A Method of Lower Limb Gait Based on Multi-sensor Data Fusion for Rehabilitation Robot

A method of gait movement of lower limbs for a rehabilitation robot is proposed in this article. The gait detection system is established using acceleration sensors, gyroscopes and magnetometers as the attitude detection devices. According to the movement mechanism of human limbs, attitude sensors are respectively installed in the lateral middle position of the left and right thighs and calves, and the collected data is transmitted to a microprocessor through IIC communication. The processing terminal displays the data curve in real time while saving the collected data. Since the acceleration sensors, magnetometers and gyroscopes all have a certain degree of zero drift, ellipsoid fitting which based on least squares method is employed to calibrate, and then quaternion method is adopted to solve the attitude algorithm. The second-order Kalman filtering for attitude compensation is used in this research work. Finally, the simulation and real experimental results proved that the ellipsoid fitting can be well calibrated, in addition, second-order Kalman filtering is capable of performing good attitude fusion. Furthermore, during walking, the characteristic states such as turning and unstable walking can be collected, which is more obvious than the states collected by the three-dimensional gait analyzer.