Effectiveness of altered parameters on amplitude modulation-based vibrotactile sensation generation

Introduction: Haptic sensation, commonly referred to as touch sensation, is used by individuals to receive information from their surroundings. Individuals using lower limb prosthetics are unable to receive haptic feedback, as amputation results in severed sensory nerves. A prominent use of haptic sensation generation is to restore sensation for lower limb amputees when walking or standing. While vibrational stimulation is typically used to generate haptic sensation, there is no standard stimulation paradigm. The purpose of this study is to assess the effectiveness of a modified amplitude modulation vibrational paradigm that incorporates a region of no stimulation, referred to as a dead band.

Materials and Methods: The study recruited 6 able-bodied participants. Each participant used a bHaptics Tactsuit vest with embedded vibrational actuators to receive haptic sensations. Stimulation amplitude followed the amplitude modulation (AM) vibration. 3 stimulation paradigms based on AM were combined with differing dead bands (none, small, and large). Training was performed to familiarize participants with sensation from the haptic vest. During all response-based trials, participants were instructed to use the stimulation provided by the haptic vest to complete a 1-dimensional tracing task. For the tracing task, stimulation pathways were generated for the vibrational actuators using MATLAB®. During each trial, the bHaptics vest would activate according to a unique stimulation pathway in a randomized order. Participants were instructed to trace the perceived stimulation pathway using an interface programmed through Unity. Each participant performed 10 response-based training trials and 20 data collection trials per stimulation paradigm. Each trial lasted 20 seconds. Participant accuracy, response time, and smoothness were used to assess effectiveness.

Results, Conclusions, and Discussion: RMSE, lag, and jerk were used to quantify accuracy, response time, and smoothness. Attached boxplots were used to visualize all data, where “Exp” refers to the stimulation paradigm without a dead band, and “DB1" and “DB2” refer to the smaller and larger dead band paradigm, respectively. A Friedman’s Test was used to assess statistical significance between stimulation paradigms, with an alpha value of 0.05 used. A statistically significant relationship (p = 0.094) was found between stimulation paradigm and RMSE, with a small dead band resulting in the most accurate position tracking. This indicates that introducing a region of no stimulation in the stimulation grid provided a reference for participants during experimentation, increasing the amount of information that can be conveyed through vibrational stimulation. Improving positional accuracy in vibrational stimulation models can be used to design better haptic feedback systems for lower limb amputees, both as a system that provides more information to the user about their environment and provides a more intuitive source of feedback. This paper shows an altered vibrational model to improve accuracy from haptic sensation. Although it is a promising start, it is limited to application with lower limb prostheses due to the isolation of using haptic stimulation without receiving input from the user. Further work in assessing the dead band paradigm with more relevant applications can be used to best provide lower-limb amputees with information about their environment.

Acknowledgements (Optional): This work was partially supported by the NSF award 1954587.