Tactile sensory feedback, or haptics, is a fundamental element of life. When prosthetic limbs are provided for persons with amputations, this key function is largely missing. Typically, prosthetic users must rely heavily on visual feedback by watching their prosthetic hand. However, it has been shown that visual feedback requires a high cognitive demand of the prosthetic user.1 Other forms of indirect feedback include listening to the sound of the prosthetic hand motors, sensing motor vibrations and changes in pressures on the residual limb, and correlating the closing velocity of the prosthetic hand with the grip force generated.2,3 However, to improve grip force accuracy and/or reduce cognitive demand when using the prosthetic hand, additional haptic feedback is needed.
To improve hand control, researchers have investigated the use of sensation modality matching (touch, pressure, shear, and temperature) for closed-loop prosthetic control. Meek et al.4 studied a force feedback system that used a motor-driven pusher to apply a perpendicular force to the user's skin in proportion to the force at the prosthetic hand. Patterson et al.5 used a pressure cuff system that applied pressure to the arm in proportion to pressure on the prosthetic hand. Sensation modality matching and somatotopic matching (feeling on the phantom hand) were tested in two individuals with upper-limb amputation with targeted reinnervation.6 These studies consistently found that the addition of any supplemental haptic feedback could improve prosthetic users' control of their grasping force. Because the direct force or pressure feedback systems have tended to be large and cumbersome with demanding power consumption,4 vibrotactile feedback may provide a better alternative for clinical use as they are small, unobtrusive, and require less power.7 Therefore, if vibrotactile feedback also improves grip force control, it could more quickly and cost-effectively be transitioned into a commercial product.
Previously, vibration producing tactors have been shown to improve object manipulation in unimpaired persons,3,8–10 prosthetic hand grasping force accuracy in unimpaired persons with prior experience in using vibrotactile feedback,7 and prosthetic hand grasping force accuracy in upper-limb prosthetic users.11 Each of these studies used different tactor vibration characteristics to convey increasing and decreasing grip force.
Sensation from vibration producing tactors can be varied by changing the amplitude, frequency (dependent on speed of vibration motors), or pulse rate (dependent on the rate that the vibration motor is turned on and off). One method for increasing the amplitude of the vibrotactile sensation is to stack more than one tactor coaxially to generate constructive vibrations.12 Results using this method demonstrated an amplitude discrimination of 75% to near 100% when assessing one to three tactors. While these results are promising, stacking tactors coaxially would make concealing these tactors in the prosthesis more challenging.
Frequency and intensity of the vibration of the eccentric rotating mass (ERM) type of tactors can be altered by changing the voltage driving the motors. A study found that the discrimination between 156 Hz and 122 Hz was 92% and that between 156 Hz and 140 Hz was 86%,12 whereas another study found that participants were able to discriminate 6 levels of vibration frequency ranging from 60 to 360 Hz13 and later used 50 to 80 Hz to improve grip force accuracy in five myoelectric prosthetic hand users.14 A combination of frequency and amplitude resulted in subjects' ability to discriminate six different stimulation patterns with an accuracy of 78%.12 However, one drawback to frequency control with current ERM vibration motor technology is its sensitivity to motor restraint15 and the inability to accurately control the frequency of the vibration. When a constant electrical signal is applied to an ERM vibration motor, a motor that is mounted more loosely to a person's arm will vibrate at a lower frequency than a motor that is more firmly restrained.15 Therefore, a change in arm position could change the motor restraint and, therefore, frequency of the motor's vibration.
Varying the pulse rate of the vibration is another vibration characteristic that can also be used to better differentiate sensations. Pulse rates are created by starting and stopping the vibration of the tactor at specific rates. Pulsing has the advantage of being independent of motor frequency. Previously, pulse rates were used to improve prosthetic hand grasping force accuracy in unimpaired persons with prior experience in using vibrotactile feedback7 and in upper-limb prosthetic users.11 The latter study, in upper-limb prosthetic users, conveyed increasing grip force by keeping the carrier frequency constant and increasing the pulse rate between 1 and 8 Hz with 8 Hz reserved for the maximum grip force of the individual with amputation.11 Overall, subjects were 80% accurate at detecting differences in 1-Hz pulse rate changes. To further improve accuracy of discrimination in vibratory feedback, the Chaubey et al. study concluded that “further improvements to the interpretation of vibratory feedback may occur if 1) amplitude and/or carrier frequency along with pulse rate of the vibratory feedback simultaneously increased with increasing [grip] force and 2) vibration sensation increased in steps with distinct force levels such as light, medium, and strong grip force rather than a unique vibration sensation with all unique forces.” This conclusion recognizes the limits of people being able to discriminate between multiple vibration patterns. To achieve better force discrimination, the number of distinct vibration patterns, intended to convey grip force, may need to be limited to three general levels: light (or contact), medium, and strong grip force.
