Research On Path Planning Algorithm For Autonomous Mobile Robot Based On Optical Nod

Currently,path planning technology for autonomous mobile robots is relatively mature and yields better results in wide scenes.However,there are some challenges in narrow spaces such as increased time consumption,longer paths,poor smoothness,and even path planning failures.Furthermore,narrow environments like overtaking on road sections and cargo transportation require faster path planning.Therefore,there is a need for more efficient path planning in such scenarios.In summary,improving the path planning effectiveness of autonomous mobile robots in narrow spaces is necessary.In this study,a new node model,called the light source node model,is proposed based on the existing mainstream path planning algorithm that relies on sampling.This model incorporates the characteristics of narrow spaces by adding attributes of light sources in the real physical environment to the traditional node model.It narrows down the sampling area to reduce the number of node samples and enhance node sampling efficiency.More precisely,since any point within the illumination range of a node can be connected to that node,there is no need to sample nodes within the illumination range in the future.In other words,this approach deviates from the traditional method of random sampling in free space and restricts the sampling area to unlit areas,minimizing the chances of invalid sampling.As a result,it significantly improves the efficiency of path planning for mobile robots in narrow spaces.The effectiveness of this approach is verified by integrating the light source node model with the traditional mainstream path planning algorithm.The specific improvement methods and simulation results are as follows: Firstly,the traditional node model of the probabilistic roadmap algorithm is optimized to the light source node model,and the sampling method is simultaneously improved.The simulation result indicates that the improved algorithm reduces the time required for path planning in narrow spaces,improves smoothness,and maintains a similar path length.Secondly,the improved artificial potential field method is combined with the light source node model and the traditional deep Q network algorithm for path planning.On the one hand,this combination incorporates the repulsive field nature of the artificial potential field method to enhance the agent’s ability to avoid obstacles,thus reducing collisions.On the other hand,the nature of the target point light source node is utilized,allowing the agent to reach the target point directly along a straight line without further exploration and training,thereby improving training efficiency.The simulation result indicates that the improved algorithm expands the application scope of the algorithm,making it suitable for path planning in narrow spaces.Finally,experimental verification of the probabilistic roadmap algorithm optimization algorithm based on optical nodes is conducted.A two-wheeled differential mobile robot model is designed and created using Solid Works software,and a physical platform is built for conducting physical experiments.The experimental results align with the simulation results,indicating that the optimized probabilistic roadmap algorithm yields better optimization effects in both narrow and ordinary space path planning.In conclusion,this study proposes a light source node model for path planning in narrow spaces and integrates it with traditional algorithms for improvement.The research results indicate that the light source node model optimizes the path planning effectiveness from various aspects.When combined with the probabilistic roadmap algorithm,the algorithm using the optical node model reduces path planning time for mobile robots in narrow spaces by approximately 50% and improves path smoothness by around 10%,resulting in better path planning outcomes.Moreover,the combination of the light source node model and the traditional deep Q network path planning algorithm expands the applicability of traditional algorithms and lays the foundation for further algorithmic optimization in the future.