupled together, the real world (users) and the virtual world are separated, and do not interact directly with each other. For example, when the task of the user is to grasp a virtual object, usually we provide an imaginary graphic hand to interact directly with the virtual object (small sphere in figure 1-b), whereas the real hand is used to control the interaction without co-existence with the object, see figure 1-b. Such indirect control makes the user think in terms of manipulating objects in remote site, and do not have the conviction that he is within the virtual environment itself. Although, representing body’s parts involved in the virtual environment with similar graphics has sufficient accuracy for interactions (Butts and McAllister 1988, Spain 1984, Beaton 1987 and Reinhardt 1990), yet, it keeps the real and the virtual worlds non fused, and felt as being two different environments. What we want to achieve is direct control over the virtual object as seen in figure |-a.
Fig. 2. Difference between stereographic and real object visual display. The first issue is of particular interest in the development of mixed and augmented reality systems. The problem of misperceiving the location of objects at different depths is especially important if one of the principal tasks of the user is to reach and grasp an object, or to align objects. The force feedback sensation has to be always displayed in accordance with the virtual environment; any mismatching between the two modalities may cause the failure of the tasks. The present paper presents the results of our investigation about this issue.
Fig. 2. Difference between stereographic and real object visual display. The first issue is of particular interest in the development of mixed and augmented reality systems. The problem of misperceiving the location of objects at different depths is especially important if one of the principal tasks of the user is to reach and grasp an object, or to align objects. The force feedback sensation has to be always displayed in accordance with the virtual environment; any mismatching between the two modalities may cause the failure of the tasks. The present paper presents the results of our investigation about this issue.
stereogram is displayed. The observer was seated on a chair inside the device facing the screen. To provide constant viewing distance and avoid cues known to affect depth perception such as perspective and motion parallax, the observer’s head was stabilized with a chinrest at predetermined distance of 70 cm from the screen. All experiments were carried out in a completely dark room to discard any aiming point or background information that the observer may use as depth reference. The random- dot stereograms were made up of a square matrix consisting of 230x230 square dots, each of which had an equal probability of being displayed or not. The square has no background, and all dots were displayed in only red color. The square displayed to the left eye had a range of disparities added to it by shifting its horizontal position. Eight crossed disparities were employed in this experiment. The disparity ranges from 20 dots to 160 dots, and each disparity was in integer multiple of 20 dots. When the observer fuse left and right image he always perceive the square being in front of the screen and at hands reach. Figure 3 gives more detail about the apparatus of the system.
stereogram is displayed. The observer was seated on a chair inside the device facing the screen. To provide constant viewing distance and avoid cues known to affect depth perception such as perspective and motion parallax, the observer’s head was stabilized with a chinrest at predetermined distance of 70 cm from the screen. All experiments were carried out in a completely dark room to discard any aiming point or background information that the observer may use as depth reference. The random- dot stereograms were made up of a square matrix consisting of 230x230 square dots, each of which had an equal probability of being displayed or not. The square has no background, and all dots were displayed in only red color. The square displayed to the left eye had a range of disparities added to it by shifting its horizontal position. Eight crossed disparities were employed in this experiment. The disparity ranges from 20 dots to 160 dots, and each disparity was in integer multiple of 20 dots. When the observer fuse left and right image he always perceive the square being in front of the screen and at hands reach. Figure 3 gives more detail about the apparatus of the system.
