Development of An Adaptive Iterative Learning Controller With Sensorless Force Estimator for The Hip-type Exoskeleton

Nowadays industrial loads lifting still heavily relies on manual operation. Long-term lifting job can hugely increase low back strains and even cause lumbar diseases. With the rapid development of exoskeleton systems, it is essential to develop a hip-type exoskeleton for the porters. This paper proposes a modified adaptive iterative learning-based control algorithm with a sensorless force estimator. The advantage of this algorithm is that the exoskeleton joint can move precisely along the desired trajectory as well as generate an external assistance torque based on the force estimator while the wearer lifting the load. A hip-type exoskeleton device has been developed to evaluate the effectiveness of this algorithm. We carried out experiment works of bending and lifting heavy loads. Experimental results show that when choosing the IEMG of LES as the evaluation criteria, our exoskeleton device with the proposed control algorithm can significantly reduce the strain force of the wearer’s lumbar over 40% when they lift heavy loads. Thus it can improve the work efficiency of wearers and reduce their risk of lumbar diseases.