This study is the part 2 of the study of Chaubey et al.11 that investigates the usefulness of providing three distinct vibration patterns to represent light, medium, and strong grip force. This study is also the first study to our knowledge to create a portable vibratory haptic feedback system for use outside of the laboratory. Testing the usefulness of vibratory haptic feedback during daily activities is an important step because improved gripping accuracy alone is irrelevant if it does not also improve the user experience in his/her daily life.
The objectives were as follows: 1) optimize tactor placement within the socket; 2) optimize unique vibration patterns to represent three grip forces; 3) create a portable vibratory haptic feedback system for use in daily life; 4) test grip accuracy differences with and without vibratory haptic feedback; 5) test the ability to perform daily gripping tasks with and without vibratory haptic feedback; and 6) survey prosthetic users' opinion of the system after using the vibratory haptic device at home.
DEVELOPMENT FOR THE NONPORTABLE HAPTIC FEEDBACK SYSTEM (OBJECTIVES 1 AND 2)
The nonportable haptic feedback system was an intermediate step needed before building the portable haptic feedback system for home use. It allowed efficient testing in objectives 1 and 2. There are many technical details in regard to building the nonportable haptic feedback system (Table 1).
TACTOR PLACEMENT AND VIBRATION DIFFERENTIATION TESTING (OBJECTIVE 1)
Two tactors were placed on the proximal, middle, and distal locations below the elbow on the residual limb as long as there was room on the residual limb to have nonoverlapping test locations that did not interfere with the myoelectrodes. Tactors were not placed over boney prominences. The settings displayed in Table 2 were used for the location testing. The subjects were blinded to the settings. Each vibration test location was compared with the previous vibration test. The subject was asked if there was a change in the vibration. If so, the subject stated whether the change involved an increase in pulse rate, decrease in pulse rate, or if it went from pulsing to continuous. In addition, the subjects were asked if they associated the change in vibration as an increase in vibration intensity or decrease in vibration intensity with the assumption that a higher intensity vibration would coincide with a higher grip force in the haptic feedback system. Tests 1 to 16 in Table 2 were performed in the same order for each subject because randomization would change how difficult it was to detect change between two consecutive tests.
OPTIMIZE UNIQUE VIBRATION PATTERNS TO REPRESENT THREE GRIP FORCES IN A MAPPING ALGORITHM (OBJECTIVE 2)
Voltage differentials that were output from the grip sensor to represent light (or contact), medium, and strong grip force were set in the software based on investigator testing with the prosthetic hands before subject testing. The voltage differential to represent light (or contact), medium, and strong grip force was chosen based on the average voltage differential at 1.5 to 2.5 lb (light), 9.0 to 11.0 lb (medium), and 19.0 to 21.0 lb (strong) of grip force as detected by the Microfet4 (Hoggan Scientific LLC, Salt Lake City, UT, USA). Both cylindrical and pinch grips were used, and testing was performed over several days to choose voltage differentials that were less influenced by the position of the prosthetic silicone glove. Once the voltage level was reached that represented a target grip force, the feedback would remain the same until a different target grip force was reached (Figure 4). For example, the light grip vibration characteristics (i.e., 2 Hz pulse, 50% duty cycle, 65% carrier frequency duty cycle) would trigger at the light grip voltage level and continue until the medium grip voltage level was reached. At the medium grip voltage level, the vibration characteristics would change to indicate medium grip force (i.e., 4 Hz pulse, 50% duty cycle, 100% carrier frequency duty cycle). Medium grip force vibration characteristics would continue until grip force increased to strong grip force (triggering strong grip force vibration characteristics) or decreased down to light grip force (triggering light grip force vibration characteristics).
To prevent the tactors from being activated and deactivated erratically when the force-sensitive resistor (FSR) fluttered around the target grip force, a 90% hysteresis (step down scaler) was applied. This meant that once the target voltage was reached, the voltage would have to drop by more than 10% of the target voltage level to change the vibration characteristics to the next level down (Figure 4). In the portable haptic feedback system, the tactors were set to turn off, regardless of the grip force level, if the grip force level remained stable for 5 or more seconds. This was to reduce the likelihood of desensitization to the vibration feedback and to not annoy the prosthetic user if the user was using the same grip for longer than 5 seconds. Once the grip changed to a different grip force level, the tactors would reactivate to represent the appropriate level.
