Experimental Evaluation of Attachment Methods for a ... - CiteSeerX
Experimental Evaluation of Attachment Methods for a Multifinger Haptic Device Gina Donlin
Rainer Leuschke Blake Hannaford BioRobotics Laboratory Department of Electrical Engineering University of Washington Seattle, WA 98195 E-mail: (rdonlin, rainer, blake)@u.washington.edu
Abstract Due to the nature of touch sensation, the interface between a user and a haptic device must be carefully designed and selected such that the user feels they are directly manipulating a virtual object. This is particularly critical for attachment of fingertips to a multi fingered device. We experimentally tested eight methods of attachment for a Multifinger Haptic Device (MFHD) previously developed in our lab. From the test we gathered objective and subjective data on force and user preference. Based on this data we were able to evaluate the quality of the attachment methods.
1. Introduction A haptic device, which allows for interaction with a virtual object, is intended to be as transparent to the user as possible. The human-device interface is critical in rendering high quality haptic experiences such that the user feels as though they are directly manipulating an object. The interface should be carefully designed so as not to interfere with touch perception. Previous work has shown that the human sense of touch is extremely sensitive. Dosher et al. showed that on average subjects recognize forces as minute as 30mN [1]. The sense of touch can also be described as layered and extremely complex. Lederman and Klatzky defined stereotypical hand motions (Exploratory Procedures, EPs) characteristic of human haptic exploration. They found that the EP’s by the subjects were predictable based on what property of the material was being discriminated [2].
The procedures (and properties) they identified were: • • • • • • • •
Lateral Motion (texture) Pressure (hardness) Static Contact (temperature) Unsupported Holding (weight) Enclosure (shape, volume) Contour Following (exact shape, volume) Part Motion Test (part motion) Function Testing (specific function)
Current haptic technology is not capable of supporting all of the EP’s that would provide a complete user experience. Most haptic devices currently focus on a narrow combination of EPs, a specific element of touch. Devices offer free hand, multifinger or single finger exploration each requiring different human-device interfaces [5]. The commercially available Phantom line of devices from Sensable Inc allows the user to perform free space exploration through a stylus type interface or single finger exploration through a thimble interface [3]. The Ruters Master II-ND takes a novel multifinger approach by placing pneumatic actuators in the hand, affixed to the users’ fingers by straps [4]. The PERCO Lab has done extensive work in single and multifinger devices, featuring a rigid cuff with fingertip pad [5]. The most common attachment methods are gloves for the free hand devices and thimble attachments for the finger devices. To our knowledge, the question of how to best attach the human to a haptic device has not been the focus of experimentation. This paper describes experiments to investigate the feasibility of eight different finger attachment methods for a multifinger haptic device (MFHD) under development at the University of Washington BioRobotics Lab [6].
2. Methods To investigate this question further, we chose eight finger attachment methods to test; some that are already in use (thimble, glove) and some that we have not yet seen in haptic devices. The human-device attachment methods are: 1. Velcro glove 2. Elastic band 3. Thick strap 4. Bias force 5. Adhesive material 6. Magnet glove 7. Two magnets 8. Thimble attachment
2.1.2 Elastic band The elastic band is made from ¼ inch elastic that is affixed to the platform and wraps over the base of the user’s fingernail. To attach to the device, the user lifts the elastic band and positions their fingertip on the platform. Their finger is held tightly to the device by the elastic.
2.1 Methods of attachment The current setup of the device consists of a 1 inch by ¾ inch rectangular metal platform for the user to rest their finger on with a ½ by ¾ inch perpendicular stop at the front. It was intended for use with a bias force, a method that is studied in our experiment. The goal of our experiment was to acquire subjective and objective results for each of the attachment methods to evaluate its potential for use with the MFHD as well as other haptic displays. Detailed descriptions of each method can be found below.
