Research And Discussion For Hand Tracking Interaction Based On Cognitive Theory

With the improvement of information technology,the research of cognitive theory has been a breakthrough development,and cognitive model has also been used wildly.The conventional control mode is no longer applicable for the development of society and the demand of bionic machinery intelligent degree increasing,and the new control concept is needed to carry on the technical innovation.At present,it is a research topic in the fields that combined with the research results of neurophysiology and neuroanatomy to study and replicate cerebellum sensory motion control system.This thesis aims to resolve the problems in the human race reaching and grasping under perturbations of object size,distance and orientation,and discusses how to reduce the disturbances as well as correcting real-time errors of the motion trajectory in movement.Firstly,this thesis aims to solve how to minimize the numbers of three main kinds of disturbed components to least: hand/wrist transport,grip aperture and hand orientation to the problems discussed above in a vector integration to endpoint(VITE)model.According to the characteristics of arm motion control model and based on the achievements of cognitive psychology,this thesis proposes a new divide-and-conquer model that is used for indicating current grasping status and triggering three main kinds of methods of when to start or stop working.When coupled with the VITE model,the resulting mode simulates the dynamic characteristic of human reaching and grasping more closely,and provides more intelligence when building hand/wrist transport models and hand orientation control simulation modules with mutually independent visual channels.An analysis of simulation results shows that the proposed model is advantageous for its better performance on perturbation rejection.Lastly,an abstract feedback correction module is proposed based on the discussions of cerebellum's functions in motion and how to correct real-time errors of trajectory when reaching and grasping as well as how the cerebellum taking part in generating trajectories under the control of VITE,and then expands the T-VITE model with the proposed module.Finally,the system is simulated by MATLAB/Simulink to assess the learning ability of the cerebellar neural network.The experimental consequences demonstrate that the expanded model has better performance on error correction in real time,it is instantaneity and accuracy in arm movement control,and is closer to the expected result.