Research On Autonomous Navigation Method For Outdoor Unmanned Vehicle

This paper introduces an autonomous navigation system for an ackermann wheeled mobile platform under unrehearsed terrain.Based on a binocular stereo camera and a single-line lidar range finder,environment is reliably explored and precisely modelled.This model is then used to relocalize the vehicle and helps performing global path planning to achieve a safe point-to-point autonomous navigation.The novel characteristic of this paper is that,mainly based on stereo vision,the system is capable of achieving core functions of a typical navigation system,such as environment sensing,localization,path planning and obstacle avoidance.The main research results include:1.A method of constructing a metric-topological mixed map from sparse visual map is proposed.2.A method of tracking and real-time obstacle avoidance based on the above-mentioned map is proposed.3.An unmanned navigation system is developed,integrating sensing,planning and obstacle avoidance.