Figure 1. An improved human-device interface is necessary for the MFHD. 2.1.1 Velcro glove The Velcro glove is made from a 1mm neoprene Seaskyn diving glove, size L. Velcro hooks are attached to platform of the device and the loops to the fingertips of the glove. Wearing the glove, the user presses the Velcro onto each platform to affix the glove to the device.
2.1.3 Thick strap The thick strap is composed of a ¾ inch elastic strap that loosely covers the user’s entire fingertip. To use this attachment method, the user simply slides their finger between the strap and the device. The thick strap was used for preliminary haptics research by Dosher [1] on one finger of the MFHD.
2.1.4 Bias force A force of magnitude of approximately one Newton points up and out to allow the user to rest their fingers on the platforms and push into the workspace to interact with virtual objects. Requiring no extra hardware, the bias force does not support opposing forces between the user and the device (weight or pulling up against a sticky surface, for example).
2.1.5 Adhesive material To attach the user using an adhesive material, we chose 3M Command strips as a method of attachment. This is a common foam adhesive tape. Since this method would need to be changed for each use, the adhesive was not preattached to the device. The user placed the adhesive on the device as part of the experimental protocol.
2.1.6 Magnet glove The magnet glove consists of an ATLAS Nitrile touch glove, size L, with rare earth magnets on the inside of the glove. On the platform of the device is an identical rare earth magnet embedded and covered by a piece of foam to make a level surface. The attraction between the two magnets hold the user’s gloved hand in place.
2.2 Experimental Protocol In order to determine whether the attachments could withstand a given amount of opposing force, as well as gather user feedback regarding comfort, ease of use and security of attachment for each method, two separate protocols were performed: a force test and a usability test. Objective and subjective data was collected from these experiments to evaluate each attachment method
2.1.7 Two-magnets Using the same rare earth magnet platform as used in the method above, a second magnet is placed on top of the user’s finger in the “twomagnets” method. The magnetic attraction between the two magnets sandwiches the user’s finger to the device.
2.2.1 Force test To test the strength of the connection between fingertip and device, we implemented a force test. The device rendered a sinusoidal vertical position trajectory but the finger, attached by one of the above methods, was held by the user in a fixed location. Gravity compensation of the device was enabled for this test. This allowed us to examine the ability of each attachment method to exert a force that pulled the device away from the stationary finger; applicable to interaction with a sticky surface or an object with attractive force. The user was asked to wear the attachment on the index finger, and hold their finger stationary at the center of the workspace. As the error between desired and actual y position increased, a proportional force was applied and resisted by the user. The force data was collected and used to determine whether the maximum force an attachment method could support was within the given threshold. The forces were calculated as follows:
2.1.8 Thimble We attached a thimble to the platform to form a finger cup for finger insertion. A thimble attachment is used in several other haptic devices including [3] [5] [7].
x = current horizontal position in the flexionextension plane [mm] y = current vertical position (same plane) [mm] xd = desired horizontal position [mm] yd= desired vertical position [mm] kp = proportionality constant [N/m] Fx= force in x direction (positive values point away from the user) [mN] Fy= force in the y direction, (positive values point up) [mN] Yd(t) =
1500mN 2π t sin ( ) 10 sec kp
Xd(t) = 0 Fx=kp*(x-xd)
Fy=kp*(y-yd) Since the user kept their finger at a constant y position, this trajectory generated an effective sinusoidal force on the finger attachment. We chose a maximum force of 1500 mN for this test. 1500 mN is well above this minimum requirement, enough to cause a noticeable strain on the finger when held at the center of the workspace and small enough to be overcome by the user so that it can be easily measured.
2.2.2 Subjective Assessment and Questionnaire For the subjective user test, we created a simple haptic plane as a surface for the user to interact with. Users were instructed to put on the attachment, enable the device and interact with the plane taking note of the following criteria: ease of use, comfort during use and security of attachment to the device. They then removed themselves from the device and were asked to give a rating from one to ten (ten being the best, one the worst) on each of the criteria. They repeated this procedure for each attachment method, and at the end selected their two favorite finger attachments for the MFHD. Nine subjects went through the protocol, including novice and experienced haptics users with a variety of hand sizes.