Results The results are summarized in figure 5. The graph shows the means and deviations of perceived depth as a function of disparity. Depth is expressed in terms of the distance interval between the observer and the perceived square. The small rectangles in the graph represent the mean of depth for each disparity. The vertical line represents the depth variation related to each disparity. The standard errors across disparities ranged from 1.5 cm to 5 cm of the mean depth and averaged 2.3 cm. Nearly all subjects responded quickly and confidently to all trial, usually viewing them for about 2 to 4 seconds before making a response. The results from the naive observers did not differ in any systematic way from those from the experienced observer. was to move his right hand forward until it becomes aligned with the right side of the square. Observers were not able to see their hands, because the room was dark and also instructed previously to not occlude the image by heir hands, figure 4. When the observer subjectively believes that his hand is at the same depth as the square he report this judgment vocally by saying “HERE”. As the hand position is tracked in real time via the SCALEABLE SPIDAR device, the experimenter on clicking a mouse button recorded it instantly. The same task was repeated randomly at least 4 times for each disparity (8 disparities in all). After each trial the observer was asked to close his eyes while the experimenter change the disparity by clicking another mouse button, this pause is about 2 to 3 seconds. TO ie a a a I aN ee a a Ee can
Results The results are summarized in figure 5. The graph shows the means and deviations of perceived depth as a function of disparity. Depth is expressed in terms of the distance interval between the observer and the perceived square. The small rectangles in the graph represent the mean of depth for each disparity. The vertical line represents the depth variation related to each disparity. The standard errors across disparities ranged from 1.5 cm to 5 cm of the mean depth and averaged 2.3 cm. Nearly all subjects responded quickly and confidently to all trial, usually viewing them for about 2 to 4 seconds before making a response. The results from the naive observers did not differ in any systematic way from those from the experienced observer. was to move his right hand forward until it becomes aligned with the right side of the square. Observers were not able to see their hands, because the room was dark and also instructed previously to not occlude the image by heir hands, figure 4. When the observer subjectively believes that his hand is at the same depth as the square he report this judgment vocally by saying “HERE”. As the hand position is tracked in real time via the SCALEABLE SPIDAR device, the experimenter on clicking a mouse button recorded it instantly. The same task was repeated randomly at least 4 times for each disparity (8 disparities in all). After each trial the observer was asked to close his eyes while the experimenter change the disparity by clicking another mouse button, this pause is about 2 to 3 seconds. TO ie a a a I aN ee a a Ee can
Fig. 5. Perceived depth for the different disparities.
Fig. 5. Perceived depth for the different disparities.
Fig. 8. Sensitivity to depth variation. change of disparity from 20 dots to 40 dots generate about 9 cm of depth variation, whereas the same change of 20 dots from 140 to 160 dots gives only 3 cm of depth variation. Ogle [8], Sperling 1983 and others researchers classified this phenomenon into two kinds of stereopsis, “patent” and “qualitative”. The first represent accurate stereopsis and occurs when disparity is small. The later represent imprecise stereopsis and occurs when disparity is big. This observation is validated by equation (3), which is the derivative function of equation (2). When the disparity is small the slop is big and gradually flattened while disparity is increasing, see figure 8. An important question arises immediately in regard to all these data and
Fig. 8. Sensitivity to depth variation. change of disparity from 20 dots to 40 dots generate about 9 cm of depth variation, whereas the same change of 20 dots from 140 to 160 dots gives only 3 cm of depth variation. Ogle [8], Sperling 1983 and others researchers classified this phenomenon into two kinds of stereopsis, “patent” and “qualitative”. The first represent accurate stereopsis and occurs when disparity is small. The later represent imprecise stereopsis and occurs when disparity is big. This observation is validated by equation (3), which is the derivative function of equation (2). When the disparity is small the slop is big and gradually flattened while disparity is increasing, see figure 8. An important question arises immediately in regard to all these data and
Fig. 10. Small squares represent the average time needed to perceive depth while using only stereopsis. The dark dots represent the perceived depth when both visual and haptic cues are provided.
Fig. 10. Small squares represent the average time needed to perceive depth while using only stereopsis. The dark dots represent the perceived depth when both visual and haptic cues are provided.
Figure 12-a: Resultant force of tension. must satisfy the following equations (Eqs).
Figure 12-a: Resultant force of tension. must satisfy the following equations (Eqs).
Figure 12-b: Motor and rotary encoder.
Figure 12-b: Motor and rotary encoder.
Figure 14-a: The fingering. Figure 14-b: Apparatus of the Scaleable- SPIDAR.
Figure 14-a: The fingering. Figure 14-b: Apparatus of the Scaleable- SPIDAR.
