Research Of Gesture Coordination System Based On Neural Network Model

Human beings have been working on machine and equipment that can replace all kinds of work,and the intelligentized degree of the machine is changing with each passing day.As an important part of robot,robot arm has become an important research topic in robotics.Nowadays,there are many arm motion control models,but all of them lack certain cognitive functions.Based on the movement and grasping of arm which is at the neurophysiology and neuroanatomy of cognitive sense,this paper presents and constructs a temporal-spatial coordination neural network brain-inspired model for hand gesture tracking.The specific work is as follows:Firstly,in order to solve the problem of tracking error in the previous cognitive cerebellar model,the concept of fuzzy set is introduced in this paper.By combining the cerebellum model with the fuzzy theory,a fuzzy cerebellar model is proposed for the control of the motion of the manipulator.This model introduces the concept of membership degree of fuzzy set in the input layer of cerebellar model to reflect the objective world more accurately.It can not only control the trajectory of the arm like the biomimetic cerebellar model,but also have higher integrity.Then,in view of the fact that the visual occlusion in the grasping movement will lead to larger grip aperture,this paper is discussed in two cases.One is grasping motion under normal circumstances,and the other is grasping under visual occlusion.On this basis,a gesturecoordination model under visual occlusion is put forward.Through MATLAB simulation,it is proved that visual occlusion will not have a great impact on the arm movement of the new model.The finger preforming part will also avoid unnecessary collisions between hands and objects by increasing the grasp aperture.Finally,a kind of brain model of spatio-temporal coordination neural network for gesture tracking is constructed.Based on the above two models,the model takes full account of the effects of grasping intention on grasping motion,and incorporates fuzzy cerebellar control model into the gesture coordination model under visual occlusion,so that the new model has certain cognitive ability.The comparison with Vilaplana's classical experimental results shows that the new model has significantly improvement in reducing time delay and increasing grasping speed and accuracy.