Research On Mobile Robot Path-planning Based On D*Lite Algorithm

As the core of the mobile robot navigation system,path planning plays an indispensable role in life service navigation,extraterrestrial exploration,unmanned driving,underwater exploration and so on.It solves the problem of "how to go" based on knowing the start position and target position.Path planning experienced the development process from knowning the environment to unknown,simple environment to the complex environment,a small map to the big map,simple artificial intelligence to advanced artificial intelligence.At present,the research of path planning for mobile robots in unknown environment has not yet formed a system.Especially in the large map and complex environment,the traditional artificial intelligence path planning method using high-resolution map representation environment,planning has blindness,complexity and planning time-consuming and other shortcomings.In this paper,how to optimize the environmental map spatial representation and improve the flexibility and efficiency of the planning to reduce the path planning time will be studied.The traditional path planning method in solving the large-map environment path planning,usually grids the large-map to small,the drawback is that in the unknown environment,the times of re-planning will be much more,and as the grid increases,the complexity of planning algorithm increases exponentially and planning is time-consuming seriously.In addition,as for complex environment,the traditional planning method can not make rational judgments according to the environmental characteristics and that leads to its low efficiency.In this paper,the traditional path planning method is analyzed in detail,and the advantages and disadvantages of each algorithm and the specific scenarios are analyzed as well.Finally,a fusion algorithm is proposed for the path planning under the complicated large-scale map in this paper,that is,using the hierarchical map to plan.Refine the planning method to get a collection of a variety of algorithms so that makes every aspect the best performance.As for the global,the fusion method uses D * Lite algorithm which has a excellent performance for the global path planning,and as for the local,the fusion method uses enhanced neural network algorithm to have the ability to autonomously learn environment.In fact,it is reasonable to adopt a variety of algorithm fusion methods to solve the path planning under certain circumstances.It can collect the advantages of a variety of algorithms,for different specific path planning to achieve local optimal,and then to complement the advantages of the algorithm between the results.In this paper,the proposed MRDS simulation platform is used to model the proposed robot,and the fusion algorithm is encapsulated as a service to the mobile robot to guide its path planning.By using the quantitative analysis method,four dimensions include the planning path length,the number of re-planning,the number of updating nodes and the planning time are compared with the traditional D * Lite algorithm path planning.The experimental results show that the fusion algorithm can greatly optimize the planning time and provide a new feasible scheme for road planning under the complex environment of large map.