GRIP FORCE ACCURACY (OBJECTIVES 2 AND 4)
To confirm that the mapping algorithm was working, grip accuracy tests were performed with the nonportable vibratory haptic feedback system using the same test structure as for the portable haptic feedback system in objective 4 (Table 3). The grip accuracy tests used the Microfet4 as the instrumented grip object. The subjects were asked to go through each of the trials described in Table 3. The subject was given 5 seconds to reach the target grip force. One practice session was allowed with the vibratory feedback so that the subjects could feel what to expect from the tactors and to verify that each subject understood the instructions of the accuracy test. Sets II and III allowed the subject to watch the custom graphic user interface (GUI) (Figure 3) and use the black feedback needle on the meter to match their actual grip forces (as detected by the Microfet4) to the red needle that showed the target grip force (2 lb, 10 lb, or 20 lb). Sets IV and V involved the use of vibratory haptic feedback without visual feedback from the GUI. Voltage differentials from the FSR and actual grip force from the Microfet4 were synchronized and recorded in the software at 18.6 Hz. The following metrics were calculated from each trial:
1. Grip force accuracy. Accuracy was calculated by determining the absolute difference between the final grip force at the end of a 5-second trial and the target grip force. This amount was then divided by the target grip force. Five seconds was an appropriate amount of time to achieve and maintain a target grip force based on previous testing.11
2. Oscillation. Oscillation calculated how much the subject over and under shot their intended final grip force. It was noted that almost all of the subjects oscillated around the target force before settling on the final grip force. Once the target grip force was past, the maximum grip force was recorded and then the minimum grip force that followed the maximum grip force was recorded. Oscillation was the difference between this maximum and minimum grip force, which was then divided by the subject's final grip force at the end of the 5-second trial. A low oscillation score indicated that the subject got to their final grip force more directly without a lot of seeking behavior of over and under gripping.
DEVELOPMENT OF THE PORTABLE VIBRATORY HAPTIC FEEDBACK SYSTEM (OBJECTIVE 3)
The portable vibratory haptic feedback system used the same prosthetic hands with the grip force sensor system as used in the nonportable vibratory haptic feedback system. However, rather than wiring the FSR output to a DAQ, the FSR voltage was transmitted wirelessly to the main controller unit on the patients' existing prosthetic arm using a Bluetooth connection (Bluegiga BLE112). Two lithium polymer cells (PL-402248–2C; AA Portable Power Corp, Richmond, CA, USA), signal conditioning circuit, analog/digital convertor, and Bluetooth transmitter were fitted under the prosthetic glove on the posterior side of the prosthetic hand distal to the wrist (Figure 5). The Bluegiga BLE112 on the prosthetic arm received the radio frequency link from the hand and the transmitted voltage thresholds representing light, medium, and strong grip from the hand, which were preset using the researcher's custom GUI. When accuracy testing with the portable haptic system, a WebSockets-based Web app was used to synchronize and record data from the haptic system and Microfet4 grip force.
Once the voltage thresholds were set for the prosthetic hand being used, input from the GUI was not needed. The information was then transferred to the firmware for processing (MSP430; Texas Instruments Inc, Dallas, TX, USA) using the mapping algorithm, the motor drive circuit, and then the tactor motors. The tactors were coated in a silicone shell to protect the subjects from the tactors warming and to protect the tactors from moisture (sweat) generated during daily use (Figure 6). The tactors were placed under the prosthetic liner and could be turned off by a switch on the prosthesis in case the tactor feedback was unhelpful or malfunctioned. The power for the Bluetooth and MSP430 in the prosthetic arm was from the myoelectric arm battery with a 3.5 voltage regulator. This design was appropriate for short-term testing at home. A more robust design will be needed with unexposed wires going to the tactors for long-term clinical use.
ACMC TEST WITH PORTABLE VIBRATORY HAPTIC SYSTEM (OBJECTIVE 5)
The Assessment of Capacity for Myoelectric Control (ACMC) was used for assessing the subject's capacity for myoelectric prosthetic hand control both with and without haptic feedback. This analysis has been validated for detecting expected change in the ability of the myoelectric prosthetic users.16 The ACMC test requires the subjects to perform various 2-handed tasks while an occupational therapist assesses their capacity for control of their myoelectric prosthesis by rating their performances on 30 items representing different aspects of myoelectric control quality. The items in the ACMC are classified into four groups: 1) gripping (12 items); 2) holding (6 items); 3) releasing (10 items); and 4) coordinating between hands (2 items). Each person's performance is rated with scores ranging from 0 to 3, where 0 = not capable, 1 = sometimes capable, capacity not established, 2 = capable on request, and 3 = spontaneously capable.
