Research On ROS-based Indoor Intelligent Wheelchair System

According to statistics,in recent years,the proportion of the global elderly population and the disabled population has been increasing,and people with disabilities have become a large group of society.Wheelchair is a common auxiliary equipment for people with mobility disorders to carry out daily activities.As a new type of wheelchair,intelligent wheelchair has gradually stepped onto the historical stage.Intelligent wheelchair integrates many related technologies in the field of robotics: Simultaneous Localization and Mapping(SLAM),path planning,machine vision,pattern recognition,motion control,multi-sensor information fusion and humancomputer interaction.This paper mainly studies SLAM algorithm and path planning algorithm in intelligent wheelchair system.The problems of the traditional location algorithm and the global path planning algorithm are discussed,and the corresponding improved algorithm is proposed.Specific research work includes:(1)Based on Rao-Blackwellised Particle Filter(RBPF)SLAM algorithm,location and mapping are decomposed into two relatively independent processes.Firstly,the motion model and observation model of intelligent wheelchair are constructed,and various description methods of environmental map are compared and selected.Aiming at the problem that the traditional Monte Carlo localization algorithm only uses a single sensor to estimate the wheelchair position and pose,an improved Monte Carlo localization algorithm is proposed.The measurement data of odometer and Inertial Measurement Unit(IMU)are fused by unscented Kalman filter to reduce the positioning error.In addition,in order to improve the convergence speed,the adaptive mechanism is introduced into the resampling process of particle swarm optimization(PSO).(2)In global path planning,traditional A* algorithm has many useless inflection points,and the path planned in multi-U obstacle indoor environment is too close to obstacles.To overcome these two shortcomings,an improved A* algorithm is proposed.The concept of neighborhood search matrix is introduced into the process of obstacle search,so that the planned path can be extended a certain safe distance from obstacles.The algorithm also optimizes the path,eliminates the useless inflection point,and overcomes the problem that the path planned by the traditional algorithm can only extend at a fixed angle.In addition,DWA local path planning algorithm and comprehensive path planning strategy are also studied.At the same time,in order to enable the self-construction of the wheelchair when it is in an unfamiliar environment,an autonomous exploration algorithm based on edge detection is integrated into the system.(3)An intelligent wheelchair system was built based on the Robot Operating System(ROS)and experimentally verified in the actual environment.In order to facilitate the maintenance of the system,the hardware platform and software platform of the system are constructed with modular ideas.The software system for intelligent wheelchairs is built by ROS.The system implements communication between processes in the form of subscription and publication,which greatly reduces the coupling degree of the algorithm module.It can flexibly and efficiently organize various algorithm modules to improve system reusability and portability.The Arduino MCU is used as the core control module of the hardware system,and the host computer and slave computer are interconnected through serial communication.The lower computer obtains the corresponding PWM value by solving the desired speed obtained by the path planning on the host computer,thereby realizing the control of the wheelchair motor.In addition,the system uses the Android APP for human-computer interaction,enabling users to more easily and quickly control the wheelchair and view maps.