Shape recognition is potentially based upon both abstract knowledge of the shap (in the case of our shapes, the directions of curvature for the two parabolas) and han movements specific to an exact configuration. The balance between these two wi depend upon the frequency of contact with a specific object. For example, you migh identify an arbitrary cup shape using one sequence of hand movements but identif the coffee cup you drink from every day using movements specific to its geometry This presented us with a dilemma: We needed to teach our participants to hapticall recognize the five shapes but did not want them to use recognition methods that relie: upon specific geometries of a given stimulus set. We resolved this by training ou participants with one set of stimuli and testing them with another. The training phas used medium-sized shapes presented in “head on” configuration whereas the testin phase used small- and large-sized shapes presented in different rotations. T]eino these stimuli we develoned a henchmark task summarized in the ney
Shape recognition is potentially based upon both abstract knowledge of the shap (in the case of our shapes, the directions of curvature for the two parabolas) and han movements specific to an exact configuration. The balance between these two wi depend upon the frequency of contact with a specific object. For example, you migh identify an arbitrary cup shape using one sequence of hand movements but identif the coffee cup you drink from every day using movements specific to its geometry This presented us with a dilemma: We needed to teach our participants to hapticall recognize the five shapes but did not want them to use recognition methods that relie: upon specific geometries of a given stimulus set. We resolved this by training ou participants with one set of stimuli and testing them with another. The training phas used medium-sized shapes presented in “head on” configuration whereas the testin phase used small- and large-sized shapes presented in different rotations. T]eino these stimuli we develoned a henchmark task summarized in the ney
University Veterinary School. An initial experiment showed no significant difference in performance between students trained by traditional methods and those trained on the simulator. However, the results also showed there was potential for improvement in traditional training methods of ovary palpation [6], as an equally low percentage of ovaries were diagnosed correctly for both training groups. —_ Guiding a user using pre-recorded movements can be broken down into 2 stages - record and playback. During the record stage, both positional and force information of a user must be stored at specific regular sample intervals. Playback of the procedure would drive the user's interaction point along the path recorded. At each stage, the driving force would equal the force recorded for the current position. By this method, a student could feel the techniques and forces involved in a correct procedure by playing back a recording of an experienced doctor or veterinarian performing the same procedure. Alternatively, a doctor could assess the performance of a student by playing back a recording of their movements. 5 ee ee SS See f nee: Oy a 2 ey ce Perr: OMe | i ie. ee Eh OR
University Veterinary School. An initial experiment showed no significant difference in performance between students trained by traditional methods and those trained on the simulator. However, the results also showed there was potential for improvement in traditional training methods of ovary palpation [6], as an equally low percentage of ovaries were diagnosed correctly for both training groups. —_ Guiding a user using pre-recorded movements can be broken down into 2 stages - record and playback. During the record stage, both positional and force information of a user must be stored at specific regular sample intervals. Playback of the procedure would drive the user's interaction point along the path recorded. At each stage, the driving force would equal the force recorded for the current position. By this method, a student could feel the techniques and forces involved in a correct procedure by playing back a recording of an experienced doctor or veterinarian performing the same procedure. Alternatively, a doctor could assess the performance of a student by playing back a recording of their movements. 5 ee ee SS See f nee: Oy a 2 ey ce Perr: OMe | i ie. ee Eh OR
tutor to have a more active role in guiding the student. An initial attempt has been made to integrate pre-recorded haptic guidance into the training environment. During the recording stage, the position of the PHANToM can be sampled at a rate of between 100 and 1000 Hz. The PHANToM's position sensors provide a representation of the current cursor position that can be used to accurately recreate the path recorded. However, recording force information at the sample points presents problems, as the PHANToM device does not have force sensors. The system implemented attempts to estimate the applied force through the reaction force from objects or effects within the scene. By introducing viscosity throughout the scene, a reaction force to any movement can be detected through the device. Playback also presents problems, as even a passive user can affect the path of the cursor, and applying the recorded force vector will not generally move the cursor along the recorded path. The PHANToM drags the user’s finger through a series of movements. Resistive forces from the user will combine with the driving force to produce deviations from the path. The current system calculates the magnitude of the recorded force and applies it towards the next sample point on the path. This however can cause instabilities when contacting objects in the scene, demonstrated in Figure 2. When touching an object the user will apply a force to counteract the reaction force from the object. Even if the user is moving perpendicular to the surface of the object, he/she is still applying a force to counter this reaction force. However, the playback force will not take account of this and will drive the user directly towards the next sample point. The reaction force will combine with the playback force to produce a net force vector that is not in the direction of the next sample point. A more complex algorithm must be developed to adapt the playback force direction depending on the reartinn farce fram the cantacted nhiect