To actively engage both hands for the ACMC test, the protocol involved setting out a simple lunch. This activity involved putting out a tablecloth, putting out plates from shelves and drawers, folding napkins, cutting donuts, pouring water into mugs, and cleaning up. Before the subjects perform the “setting out lunch” activity for the ACMC test, the activity was critiqued and approved as appropriate by an ACMC expert, Dr Hermansson.16,17 Subjects performed the setting out lunch activity twice: once with vibratory haptics on and once with vibratory haptics off. A coin flip determined the order of the ACMC test trials. Each trial was video recorded. Because experience has been shown to be important to the reliability of the ACMC,17 Dr Hermansson consulted by viewing the videos and performing the ACMC test analysis while remaining blinded to whether the vibratory haptics were on or off. The Rasch analysis was performed to achieve an overall score for each ACMC test.
DAILY USE OF PORTABLE VIBRATORY HAPTIC SYSTEM (OBJECTIVE 6)
After the ACMC test, subjects took home the portable vibratory haptic system for use up to 3 days. After use, the subject responded to the following questions where the rating could be an integer between 0 (poor) and 4 (excellent):
- How comfortable was the vibrotactile feedback?(0 = extremely uncomfortable, 1 = somewhat uncomfortable, 2 = neutral, 3 = reasonably comfortable, or 4 = extremely comfortable)
- How well did you achieve the correct grip pressure with the vibrotactile feedback?(0 = not correct, 1 = somewhat correct, 2 = neutral, 3 = reasonably correct, 4 = very correct)
- Rate your ability to operate a myoelectric hand and react to the vibratory sensation at the same time?(0 = not well, 1 = somewhat well, 2 = neutral, 3 = reasonably well, 4 = extremely well)
The following open-ended questions were asked:
- 1) Overall, what was your impression of having vibratory haptic feedback?
- 2) If a commercial product was pursued, how would you like it to be different?
For determining if the proximal, middle, or distal placement of the tactors on the residual limb under the socket had a statistical advantage, a paired t-test was used to compare the two top performing sites. Grip force accuracy was determined by comparing the accuracy error percentage and oscillation score between set I (control) and sets IV and V with the vibratory haptic feedback. To account for the lack of independence introduced by using repeated measurements on the same study subjects, a mixed effects linear regression model was used. Subjects were included as a random effect to allow inferences to the population of subjects, with each subject having his or her own intercept. The remaining factors were treated as fixed effects to make inferences to the specific values of the factors: sets (I, IV, V) and target force (light, medium, strong). The factor trial (1, 2, or 3) was treated as random effect, being nested within subject, for sets that did not provide feedback (set I). For sets that did provide feedback, the factor trial was first included as a fixed effect to assess a sequential learning effect, and then included as a random effect, simply nested within subject, if the learning effect was not present or inconsequential. To analyze the ACMC test, the ACMC scores were compared using a paired t-test. Qualitative statistics were reported for the survey scores, and responses to the open-ended questions were summarized.
Research protocols were approved by the Western Institutional Review Board, and six transradial subjects were recruited for this study. Their average age was 59 ± 7.4 years. The inclusion criteria were myoelectric prosthetic users with a transradial amputation between the ages of 21 and 75. Subjects also had to be at least 3 years after the amputation, comfortable in their existing socket, comfortable with a computer, and able to perform test activities with their myoelectric hand for 2 hrs. Potential subjects were excluded if they had extensive upper-limb neuropathy that would interfere with vibration detection, regardless of where the tactors are placed; had a confounding injury or musculoskeletal problem; were pregnant, had a skin condition that lead to abnormal sensitivity; or could not read and understand English.
