A Real-Time Assistance Control Strategy for Active Knee Exoskeleton

Climbing stairs under weight-bearing conditions increases the risk of firefighters’ knee pain and even synovial damage. The active knee exoskeleton can help firefighters indirectly lift the load by providing a positive auxiliary torque on the wearer’s knee joint, and reduce the load on the knee patella tendon and quadriceps during climbing stairs. Additional force reduces the risk of knee injury. Most of the existing active knee exoskeletons cannot automatically adapt to the power-assisted mode switching in the process of climbing stairs and walking on the ground, and cannot establish accurate motor control models and dynamic models used in human-robot interaction, resulting in unsatisfactory control results. In this paper, we propose a simple but effective real-time assist control strategy for knee exoskeleton control. This strategy uses only exoskeleton integrated sensors and wirelessly transmitted gait detection insoles, without any extra sensors to capture human motion intentions. It takes the human-robot interaction force as the control target, and establishes a multi-mode switching control system with discrete dynamic variables (such as gait) and continuous dynamic variables (such as joint position and joint speed) of human-robot interaction system, so as to save the complex motor control model and human-robot dynamic model. This makes the auxiliary assistance control strategy of knee joint more practical for applications in the task environment where the ground and stairs are interlaced. A healthy subject participated in this study to test the effectiveness of the algorithm. The experiment demonstrated a significant reduction of metabolic consumption for climbing stairs with a load of 20 kg comparing Knee-Exo On to Knee-Exo Off (3.3%). These results validate the promise of applying the proposed real-time assistance control strategy for knee exoskeleton control aiming at climbing stairs assistance.