tutor to have a more active role in guiding the student. An initial attempt has been made to integrate pre-recorded haptic guidance into the training environment. During the recording stage, the position of the PHANToM can be sampled at a rate of between 100 and 1000 Hz. The PHANToM's position sensors provide a representation of the current cursor position that can be used to accurately recreate the path recorded. However, recording force information at the sample points presents problems, as the PHANToM device does not have force sensors. The system implemented attempts to estimate the applied force through the reaction force from objects or effects within the scene. By introducing viscosity throughout the scene, a reaction force to any movement can be detected through the device. Playback also presents problems, as even a passive user can affect the path of the cursor, and applying the recorded force vector will not generally move the cursor along the recorded path. The PHANToM drags the user’s finger through a series of movements. Resistive forces from the user will combine with the driving force to produce deviations from the path. The current system calculates the magnitude of the recorded force and applies it towards the next sample point on the path. This however can cause instabilities when contacting objects in the scene, demonstrated in Figure 2. When touching an object the user will apply a force to counteract the reaction force from the object. Even if the user is moving perpendicular to the surface of the object, he/she is still applying a force to counter this reaction force. However, the playback force will not take account of this and will drive the user directly towards the next sample point. The reaction force will combine with the playback force to produce a net force vector that is not in the direction of the next sample point. A more complex algorithm must be developed to adapt the playback force direction depending on the reartinn farce fram the cantacted nhiect
Fig. 1. Placement of the tactile actuators on the back for the spatial sensitivity experiments (scale is cm). A potential interesting body locus for a tactile navigation display is the torso because of its large surface, its 3D form, and its possible ego-centric ‘view’. Furthermore, information presented to the torso is not likely to interfere with tactile information presentation to, for example, the hands. A simple tactile display could consist of a number of actuators located in a horizontal plane. By stimulating a certain area, the display could indicate a direction, e.g., to a point or object of interest.
Fig. 1. Placement of the tactile actuators on the back for the spatial sensitivity experiments (scale is cm). A potential interesting body locus for a tactile navigation display is the torso because of its large surface, its 3D form, and its possible ego-centric ‘view’. Furthermore, information presented to the torso is not likely to interfere with tactile information presentation to, for example, the hands. A simple tactile display could consist of a number of actuators located in a horizontal plane. By stimulating a certain area, the display could indicate a direction, e.g., to a point or object of interest.
Fig. 3. Top view of the set-up for the direction discrimination task. With a dial, the observer can position a cursor (a spot of light projected from above) along a white circle drawn on the table. The cursor should be positioned such that it indicates the direction associated with the tactile stimulus.
Fig. 3. Top view of the set-up for the direction discrimination task. With a dial, the observer can position a cursor (a spot of light projected from above) along a white circle drawn on the table. The cursor should be positioned such that it indicates the direction associated with the tactile stimulus.
Fig. 5. The Internal Reference Points for the ten observers in the tactile direction determinatiot task. The numbers indicate the individual observers. The results of this experiment were interesting in several ways. First of all, none of the participants had any trouble with the task. This is noteworthy since a point stimulus does not contain any explicit direction information. The strategy people use is probably similar to the one known in visual perception, namely using a perceptual ego-—centre as a second point. Several authors determined the visual ego—centre (e.g., [11]), that can be defined as the position in space at which a person experiences him or herself to be. Identifying an ego-centre or internal reference point is important, because it correlates physical space and phenomenal space. A second reason to determine the internal reference point in this tactile experiment was the striking bias that all participants showed in their responses, namely a bias towards the sagittal plane, see Figure 5 for an example. This means that stimuli on the frontal side of the torso were perceived as directions coming from a point closer to the navel, and stimuli on the dorsal side of the torso were perceived as coming from a point closer to the spine. Further research [15] showed that this bias was not caused by the experimental set-up, the visual system, the subjective location of the stimuli, or other anomalies.
Fig. 5. The Internal Reference Points for the ten observers in the tactile direction determinatiot task. The numbers indicate the individual observers. The results of this experiment were interesting in several ways. First of all, none of the participants had any trouble with the task. This is noteworthy since a point stimulus does not contain any explicit direction information. The strategy people use is probably similar to the one known in visual perception, namely using a perceptual ego-—centre as a second point. Several authors determined the visual ego—centre (e.g., [11]), that can be defined as the position in space at which a person experiences him or herself to be. Identifying an ego-centre or internal reference point is important, because it correlates physical space and phenomenal space. A second reason to determine the internal reference point in this tactile experiment was the striking bias that all participants showed in their responses, namely a bias towards the sagittal plane, see Figure 5 for an example. This means that stimuli on the frontal side of the torso were perceived as directions coming from a point closer to the navel, and stimuli on the dorsal side of the torso were perceived as coming from a point closer to the spine. Further research [15] showed that this bias was not caused by the experimental set-up, the visual system, the subjective location of the stimuli, or other anomalies.