TACTOR PLACEMENT AND VIBRATION DIFFERENTIATION TESTING RESULTS
This test confirmed that in general the distal end of the residual limb is the least sensitive to vibratory feedback due to scar tissue and/or damaged nerves from the amputation. In those with sensation to touch both proximally and in the middle position, the proximal and middle position could equally detect a change in vibration but, on average, the proximal position was more sensitive to vibration intensity interpretation (increase in pulse rate or carrier frequency duty cycle was associated with increase in grip force). The amount of improvement in intensity interpretation in the proximal location was on average 13%. This improvement was not statistically significant (P = 0.10), indicating that tactor placement should be checked proximally then in the middle of the residual limb for appropriate tactor placement. Testing also confirmed that each subject could detect differences in the vibration sensation with the following characteristics: 1) 2 Hz pulse rate, 50% duty cycle, 65% to 80% carrier frequency duty cycle; 2) 4 Hz pulse rate, 50% duty cycle, 100% carrier frequency duty cycle; 3) 100% duty cycle (continuous vibration), 90% to 100% carrier frequency duty cycle (Table 2).
GRIP ACCURACY TESTS WITH NONPORTABLE AND PORTABLE VIBRATORY HAPTIC SYSTEM RESULTS
The accuracy tests with the haptic versus no haptic feedback trials demonstrated that haptic feedback statistically improved the grip force accuracy by 129% (P = 0.024, adjusted P = 0.041) for the light grip force target (2 lb) in the nonportable system and by 21% (P = 0.029, adjusted P = 0.051) for medium grip force target (10 lb) in the portable system (Figure 7). There was no significant improvement for strong grip force levels.
ACMC TEST WITH PORTABLE VIBRATORY HAPTIC SYSTEM RESULTS
Five research subjects were available for the ACMC test. All five subjects needed a right-hand prosthesis so only hand serial number 1039 was used. The test was performed with and without haptics. The first subject to use the portable vibratory haptic system had difficulty using the myoelectric control (opening and closing the hand), so his data were not included in the statistics. With the four remaining subjects, there was a 1.22-point improvement in using the haptic feedback (P = 0.27). This improvement was not statistically significant but was above the minimal detectable change.18
SURVEY RESULTS WITH PORTABLE VIBRATORY HAPTIC SYSTEM FOR DAILY USE
The small size of the battery for the prosthetic hand to control the tactors lasted 1 day for subject 1 and about 3 days for the other subjects. All five participants using the haptic system at home found value in the system. Using a 5-point Likert scale (0–4 score) with 4 being the best possible score, the participants rated the haptic system comfort 3.2, correctness of grip feedback 3.8, and ease of operation 3.8. The lower comfort score was because one participant thought the vibration feedback was too strong (distracting). He preferred a less intense vibration sensation. Specific comments on the system included the following:
Research subject 1 commented how he enjoyed shaking hands with his 5-year-old granddaughter while knowing he was only applying a light grip force. Research subject 2 commented that this device “could catapult haptic users into being more valuable employees in the workforce” and “tools such as haptics help prosthetic users become more confident and will help them achieve more.” Research subject 3 commented that if it became a product, he would use it all the time as he enjoyed receiving feedback for all of his activities. Research subject 4 was pleased to receive haptic feedback when his prosthetic hand accidently made contact with the floor while he was working on cars. The feedback alerted him to his unintended arm orientation. However, it should be noted that subject 4 would not routinely use the system because he is very confident with his gripping without haptic feedback. Subject 4 did state that he would have found more value in the system when he was a new user. Research subject 5 did not test the device at home. Research subject 6 felt that the haptic feedback helped him locate medium grip force.
Various types of devices have been used to test vibratory haptic feedback in the laboratory, but this is the first study to our knowledge to test a vibratory haptic feedback system both in the laboratory and in the real world. Unique aspects of this study were the optimization of vibration patterns, enhanced understanding of tactor placement, and the development and testing of a portable haptic feedback system.
The pulse rates chosen were 2 Hz pulse rate for light grip, 4 Hz pulse rate for medium grip, and continuous vibration for strong grip. The subjects easily differentiated these pulse rates. However, while increasing the pulse rate toward a continuous vibration was associated with an increasing grip strength for most, one subject felt that the continuous vibration should be associated with the light grip instead of the strong grip force. This indicated that what is intuitive to most will not be intuitive to all in regard to interpretation of vibration characteristics. Therefore, training with the system or having flexibility in matching vibration characteristics to different grip force categories will be helpful in a commercial device.
The experimentation on tactor placement demonstrated that the proximal location was better for vibration discrimination, although not statistically different. Because each residual limb had different locations of scar tissue and/or less sensitive areas, tactor placement will likely need to remain customizable.
