Optimizing prosthesis design to maximize user satisfaction using a tethered robotic ankle-foot prosthesis

Lower-limb prostheses are designed based on observations of how users on average respond to different design features. Prostheses are then marketed on the basis that certain features are appropriate for certain types of individuals, with few options for user customization. This process is unlikely to provide individual users with devices that best suit their needs since it is unclear how to best categorize users and which design features are most important [1, 2]. This process is also slow to accommodate disruptive technologies since it requires time to develop a body of observations about new devices that practitioners are hesitant to prescribe. Prosthesis designs could instead be optimized for individual users, producing customized designs that are likely to be preferable to off-theshelf designs. Using a traditional design approach, this would require costly rapid prototyping and evaluation of candidate designs. Instead, we demonstrate the use of a tethered robotic ankle-foot prosthesis [3] as a tool for rapid exploration of candidate designs. We devised a strategy for systematically exploring a space of possible device behaviors and identifying which are user preferred. The resultant optimized designs could then be sent to prosthesis manufacturers for physical implementation using traditional processes.