Fig. 1. Geometric information: Directions in 3D space. It has been shown that a pointlike vibro- tactile stimulation on the torso is easily understood by people as an external direction in space. This can be used to effectively indicate directions in 3D space to pilots, such as those associated with other aircraft. It is hypothesized that several of such directions can be distinguished simultaneously.
Fig. 1. Geometric information: Directions in 3D space. It has been shown that a pointlike vibro- tactile stimulation on the torso is easily understood by people as an external direction in space. This can be used to effectively indicate directions in 3D space to pilots, such as those associated with other aircraft. It is hypothesized that several of such directions can be distinguished simultaneously.
These contemplations can be used to derive a human factors research agenda for investigating the advantages and disadvantages of using tactile versus visual and auditory information presentation in the cockpit. For instance, additional studies on multi-modal attention and processing capacity need be performed before it can be confirmed that tactile stimulation can be used to overcome current processing overload problems. The second half of this paper discusses an experiment along one of the other lines, namely vibro-tactile perception under high G-load.
These contemplations can be used to derive a human factors research agenda for investigating the advantages and disadvantages of using tactile versus visual and auditory information presentation in the cockpit. For instance, additional studies on multi-modal attention and processing capacity need be performed before it can be confirmed that tactile stimulation can be used to overcome current processing overload problems. The second half of this paper discusses an experiment along one of the other lines, namely vibro-tactile perception under high G-load.
Table 1. Five different G-profiles used during the experiment.
Table 1. Five different G-profiles used during the experiment.
Fig. 3. Reaction time as a function of the G-load for different profiles.
Fig. 3. Reaction time as a function of the G-load for different profiles.
Fig. 4. Proportion correct responses as a function of the G-load for different G-profiles.
Fig. 4. Proportion correct responses as a function of the G-load for different G-profiles.
and audio, force information is very difficult to send through air. To form a 3D force at a certain point, say point A, lead a “hard” mechanical device from a “force source” to point A may be the only “simple” and precise way. If A is moveable, then the force display device will become much more complicated in structure compared with video and audio display. Particularly, when the virtual environment workspace becomes larger, that is the point A may go far away from the force source, the haptic device structural streneth needs to be enhanced to keen the precision. This enhancement usually makes the whole system bulky, heavy and expensive, as well limits the user’s moving freedom. On the other hand, the machinery based forces displays are usually low dynamic performance. In a mechanical system, the dynamic per formance is mainly decided by system’s weight and moment of inertia. As the haptic devices in human-scale virtual environment are heavy, they would have lower dynamic performance than the ones in a relatively small system, desktop devices. Un fortunately, the task in large working space tends to need higher moving speed and bigger acceleration. How to balance precision and dynamic performance? While improving both of them are the key points to realize usable and accurate force display in human-scale virtual environment.
and audio, force information is very difficult to send through air. To form a 3D force at a certain point, say point A, lead a “hard” mechanical device from a “force source” to point A may be the only “simple” and precise way. If A is moveable, then the force display device will become much more complicated in structure compared with video and audio display. Particularly, when the virtual environment workspace becomes larger, that is the point A may go far away from the force source, the haptic device structural streneth needs to be enhanced to keen the precision. This enhancement usually makes the whole system bulky, heavy and expensive, as well limits the user’s moving freedom. On the other hand, the machinery based forces displays are usually low dynamic performance. In a mechanical system, the dynamic per formance is mainly decided by system’s weight and moment of inertia. As the haptic devices in human-scale virtual environment are heavy, they would have lower dynamic performance than the ones in a relatively small system, desktop devices. Un fortunately, the task in large working space tends to need higher moving speed and bigger acceleration. How to balance precision and dynamic performance? While improving both of them are the key points to realize usable and accurate force display in human-scale virtual environment.
