Research And Implementation Of Self-driving Along The Road For Autonomous Wheelchairs

According to the demographic data of our country in recent years,in 2018,China’s total population over 65 years old accounted for 11.94%.The daily travel of these elderly people has a greater dependence on wheelchairs,which has made society’s demand for wheelchairs more and more.The intelligence level of wheelchairs is getting higher and higher.Based on the traditional electric wheelchair,this article optimizes the design of the control system structure of the wheelchair,and uses technologies such as multi-sensor data fusion to realize that the wheelchair can automatically drive along the roadside in outdoor environments,which greatly reduces the difficulty of control and work intensity of wheelchair users outdoors.Because the shape of the roadside is not an accurate value,the height,width,and shape of the roadside feature will vary greatly in different scene.In addition,obstacles will block the roadside feature.This brings great difficulties to data collection and roadside identification.This article designs and implements an intelligent wheelchair automatic driving along the road from three aspects: hardware system,navigation along a regular roadside,and navigation under complex road conditions.The hardware system of this paper adopts a modular design,which is divided into a main control module,a lidar module,and a ranging module.The main control module is responsible for regulating and controlling the overall resources of the system;the laser radar is installed in front of the wheelchair to detect the information about the road surface and obstacles in front;the ranging module is available in four directions of the wheelchair,and is mainly responsible for detecting the distance information between the wheelchair and the environment.The three modules cooperate with each other to realize the perception and processing of surrounding environment information.Automated driving along a regular road is achieved by combining lidar with an improved artificial potential field method.This paper establishes three coordinate systems of navigation,sensors,and wheelchairs,and integrates the data obtained by sensors.The point cloud data is preprocessed by filtering,screening,etc.,and the data is mapped to a twodimensional plane using arc modeling,and the roadside features are identified from the point cloud map by combining the horizontal index and radial index.Finally,the RANSAC algorithm is used for fitting to obtain the exact position along the road.Introducing the roadside force field into the artificial potential field method makes it more satisfy the scene requirements,and realizes the automatic driving along the roadside when the roadside features are obvious.Autonomous driving under complex road conditions uses fuzzy neural network algorithms for navigation.Due to the uncontrollable external environment,the sensor system cannot obtain the exact information of the edge.In this paper,a method based on fuzzy neural network is used to drive along the road.The input contains the distance data of the two sides of the wheelchair to the roadside and the changing trend.The output of the fuzzy neural network is used to give control instructions for the movement of the wheelchair to achieve automatic driving along the roadside.After many tests in the campus environment,the roadside automatic driving system designed in this article can successfully control the wheelchair along the roadway in a variety of environments.Basically achieve the controlled driving effect of wheelchair users.