3. Results Figure 2 shows the results from the force test. The graphs give force as a function of time for both x and y forces. The x values were as expected, close to zero since the generated trajectory was vertical. Some of the graphs show failure of the attachment; these plots show the device disconnecting from the user, immediately followed by oscillation in force until the desired position once again matched the actual position. Since the bias force method does not allow the user to exert opposing force on the device, the attachment could not be held at the center of the workspace. The desired position was always equal to the actual position, so no additional force was rendered. Results from the force test show the Velcro glove, elastic band, thick strap, magnet glove and thimble were all able to withstand forces at or above approximately 1500 mN. The data shows that the bias force, adhesive material and two magnets attachment methods have a force threshold lower than 1500 mN.
Figure 2. Force versus time for the eight force tests. Applied force vs. time for each finger attachment method. Force appears sinusoidal when finger is
successfully attached to device. Break of attachment is followed by a transient as the device goes to desired position. Green – vertical force, Blue – horizontal force. In the subjective test, each user was asked to give a 0-10 score for three attributes (east of use, comfort, and security) for each attachment method, and the total was computed. The mean and standard deviation are calculated for the total scores for each method. On a separate scale, attachment methods were ranked based on user preference at the end of the test. An attachment chosen as a user’s first choice received two points, second choice received one point and all others received zero points. Table 1 gives a summary of the results from the user experiment. Fingertip Attachment Method Velcro Glove Thimble Magnet Glove Elastic Band Adhesive Material Thick Strap Bias Force Two magnets
Total score (mean) 20.67
Standard deviation 3.54
Score based on preference 9
20.22 19.89
2.59 4.54
2 5
19.11 15.44
3.06 5.81
6 3
14.78 14.67 14.22
2.22 2.83 3.19
0 0 2
Table 1: Results from user questionnaire
4. Discussion According to the results, the preferred finger attachment method for the MFHD is the Velcro glove. Users liked that it fits comfortably and is easy to use, however, the Velcro tends to separate slightly from the device during both the force test and user test. While not leading to failure of the force test, this separation is problematic as the Velcro gives an audible and tactile sensation of separating from the device when the angle of the finger changes with respect to the plane of the platform. In terms of comfort, the material of the glove, neoprene, was not very breathable and some commented that this would cause discomfort if worn for an extended period of time.
Figure 3. Separation between the Velcro and the device. While being manipulated in free space users found the thick strap method to be too loose, making it unresponsive to the user’s movements. The bias force was the most difficult to use. Since the end joint is free to rotate, the user’s finger tended to rotate over the top of the device and fall off, making the device impossible to control.
Figure 4. Fingertip rotation with bias force attachment method This issue of pad rotation was evident in many of the methods, but without a secure attachment method only the bias force method was completely unusable. For methods that held the finger firmly, like the elastic band, the device did not come off the finger. Rotation around that axis caused a change in elevation of the fingertip, and users commented that this distorted the sensation of the haptic plane. In our tests, the thimble scored well for ease of use and comfort; however, the generic off-the-shelf thimble that was used in this experiment would not be appropriate for use with all four fingers. The users only used their index finger in this experiment, and there are large size discrepancies between the fingers of just one user’s hand, even more so between users. A thimblelike attachment that was either adjustable or available in different sizes would be necessary to be considered further. When using the two magnets method, we found that there was a noticeable magnetic effect in the back corner of the workspace, where the magnets were attracted to the magnets in the device. While this was
not noticeable in the magnetic glove, the effect was so strong in the affected area with the two magnets method that the attachment stuck directly to the device and was removed with two hands. The magnet glove could be greatly improved if the type of glove used and/or size were changed. We are currently working on implementing a tighter glove with a fit similar to the Velcro glove. If this method is chosen as a potential candidate, an in-depth study will be necessary to ascertain and compensate for the magnetic interactions between device and finger attachment in software.
