Application of machine learning towards design optimisation of bio-inspired transfemoral prosthetic socket for robotic leg test rig

Sabau, Panashe; Jie Chong, Jun; Jafari, Aghil; Agrawal, Subham; Semasinghe, Chathura; Etoundi, Appolinaire

Abstract

In the past century many medical advancements in prosthetics have been achieved, however, discomfort in prosthetic socket remains one of the toughest challenges faced by both amputees and prosthetists. Wearing an uncomfortable socket can lead to users discontinuing use of their socket and subsequently reducing their long-term mobility; negatively impact their psychological health; and prolong rehabilitation. This paper continues the research conducted in earlier publications [1], [2], which introduced the concept of an automated ISO standard robotic testing rig to test a full artificial limb prosthesis (a bio-inspired transfemoral prosthetic socket attached to robotic prosthetic joints and an ankle joint). This paper presents an automated method of designing the bio-inspired socket using artificial intelligence to reduce discomfort and the design time of new or existing full artificial lower limbs using qualitative and quantitative data. The socket will be tested in a gait simulation shown in the figure 7, to safely achieve desirable walking velocities, step length, safety and comfort while consequentially reducing the physical testing on patients and consequentially reduce physical testing on patients.

Citation

Sabau, P., Jie Chong, J., Jafari, A., Agrawal, S., Semasinghe, C., & Etoundi, A. (in press). Application of machine learning towards design optimisation of bio-inspired transfemoral prosthetic socket for robotic leg test rig

• Conference Name: 2020 20th International Conference on Control, Automation and Systems (ICCAS 2020)
• Conference Location: Busan, Korea
• Start Date: Oct 13, 2020
• End Date: Oct 16, 2020
• Acceptance Date: Aug 19, 2020
• Deposit Date: Sep 16, 2020
• Keywords: Biomedical engineering, Prosthetics, Robotics and Automation
• Public URL: https://uwe-repository.worktribe.com/output/6690249