The portable vibratory haptic system functioned as intended. It was able to convey three distinct grip force ranges. Attempting to convey more than three grip forces in future research may be difficult because there is variability in how a prosthetic user grips and the prosthetic glove characteristics that can alter the FSR output for the same grip force. Therefore, more than three grip force ranges would be less reliable with this design. However, using three grip force ranges improved the grip force accuracy for medium grip force in the laboratory for the portable system. Although the improvement was statistically significant, it should be noted that the system improved the grip force accuracy, a great deal in some patients and not at all in others. This indicated that the current system would be beneficial for most but not all prosthetic users. For the subject that thought the vibration sensation was too strong, this could be addressed by lowering the duty cycle of the carrier frequency, by using a smaller ERM tactor motor, or by mounting the tactor differently. Nevertheless, all subjects gave the portable vibratory system favorable reviews when using the system in the real world. This supports the continued progress toward creating a commercially viable vibratory haptic system for upper-limb prosthetic use.
This study was a successful proof of concept study on the development of a portable haptic feedback system using vibration on the arm to convey grip force at the prosthetic hand. It demonstrated that a portable vibratory haptic system could improve grip force accuracy and be used successfully in daily life to perform everyday tasks. Next steps would be to create a more robust prototype with longer battery life. This would allow longer testing in the real world. In addition, a study with a larger sample size would better predict how the general population would feel about the system.
1. Gonzalez J, Soma H, Sekine M, Yu W. Psycho-physiological assessment of a prosthetic hand sensory feedback system based on an auditory display: a preliminary study. J Neuroeng Rehabil
2. Schofield JS, Evans KR, Carey JP, Hebert JS. Applications of sensory feedback in motorized upper extremity prosthesis: a review. Expert Rev Med Devices
3. Saunders I, Vijayakumar S. The role of feed-forward and feedback processes for closed-loop prosthesis control. J Neuroeng Rehabil
4. Meek SG, Jacobsen SC, Goulding PP. Extended physiologic taction: design and evaluation of a proportional force feedback system. J Rehabil Res Dev
5. Patterson PE, Katz JA. Design and evaluation of a sensory feedback system that provides grasping pressure in a myoelectric hand. J Rehabil Res Dev
6. Kim K, Colgate JE. Haptic feedback enhances grip force control of sEMG-controlled prosthetic hands in targeted reinnervation amputees. IEEE Trans Neural Syst Rehabil Eng
7. Chatterjee A, Chaubey P, Martin J, Thakor N. Testing a prosthetic haptic feedback simulator with an interactive force matching task. J Prosthet Orthot
8. Stepp CE, Matsuoka Y. Vibrotactile
sensory substitution for object manipulation: amplitude versus pulse train frequency modulation. IEEE Trans Neural Syst Rehabil Eng
9. Stepp CE, Matsuoka Y. Relative to direct haptic feedback, remote vibrotactile
feedback improves but slows object manipulation. Conf Proc IEEE Eng Med Biol Soc
10. Ninu A, Dosen S, Muceli S, et al. Closed-loop control of grasping with a myoelectric hand prosthesis: which are the relevant feedback variables for force control? IEEE Trans Neural Syst Rehabil Eng
11. Chaubey P, Rosenbaum-Chou T, Daly W, Boone D. Closed-loop vibratory haptic feedback in upper-limb prosthetic users. J Prosthet Orthot
12. Cipriani C, D'Alonzo M, Carrozza MC. A miniature vibrotactile
sensory substitution device for multifingered hand prosthetics. IEEE Trans Biomed Eng
13. Pylatiuk C, Mounier S, Kargov A, et al. Progress in the development of a multifunctional hand prosthesis. Conf Proc IEEE Eng Med Biol Soc
14. Pylatiuk C, Kargov A, Schulz S. Design and evalutation of a low-cost force feedback system for myoelectric prosthetic hands. J Prosthet Orthot
15. Cohen J, Niwa M, Linderman R, et al. A closed-loop tactor frequency control system for vibrotactile
feedback. Portland: ACM CHI 2005; 2005.
16. Hermansson LM, Fisher AG, Bernspang B, Eliasson AC. Assessment of capacity for myoelectric control: a new Rasch-built measure of prosthetic hand control. J Rehabil Med
17. Hermansson LM, Bodin L, Eliasson AC. Intra- and inter-rater reliability of the assessment of capacity for myoelectric control. J Rehabil Med
18. Lindner HY, Langius-Eklof A, Hermansson LM. Test-retest reliability and rater agreements of assessment of capacity for myoelectric control version 2.0. J Rehabil Res Dev