Experimental Study of Hand Motion Recognition Method Based on SOA-ELM

To improve the accuracy of human hand motion recognition, a hand motion recognition method based on seeker optimization algorithm and extreme learning machine (SOA-ELM) was proposed, which took the advantage of the principle of SOA to search the global optimal solution and the ability of ELM to deal with nonlinear relation. Firstly, surface electromyography of 4 kinds of hand motions, such as wrist down, wrist up, hand grasps, hand extension were acquired, and the integral value and root mean square value of myoelectricity were extracted from 4 kinds of surface electromyography, which were taken as the feature values. Secondly, the feature values were used to train ELM, and the seeker optimization algorithm was used to find the optimal input layer weight and hidden layer threshold of ELM. Finally, 4 kinds of hand motions were identified by SOA-ELM.Experimental results show that SOA-ELM is more effective than particle swarm optimization-extreme learning machine (PSO-ELM) for recognition of 4 kinds of hand motions.