mr t d The device is derived from the original desktop SPIDAR device, which was introduced late in 1990 by Makoto Sato et al (Ishii 1994). As shown in figure 2, Scaleable-SPIDAR is delimited by a cubic frame that enclose a cave-like space, where he operator can move around to perform large scale movements. The experimental prototype is 27m’ size (3m x 3m x 3m). Within this space, different aspect of force feedback sensations associated mainly with weight, contact and inertia can be isplayed to the operator’s hands by means of tensioned strings. The front side of the d evice holds a large screen, where a computer-generated virtual world is projected. Providing such a combination of haptic and visual feedback cues is indispensable to lets the operator’s eyes and hands work in concert to explore and manipulate objects populating the virtual environment. The device uses tensioned string techniques to track hands position as well as to provide haptic feedback sensations. The approach consists mainly on applying appropriate tensions to the four strings supporting each fingering worn by the operator. The force feedback felt on the operator’s hand is the same as the resultant force of tension from strings at the center of the fingering; next subsection gives more detail about forces and position computation. In order to control the tension and length of each string, one extremity is connected to the fingering and the other end is wounded around a pulley, which is driven by a DC motor. By controlling the power applied to the motor, the system can create appropriate tension all the time. A rotary
mr t d The device is derived from the original desktop SPIDAR device, which was introduced late in 1990 by Makoto Sato et al (Ishii 1994). As shown in figure 2, Scaleable-SPIDAR is delimited by a cubic frame that enclose a cave-like space, where he operator can move around to perform large scale movements. The experimental prototype is 27m’ size (3m x 3m x 3m). Within this space, different aspect of force feedback sensations associated mainly with weight, contact and inertia can be isplayed to the operator’s hands by means of tensioned strings. The front side of the d evice holds a large screen, where a computer-generated virtual world is projected. Providing such a combination of haptic and visual feedback cues is indispensable to lets the operator’s eyes and hands work in concert to explore and manipulate objects populating the virtual environment. The device uses tensioned string techniques to track hands position as well as to provide haptic feedback sensations. The approach consists mainly on applying appropriate tensions to the four strings supporting each fingering worn by the operator. The force feedback felt on the operator’s hand is the same as the resultant force of tension from strings at the center of the fingering; next subsection gives more detail about forces and position computation. In order to control the tension and length of each string, one extremity is connected to the fingering and the other end is wounded around a pulley, which is driven by a DC motor. By controlling the power applied to the motor, the system can create appropriate tension all the time. A rotary
Fig. 3. Resultant force of tension. Fig. 4. Motor and rotary encoder. Where a; represents the tension value of each string. By controlling all of the a; the resultant force of any magnitude in any direction can be composed.
Fig. 3. Resultant force of tension. Fig. 4. Motor and rotary encoder. Where a; represents the tension value of each string. By controlling all of the a; the resultant force of any magnitude in any direction can be composed.
After differences between the respective adjacent two equations among equation (1)-(4) and solve the simultaneous equations, we can obtain the position of a fingering (hand) as the following equation (5):
After differences between the respective adjacent two equations among equation (1)-(4) and solve the simultaneous equations, we can obtain the position of a fingering (hand) as the following equation (5):
is used to control an 8-bits D/A, A/D converter and a VME bus, which control respectively the currents entering the motors and detect the changes occurred on each rotary encoder. The PC is connected to a graphics workstation that provides a real- time video image of the virtual world. Figure 7 shows the apparatus of the prototype. Fig. 6. The fingering pointer.
is used to control an 8-bits D/A, A/D converter and a VME bus, which control respectively the currents entering the motors and detect the changes occurred on each rotary encoder. The PC is connected to a graphics workstation that provides a real- time video image of the virtual world. Figure 7 shows the apparatus of the prototype. Fig. 6. The fingering pointer.
Fig. 7. Apparatus of Scaleable-SPIDAR.
Fig. 7. Apparatus of Scaleable-SPIDAR.
Table 1. Comparision with other haptic devices. Experiments and Application In this section, the implementation of a haptic feedback experience with Scaleable- SPIDAR and an evaluation application are described.
Table 1. Comparision with other haptic devices. Experiments and Application In this section, the implementation of a haptic feedback experience with Scaleable- SPIDAR and an evaluation application are described.