2. 3. 4. 5.
6.
Figure 5. The corner of the workspace is problematic for the magnetic finger attachment methods. The next steps for the finger attachment study include modification of the magnet glove, elastic band and thimble. The magnet glove will use a smaller glove with magnets permanently attached to the glove in the correct orientation for use with the device. Since the thimble method has worked so well for others, we will modify our current setup to give it a more general fit for different sized fingers. We plan to perform experiments with the magnetic attachments to determine if they can easily be implemented without distortion of perception in the corner of the workspace. This experiment gave valuable insight into the feasibility of each attachment method for use with the MFHD as well as other haptic devices.
5. Acknowledgements The authors gratefully acknowledge the support from the National Science Foundation (grant IIS-0303750). The authors would also like to thank Mitch Lum, Levi Miller and Hawkeye King for their help with this project and an anonymous reviewer.
6. References: 1.
J. Dosher, G. Lee, and B. Hannaford. How low can you go? Detection thresholds for small haptic effects. In Margret McLaughlin, editor, Touch in Virtual Environments, Proc. USC Workshop on Haptic Interfaces. Prentice Hall, Feb 23, 2001.
S. Lederman and R. Klatzky. Hand movements: a window into haptic object recognition. Cognitive Psychology, 19(3):342-368, 1987. www.sensable.com M. Bouzit, G. Popescu, G. Burdea and R. Boin, The Ruters Master II-ND Force Feedback Glove. IEEE/ASME Transactions on Mechatronics. G. Cini, A. Frisoli, S. Marcheschi, F. Salsedo and M. Bergamasco. A novel fingertip haptic device for display of local contact geometry. Proc. Of the First Joint Eurohaptics Conference and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator System, 2005 R. Leuschke, E.K.T. Kurihara, J. Dosher and B. Hannaford. High fidelity multi-finger haptic display. Proceedings World Haptics Congress, 2005, pages 606-608, March 2005. 0H
Rainer Leuschke Blake Hannaford BioRobotics Laboratory Department of Electrical Engineering University of Washington Seattle, WA 98195 E-mail: (rdonlin, rainer, blake)@u.washington.edu
Abstract Due to the nature of touch sensation, the interface between a user and a haptic device must be carefully designed and selected such that the user feels they are directly manipulating a virtual object. This is particularly critical for attachment of fingertips to a multi fingered device. We experimentally tested eight methods of attachment for a Multifinger Haptic Device (MFHD) previously developed in our lab. From the test we gathered objective and subjective data on force and user preference. Based on this data we were able to evaluate the quality of the attachment methods.
1. Introduction A haptic device, which allows for interaction with a virtual object, is intended to be as transparent to the user as possible. The human-device interface is critical in rendering high quality haptic experiences such that the user feels as though they are directly manipulating an object. The interface should be carefully designed so as not to interfere with touch perception. Previous work has shown that the human sense of touch is extremely sensitive. Dosher et al. showed that on average subjects recognize forces as minute as 30mN [1]. The sense of touch can also be described as layered and extremely complex. Lederman and Klatzky defined stereotypical hand motions (Exploratory Procedures, EPs) characteristic of human haptic exploration. They found that the EP’s by the subjects were predictable based on what property of the material was being discriminated [2].
