Evaluation on interaction ability of a walking robotic suit with synchronization based control

We proposed a framework of using neural oscillators to control a robotic suit which learns to synchronize with the human user´s motion. The robotic suit is a lower limb exoskeleton consisting of four actuators which are fixed to the places where the human hip joints and knee joints are. Each actuator has a built-in harmonic drive gear which is equipped with a torque sensor fixed on the flex-spline. The torque sensor helps to measure the mutual joint torque, which will be generated by once there is any differences between human´s and suit´s movement. The neural oscillator connected to each joint controls each actuator through synchronizing the actuator´s desired joint angle with the feedback mutual joint torque. Proportional-integral-derivative (PID) control is then used to determine the real joint angle of each actuator. We have conducted experiments to show whether: 1) the synchronous movement of robotic suit is realized; 2) the synchronized movement has the application potential for walking assistance and rehabilitation. The results showed the potential of applying the synchronization based control to the robotic suit for walking assistance and rehabilitation purposes.