Condition 3: Force Feedback Cues Only: after the operator has remembered yuttons’ positions in his mind, the hand’s visual feedback cues are disabled; thus yperator can not “see” the position of his hand in the simulated scene. That is he 10t know whether his hand is moving close to the button or not, but only “feel” fo eedback reactions when the hand runs over the virtual wall or the buttons. ‘our right-handed subjects participated in the experiment, including two of suthors. None of them reported any haptic deficiencies. Although it was not necess or the experiment, all subjects were familiar with haptic devices and virt vironment. Each subject was told about the three different conditions and the task ye performed. There was three different sessions of trials for all subjects. In e: session the red button should be pushed successfully 40 times. Before any sessio short time of practice was given.
Condition 3: Force Feedback Cues Only: after the operator has remembered yuttons’ positions in his mind, the hand’s visual feedback cues are disabled; thus yperator can not “see” the position of his hand in the simulated scene. That is he 10t know whether his hand is moving close to the button or not, but only “feel” fo eedback reactions when the hand runs over the virtual wall or the buttons. ‘our right-handed subjects participated in the experiment, including two of suthors. None of them reported any haptic deficiencies. Although it was not necess or the experiment, all subjects were familiar with haptic devices and virt vironment. Each subject was told about the three different conditions and the task ye performed. There was three different sessions of trials for all subjects. In e: session the red button should be pushed successfully 40 times. Before any sessio short time of practice was given.
Fig. 9. Mean task time of the different conditions. It was found also, that in condition 1 80% of the TTs is devoted exclusively to push he button, whereas only 20% of the TTs is needed for positioning the hand in front of he red button. Nearly the opposite situation is present for trials done under condition 3, where 30% of the TTs is devoted exclusively for the pushing task and the other 70% of times are used to localize the targeted button. Also, as it can be seen from figure 9 that, the TTs variance are smaller when force feedback is available. The difference of time spending in both condition 1 and 3 is significant. In the former one subiects have mainlv a lack of denth perception. but good navigational
Fig. 9. Mean task time of the different conditions. It was found also, that in condition 1 80% of the TTs is devoted exclusively to push he button, whereas only 20% of the TTs is needed for positioning the hand in front of he red button. Nearly the opposite situation is present for trials done under condition 3, where 30% of the TTs is devoted exclusively for the pushing task and the other 70% of times are used to localize the targeted button. Also, as it can be seen from figure 9 that, the TTs variance are smaller when force feedback is available. The difference of time spending in both condition 1 and 3 is significant. In the former one subiects have mainlv a lack of denth perception. but good navigational
In order to control the ball and perceive haptically its has to wear the two fingering on both hands. As the p virtual ball appropriate forces are displayed, such as its held tight enough or open her/his hands, the ball will hand's velocity and orientation while freeing the ball. If t weig ayer frame the system tracks the hands’ position, and when they come into contact with th weig he ba hoop, it may rebound from the backboard, basket's ring or objects surrounding th fall down and bounds on th floor. After making a shot, the ball begins a free falling move ht and shape, the playe start moving inside th ) ht. If the player doesn ment determined by th 1 doesn't go through th playground. The virtual ball is designed 40cm in diameter and weights 300g. To show the force feedback effectiveness in such ski Iful operation, we asked tw
In order to control the ball and perceive haptically its has to wear the two fingering on both hands. As the p virtual ball appropriate forces are displayed, such as its held tight enough or open her/his hands, the ball will hand's velocity and orientation while freeing the ball. If t weig ayer frame the system tracks the hands’ position, and when they come into contact with th weig he ba hoop, it may rebound from the backboard, basket's ring or objects surrounding th fall down and bounds on th floor. After making a shot, the ball begins a free falling move ht and shape, the playe start moving inside th ) ht. If the player doesn ment determined by th 1 doesn't go through th playground. The virtual ball is designed 40cm in diameter and weights 300g. To show the force feedback effectiveness in such ski Iful operation, we asked tw
Fig. 11. Distance between the two hands with force feedback.
Fig. 11. Distance between the two hands with force feedback.
