Research On Brain-controlled Intelligent Wheelchair Robot Based On RO

With the deepening of the global aging society,the number of elderly people and disabled people who have lost their ability to move is increasing,and the demand for intelligent wheelchair robots is increasing.Among them,brain-controlled intelligent wheelchair robots integrate brain-computer interface,control science and human-computer interaction,and other technologies,which have become the research focus for many scientific research institutions and workers.In this paper,the brain-controlled intelligent wheelchair robot system based on the Robot Operating System(ROS)is studied,and the acquisition,preprocessing,feature extraction,and classification recognition of EEG signals are completed,the control mode of the intelligent wheelchair robot is optimized,the autonomous navigation and path planning of the intelligent wheelchair robot were completed,and the universal application of the brain-controlled intelligent wheelchair robot in multiple scenarios was realized.The main work contents and innovations are as follows:1.Aiming at the problem of poor accuracy of EEG signal acquisition,a steady-state visual stimulation paradigm and stimulation interface were designed.The generation mechanism,signal characteristics,and factors inducing the formation of steady-state visual evoked potentials were comprehensively analyzed,and two acquisition channels O1 and O2 were selected.The experimental method and stimulation paradigm of EEG signal acquisition is designed,and the steady-state visual stimulation interface is designed independently by using Psychtoolbox,which realizes the efficient acquisition of EEG signals.2.Aiming at the problem of EEG signal interference,an EEG signal processing algorithm based on ICA-PSD-SVM is proposed.Firstly,in order to solve the pseudo-signal such as ocular electrical signals doped in EEG signals,the method principle of Independent Component Correlation Algorithm(ICA)was studied,and the EEG signals were separated blindly by experiments,and pure EEG signals were obtained.Then,on the basis of the ICA processing of EEG signals,the Power Spectral Density(PSD)and Support Vector Machine(SVM)are fused,the ICA-PSD-SVM EEG signal processing algorithm is proposed,and the classification and recognition experiments of EEG signals are carried out by applying the algorithm,which improves the accuracy of classification recognition.3.For the single problem of wheelchair control,a genetic-A*-D comprehensive path planning algorithm is proposed.Firstly,the principle of the A* algorithm and the implementation process of path planning are analyzed,aiming at the problem of large search space and long operation time of the A* algorithm,the A* algorithm is improved by integrating the genetic algorithm,and the A* algorithm is experimentally compared with the A* algorithm in the same environment,which reduces the amount of time and space operation and optimizes the global path planning.Then,aiming at the problem of poor dynamic obstacle avoidance effect of the A* algorithm,the DWA algorithm is integrated to optimize the local path planning.Finally,the genetic-A*-D comprehensive path planning algorithm is experimentally analyzed,the feasibility of the algorithm is verified,and good local dynamic obstacle avoidance and intelligent global path planning are realized.4.An experimental platform for intelligent wheelchairs was built and verified experiments.The overall scheme of the intelligent wheelchair system is designed,with the ITX industrial computer equipped with Core i5 processor as the upper computer,NXP minimum system board as the lower computer,and drawing the main control drive integrated PCB circuit board,completing the electric wheelchair modification,and conducting verification experiments,the experimental results show that the proposed method has good practicability and feasibility.