Research On Mobile Robot's Path Planning Technology In Complex Environments

With the rapid development of technologies such as artificial intelligence,big data and sensors,some artificial intelligence experts and investors in China predict that autonomous vehicles will be the most suitable technical product in this round of intelligent wave.Autonomous vehicles,once truly commercialized,will have obvious benefits in terms of reducing the incidence of traffic accidents,improving transportation efficiency and saving labor costs etc.Automatic driving technology can be further divided into four dimensions:low-speed loading,high-speed loading,low-speed manned and high-speed manned vehicles.Before the commercialization of the real high-speed manned autonomous vehicles,autonomous vehicles with low-speed loading is more realistic.Some domestic and foreign logistics-related enterprises have tried this and are constantly iterating research and development.Automatic driving technology is a subset of mobile robot technology.Mobile robot technology includes many aspects,such as environment perception,autonomous localization and path planning etc.Aiming at the problem of low safety and stability of mobile robot and low flexibility of obstacle avoidance,regarding the vehicles as the research carrieres,this paper proposes a new framework of path planning system,which designs a set of path planning system that can be used in complex environment many aspects of existing algorithms to improve the safety,stability and flexibility of the entire robot system.The current path planning system consists of three parts:the global path planning subsystem,the local path planning subsystem and the trajectory tracking control subsystem connected in turn.The entire path planning system is designed to meet as many other criteria as possible in the case of security,such as short distances and low energy consumption.Global path planning generally can not do real-time path adjustment.Realtime obstacle avoidance depends on the local path planning subsystem to complete,so once the local path planning subsystem fails,there will be a big collision risk.In this paper,based on the classical path planning framework the collision detector is added to avoid the collision in case of emergency.Thus,forms the double protection function of the online local path planning subsystem and the collision detector.While fully improving safety,it will not affect the real-time performance of the system.The main achievements of the paper are summarized as follows:(1)For the global path planning subsystem,by analyzing the advantages and disadvantages of the classical path planning algorithm,combined with the actual application scenarios,use the A*algorithm as a global path planning algorithm,combine with the characteristics of the circle within the cycloid and smooth the planned path,which increases the smoothness and safety of the path without reducing the planning efficiency and facilitate the effective tracking of the global path in the local path planning subsystem.(2)For the local path planning subsystem,taking into account the practical application scenarios,the hybrid A*global path planning algorithm is proposed as the local path planning algorithm,thus ensuring a small number of adjustment parameters,which have low computational resource consumption and can escape local minima and other advantages,to avoid such issues missing turning time due to the untimely adjustment of speed.At the same time,the Reeds-Shepp path curve is applied to improve the search efficiency,and the gradient descent method is used to smooth the trajectory.(3)For the trajectory tracking control subsystem,the classic Pure-Pursuit method is combined with the Stanley method,which makes the trajectory tracer adapt to the changes of different trajectories,and has strong anti-interference and universality.The collision detection algorithm is designed.When the anti-collision function in the upper layer fails,it can still respond quickly to the dangerous obstacles in the vicinity of the robot,thus effectively avoiding the accidental occurrence.(4)The software simulation platform of the autonomous navigation mobile robot based on automobile model and the hardware physical platform of four-wheel differential ground mobile robot are built,and the simulation and the verification of the designed path planning system are carried out.The experimental results show that the designed path planning system has good security,stability and flexibility.