The procedures (and properties) they identified were: • • • • • • • •
Lateral Motion (texture) Pressure (hardness) Static Contact (temperature) Unsupported Holding (weight) Enclosure (shape, volume) Contour Following (exact shape, volume) Part Motion Test (part motion) Function Testing (specific function)
Current haptic technology is not capable of supporting all of the EP’s that would provide a complete user experience. Most haptic devices currently focus on a narrow combination of EPs, a specific element of touch. Devices offer free hand, multifinger or single finger exploration each requiring different human-device interfaces [5]. The commercially available Phantom line of devices from Sensable Inc allows the user to perform free space exploration through a stylus type interface or single finger exploration through a thimble interface [3]. The Ruters Master II-ND takes a novel multifinger approach by placing pneumatic actuators in the hand, affixed to the users’ fingers by straps [4]. The PERCO Lab has done extensive work in single and multifinger devices, featuring a rigid cuff with fingertip pad [5]. The most common attachment methods are gloves for the free hand devices and thimble attachments for the finger devices. To our knowledge, the question of how to best attach the human to a haptic device has not been the focus of experimentation. This paper describes experiments to investigate the feasibility of eight different finger attachment methods for a multifinger haptic device (MFHD) under development at the University of Washington BioRobotics Lab [6].
2. Methods To investigate this question further, we chose eight finger attachment methods to test; some that are already in use (thimble, glove) and some that we have not yet seen in haptic devices. The human-device attachment methods are: 1. Velcro glove 2. Elastic band 3. Thick strap 4. Bias force 5. Adhesive material 6. Magnet glove 7. Two magnets 8. Thimble attachment
2.1.2 Elastic band The elastic band is made from ¼ inch elastic that is affixed to the platform and wraps over the base of the user’s fingernail. To attach to the device, the user lifts the elastic band and positions their fingertip on the platform. Their finger is held tightly to the device by the elastic.
2.1 Methods of attachment The current setup of the device consists of a 1 inch by ¾ inch rectangular metal platform for the user to rest their finger on with a ½ by ¾ inch perpendicular stop at the front. It was intended for use with a bias force, a method that is studied in our experiment. The goal of our experiment was to acquire subjective and objective results for each of the attachment methods to evaluate its potential for use with the MFHD as well as other haptic displays. Detailed descriptions of each method can be found below.
Figure 1. An improved human-device interface is necessary for the MFHD. 2.1.1 Velcro glove The Velcro glove is made from a 1mm neoprene Seaskyn diving glove, size L. Velcro hooks are attached to platform of the device and the loops to the fingertips of the glove. Wearing the glove, the user presses the Velcro onto each platform to affix the glove to the device.
2.1.3 Thick strap The thick strap is composed of a ¾ inch elastic strap that loosely covers the user’s entire fingertip. To use this attachment method, the user simply slides their finger between the strap and the device. The thick strap was used for preliminary haptics research by Dosher [1] on one finger of the MFHD.
2.1.4 Bias force A force of magnitude of approximately one Newton points up and out to allow the user to rest their fingers on the platforms and push into the workspace to interact with virtual objects. Requiring no extra hardware, the bias force does not support opposing forces between the user and the device (weight or pulling up against a sticky surface, for example).
2.1.5 Adhesive material To attach the user using an adhesive material, we chose 3M Command strips as a method of attachment. This is a common foam adhesive tape. Since this method would need to be changed for each use, the adhesive was not preattached to the device. The user placed the adhesive on the device as part of the experimental protocol.
2.1.6 Magnet glove The magnet glove consists of an ATLAS Nitrile touch glove, size L, with rare earth magnets on the inside of the glove. On the platform of the device is an identical rare earth magnet embedded and covered by a piece of foam to make a level surface. The attraction between the two magnets hold the user’s gloved hand in place.
2.2 Experimental Protocol In order to determine whether the attachments could withstand a given amount of opposing force, as well as gather user feedback regarding comfort, ease of use and security of attachment for each method, two separate protocols were performed: a force test and a usability test. Objective and subjective data was collected from these experiments to evaluate each attachment method
2.1.7 Two-magnets Using the same rare earth magnet platform as used in the method above, a second magnet is placed on top of the user’s finger in the “twomagnets” method. The magnetic attraction between the two magnets sandwiches the user’s finger to the device.
