Design And Implementation Of Indoor Autonomous Positioning And Navigation System For Wheelchair Robot

Since the beginning of this century,the aging degree of China's population has been deepening.The number of disabled people and chronic patients is on the rise,and the demand for smart wheelchairs is increasing year by year.This thesis takes the intelligent wheelchair as the research subject,expands the wheelchair,a single mobile tool,into a mobile service robot.The intelligent wheelchair robot is trained with multi-sensor data,so that it has the ability of accurate environment perception,accurate positioning and navigation in the indoor environment.Compared with outdoor positioning and navigation,indoor environment lacks accurate GPS and other navigation signals,and requires higher positioning accuracy.At the same time,the indoor environment is relatively complex,and there is no traffic indicator line and roadside like those commonly used on the outdoor roads and other reference objects to assist navigation,which also brings great challenges to the design of indoor positioning and navigation system.In this thesis,the environment perception and navigation system of intelligent wheelchair robot is designed from three aspects of environment perception,indoor positioning and path planning.The sensors used are lidar,binocular camera and ultrasonic probe.Aiming at the problem of environment perception,this thesis designs and implements an indoor 3d environment perception system which is suitable for low-cost sensors by using the multi-sensor fusion method.The wheelchairs acquire the point cloud image from the fixed direction lidar installed in the front of the car body,and integrate with the motion parameters of the car body,so as to conduct high-precision reconstruction of the 3d environment around the car body that can be detected by the laser.Ultrasonic probe is used to construct the unmeasurable laser environment around the vehicle body.A binocular camera is also added to the system for feature recognition.The combination of the three forms a reliable perception of the surrounding environment.In this thesis,static map and dynamic map are combined to solve the problem of autonomous positioning of the wheelchair robot.Taking the entrance of the room as the origin of the environment coordinate system,the static environment map is constructed by occupying grid method.The dynamic map is constructed with the wheelbase of the wheelset as the origin.Firstly,the wheelchair is matched with the template image by using the image collected at the entrance,and then the room where the wheelchair is currently located is determined by combining the matching points,error rate and other parameters.Then,the position of the indoor marker is identified,and the marker is used as the reference to calculate and correct the position of the wheelchair in the current room,and the target orientation is confirmed by the linear velocity and angular velocity of the wheelchair itself.While completing indoor path planning of wheelchair robot,A star algorithm is optimized in this thesis to increase the weight of turning time.According to the investigation of the parameters such as the number of turns and the length of the path and the time used,it is verified that this optimization is effective in reducing the walking distance and turning angle of the intelligent wheelchair robot.In the selection of obstacle avoidance strategy,the obstacle scene is classified and predicted.According to the distance between the obstacle and the wheelchair,the obstacle avoidance area and the danger area are divided to shorten its travel time without touching obstacles.In the experimental environment,the wheelchair robot designed in this thesis can complete the indoor autonomous positioning and navigation.After many experiments,it can achieve the experimental effect similar to that of the wheelchair assisted by others.