Human-machine inclusiveness in the process of lower limb rehabilitation is a key problem in interaction

and aiming at the force mutation and tracking trajectory precision problems existing in rehabilitation mechanism

an admittance-model predictive control(A-MPC)two-layer control strategy composed of the outer loop of the admittance controller and the inner loop of the model predictive control(MPC) controller is proposed.Force/position hybrid control strategy is adopted

and the variable parameter design of the admittance controller is carried out based on the Sigmoid function in the outer loop

which transforms the human-machine interaction torque into the expected trajectory of joint angular displacement correction.The inner loop is an MPC controller considering constraints

which can quickly and in real time adjust the trajectory with a known dynamic model of the rehabilitation mechanism and multiple constraints.Through the construction of the simulation model and the actual machine operation test

the torque mutation under the control strategy of A-MPC decreases by 41.69% on average

and the tracking precision is increased by 133.70% on average.This verifies that the designed strategy effectively improves the flexibility of the rehabilitation process and has high trajectory tracking precision.