2.2.1 Force test To test the strength of the connection between fingertip and device, we implemented a force test. The device rendered a sinusoidal vertical position trajectory but the finger, attached by one of the above methods, was held by the user in a fixed location. Gravity compensation of the device was enabled for this test. This allowed us to examine the ability of each attachment method to exert a force that pulled the device away from the stationary finger; applicable to interaction with a sticky surface or an object with attractive force. The user was asked to wear the attachment on the index finger, and hold their finger stationary at the center of the workspace. As the error between desired and actual y position increased, a proportional force was applied and resisted by the user. The force data was collected and used to determine whether the maximum force an attachment method could support was within the given threshold. The forces were calculated as follows:
2.1.8 Thimble We attached a thimble to the platform to form a finger cup for finger insertion. A thimble attachment is used in several other haptic devices including [3] [5] [7].
x = current horizontal position in the flexionextension plane [mm] y = current vertical position (same plane) [mm] xd = desired horizontal position [mm] yd= desired vertical position [mm] kp = proportionality constant [N/m] Fx= force in x direction (positive values point away from the user) [mN] Fy= force in the y direction, (positive values point up) [mN] Yd(t) =
1500mN 2π t sin ( ) 10 sec kp
Xd(t) = 0 Fx=kp*(x-xd)
Fy=kp*(y-yd) Since the user kept their finger at a constant y position, this trajectory generated an effective sinusoidal force on the finger attachment. We chose a maximum force of 1500 mN for this test. 1500 mN is well above this minimum requirement, enough to cause a noticeable strain on the finger when held at the center of the workspace and small enough to be overcome by the user so that it can be easily measured.
2.2.2 Subjective Assessment and Questionnaire For the subjective user test, we created a simple haptic plane as a surface for the user to interact with. Users were instructed to put on the attachment, enable the device and interact with the plane taking note of the following criteria: ease of use, comfort during use and security of attachment to the device. They then removed themselves from the device and were asked to give a rating from one to ten (ten being the best, one the worst) on each of the criteria. They repeated this procedure for each attachment method, and at the end selected their two favorite finger attachments for the MFHD. Nine subjects went through the protocol, including novice and experienced haptics users with a variety of hand sizes.
3. Results Figure 2 shows the results from the force test. The graphs give force as a function of time for both x and y forces. The x values were as expected, close to zero since the generated trajectory was vertical. Some of the graphs show failure of the attachment; these plots show the device disconnecting from the user, immediately followed by oscillation in force until the desired position once again matched the actual position. Since the bias force method does not allow the user to exert opposing force on the device, the attachment could not be held at the center of the workspace. The desired position was always equal to the actual position, so no additional force was rendered. Results from the force test show the Velcro glove, elastic band, thick strap, magnet glove and thimble were all able to withstand forces at or above approximately 1500 mN. The data shows that the bias force, adhesive material and two magnets attachment methods have a force threshold lower than 1500 mN.
Figure 2. Force versus time for the eight force tests. Applied force vs. time for each finger attachment method. Force appears sinusoidal when finger is
successfully attached to device. Break of attachment is followed by a transient as the device goes to desired position. Green – vertical force, Blue – horizontal force. In the subjective test, each user was asked to give a 0-10 score for three attributes (east of use, comfort, and security) for each attachment method, and the total was computed. The mean and standard deviation are calculated for the total scores for each method. On a separate scale, attachment methods were ranked based on user preference at the end of the test. An attachment chosen as a user’s first choice received two points, second choice received one point and all others received zero points. Table 1 gives a summary of the results from the user experiment. Fingertip Attachment Method Velcro Glove Thimble Magnet Glove Elastic Band Adhesive Material Thick Strap Bias Force Two magnets
Total score (mean) 20.67
Standard deviation 3.54
Score based on preference 9
20.22 19.89
2.59 4.54
2 5
19.11 15.44
3.06 5.81
6 3
14.78 14.67 14.22
2.22 2.83 3.19
0 0 2
Table 1: Results from user questionnaire
4. Discussion According to the results, the preferred finger attachment method for the MFHD is the Velcro glove. Users liked that it fits comfortably and is easy to use, however, the Velcro tends to separate slightly from the device during both the force test and user test. While not leading to failure of the force test, this separation is problematic as the Velcro gives an audible and tactile sensation of separating from the device when the angle of the finger changes with respect to the plane of the platform. In terms of comfort, the material of the glove, neoprene, was not very breathable and some commented that this would cause discomfort if worn for an extended period of time.
Figure 3. Separation between the Velcro and the device. While being manipulated in free space users found the thick strap method to be too loose, making it unresponsive to the user’s movements. The bias force was the most difficult to use. Since the end joint is free to rotate, the user’s finger tended to rotate over the top of the device and fall off, making the device impossible to control.
Figure 4. Fingertip rotation with bias force attachment method This issue of pad rotation was evident in many of the methods, but without a secure attachment method only the bias force method was completely unusable. For methods that held the finger firmly, like the elastic band, the device did not come off the finger. Rotation around that axis caused a change in elevation of the fingertip, and users commented that this distorted the sensation of the haptic plane. In our tests, the thimble scored well for ease of use and comfort; however, the generic off-the-shelf thimble that was used in this experiment would not be appropriate for use with all four fingers. The users only used their index finger in this experiment, and there are large size discrepancies between the fingers of just one user’s hand, even more so between users. A thimblelike attachment that was either adjustable or available in different sizes would be necessary to be considered further. When using the two magnets method, we found that there was a noticeable magnetic effect in the back corner of the workspace, where the magnets were attracted to the magnets in the device. While this was
not noticeable in the magnetic glove, the effect was so strong in the affected area with the two magnets method that the attachment stuck directly to the device and was removed with two hands. The magnet glove could be greatly improved if the type of glove used and/or size were changed. We are currently working on implementing a tighter glove with a fit similar to the Velcro glove. If this method is chosen as a potential candidate, an in-depth study will be necessary to ascertain and compensate for the magnetic interactions between device and finger attachment in software.
2. 3. 4. 5.
6.
Figure 5. The corner of the workspace is problematic for the magnetic finger attachment methods. The next steps for the finger attachment study include modification of the magnet glove, elastic band and thimble. The magnet glove will use a smaller glove with magnets permanently attached to the glove in the correct orientation for use with the device. Since the thimble method has worked so well for others, we will modify our current setup to give it a more general fit for different sized fingers. We plan to perform experiments with the magnetic attachments to determine if they can easily be implemented without distortion of perception in the corner of the workspace. This experiment gave valuable insight into the feasibility of each attachment method for use with the MFHD as well as other haptic devices.
5. Acknowledgements The authors gratefully acknowledge the support from the National Science Foundation (grant IIS-0303750). The authors would also like to thank Mitch Lum, Levi Miller and Hawkeye King for their help with this project and an anonymous reviewer.
6. References: 1.
J. Dosher, G. Lee, and B. Hannaford. How low can you go? Detection thresholds for small haptic effects. In Margret McLaughlin, editor, Touch in Virtual Environments, Proc. USC Workshop on Haptic Interfaces. Prentice Hall, Feb 23, 2001.
S. Lederman and R. Klatzky. Hand movements: a window into haptic object recognition. Cognitive Psychology, 19(3):342-368, 1987. www.sensable.com M. Bouzit, G. Popescu, G. Burdea and R. Boin, The Ruters Master II-ND Force Feedback Glove. IEEE/ASME Transactions on Mechatronics. G. Cini, A. Frisoli, S. Marcheschi, F. Salsedo and M. Bergamasco. A novel fingertip haptic device for display of local contact geometry. Proc. Of the First Joint Eurohaptics Conference and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator System, 2005 R. Leuschke, E.K.T. Kurihara, J. Dosher and B. Hannaford. High fidelity multi-finger haptic display. Proceedings World Haptics Congress, 2005, pages 606-608, March 2005. 0H
