A preliminary study of the efficacy of bimodal versus unimodal stumble recovery responses in a powered knee prosthesis

Abstract Transfemoral prosthesis users demonstrate a higher fall rate due to tripping than able-bodied controls in previous laboratory studies. In particular, early swing demonstrates the greatest disparity, where able-bodied controls typically utilize an elevating strategy to cross the obstacle in the same stride that the perturbation occurs, rather than the lowering strategy, where swing is ended prematurely and the obstacle is crossed in the following stride. However, due to the passive nature of most commercial knee prostheses, the elevating strategy is largely inaccessible to prosthesis users, potentially contributing to the increased fall rate in early swing. To investigate the effects of reintroducing the elevating strategy to transfemoral prosthesis users, a bimodal stumble recovery controller was developed for a powered knee prosthesis that utilized the elevating and lowering recovery strategies, selected based on the post-impact kinematics of the prosthesis. The Bimodal controller was compared to a unimodal controller that only used the lowering strategy. Three transfemoral prosthesis users underwent a series of treadmill-based obstacle perturbations with each controller following an acclimation period. All participants successfully used the elevating response in the early swing phase. On average, the elevating response reduced the disturbance to participants’ trunk kinematics and the reliance on harness support. While the Bimodal controller sometimes resulted in a recovery strategy mismatch for two participants, the mismatch still resulted in outcome metrics comparable to the unimodal controller. Overall, results suggest that the inclusion of the elevating and lowering strategies may improve stumble recovery outcomes for some transfemoral prosthesis users.