of being ‘with’ an object. Considering the other components of immersion as suggested by Slater and Wilbur [11]. The quality of a display (vivid), the range of sensory modalities (extensive) and the correspondence between the user’s actions and displayed information (matching) are all aspects of how naturally a display supports a user. These components are not unique requirements for immersive displays. Indeed the only other factor unique to immersion apart from the ability to surround a user, is the extent to which a user is removed from reality (inclusion). Thus the difference in presence between immersive and non-immersive displays results from a display surrounding and isolating a user. However, some tasks do not require the user to be surrounded or isolated. Presence forms an important subjective measure of a user’s virtual experience although it is only useful in relation to performance [3]. It is assumed that the mor natural the display feels, the greater its usefulness [9]. This naturalness may bette enable a user to utilize ‘real-world’ skills in a virtual environment although it ma also help to transfer learning from the virtual environment back into the real-worl [9]. The conventional definition of presence suggests that non-immersive displays ar inadequate, even for tasks that do not require the user to be surrounded and isolated Therefore, a new measure is needed to assess presence for non-immersive display that will more closely consider task requirements and how naturally a display support auser. “/t is here” is the idea that a display medium brings an object or person to th user [9]. This idea has only been investigated for conventional television programmes where it assesses the belief that the actual object being displayed exists within th television set. However, this concept can be extended to provide a measure for non immersive displays where the object appears to be in the user’s physical environment instead of inside the display.
of being ‘with’ an object. Considering the other components of immersion as suggested by Slater and Wilbur [11]. The quality of a display (vivid), the range of sensory modalities (extensive) and the correspondence between the user’s actions and displayed information (matching) are all aspects of how naturally a display supports a user. These components are not unique requirements for immersive displays. Indeed the only other factor unique to immersion apart from the ability to surround a user, is the extent to which a user is removed from reality (inclusion). Thus the difference in presence between immersive and non-immersive displays results from a display surrounding and isolating a user. However, some tasks do not require the user to be surrounded or isolated. Presence forms an important subjective measure of a user’s virtual experience although it is only useful in relation to performance [3]. It is assumed that the mor natural the display feels, the greater its usefulness [9]. This naturalness may bette enable a user to utilize ‘real-world’ skills in a virtual environment although it ma also help to transfer learning from the virtual environment back into the real-worl [9]. The conventional definition of presence suggests that non-immersive displays ar inadequate, even for tasks that do not require the user to be surrounded and isolated Therefore, a new measure is needed to assess presence for non-immersive display that will more closely consider task requirements and how naturally a display support auser. “/t is here” is the idea that a display medium brings an object or person to th user [9]. This idea has only been investigated for conventional television programmes where it assesses the belief that the actual object being displayed exists within th television set. However, this concept can be extended to provide a measure for non immersive displays where the object appears to be in the user’s physical environment instead of inside the display.
Fig. 2. “Table-top Spatially Augmented Reality” [10] and “The HapticScreen” [6]. [he physical model could either be a static surrogate object [10] or a dynamically leforming physical simulation [6]. Both of these examples use relatively low- esolution objects although the projected image provides a more realistic visual ‘epresentation of the object’s surface (Fig 2). In this context, object-presence is a neasure of how much the presented object seems to exist, i.e. the combination of yhysical and visual information, not the existence of the physical object alone.
Fig. 2. “Table-top Spatially Augmented Reality” [10] and “The HapticScreen” [6]. [he physical model could either be a static surrogate object [10] or a dynamically leforming physical simulation [6]. Both of these examples use relatively low- esolution objects although the projected image provides a more realistic visual ‘epresentation of the object’s surface (Fig 2). In this context, object-presence is a neasure of how much the presented object seems to exist, i.e. the combination of yhysical and visual information, not the existence of the physical object alone.
Fig. 1. Visualisation of the Lumetila — Virtual Space (image by Tiina Kymaldinen, VTT Information Technology).
Fig. 1. Visualisation of the Lumetila — Virtual Space (image by Tiina Kymaldinen, VTT Information Technology).
The LumePong game is based on the well-known Pong game. In our virtual space version there is a real time 3D graphics environment where the player controls a ‘racquet? by moving around on the floor. The floor has pressure sensors for recognising the player’s movements: when a player goes forward, the ‘racquet’ goes up, when he moves backwards the ‘racquet’ goes down, and so on.
The LumePong game is based on the well-known Pong game. In our virtual space version there is a real time 3D graphics environment where the player controls a ‘racquet? by moving around on the floor. The floor has pressure sensors for recognising the player’s movements: when a player goes forward, the ‘racquet’ goes up, when he moves backwards the ‘racquet’ goes down, and so on.
Definitions and Scales for Handfeel Properties
Definitions and Scales for Handfeel Properties
Table 3. Handle Techniques. Effective Evaluation of Properties
Table 3. Handle Techniques. Effective Evaluation of Properties
Related topics:
Computer ScienceHuman Computer InteractionHaptic Technology
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved