Research On Intelligent Algorithm Of Mobile Robot Path Planning In Substation Environment

Substation is an important part of my country's power grid,which plays an indispensable role in the process of voltage conversion,current conversion,transmission and transmission of the circuit.Therefore,ensuring the safety of substations plays a vital part in the livelihood of our society.In this regard,the inspection of substations has become the daily work of electric power workers.However,human-based inspection work is a mechanical and complex work,and is limited by the professional skill level,psychological quality and physical function of the inspectors,which is very detrimental to the safety of the substation.On the other hand,since 2012,the development of artificial intelligence has provided new ideas for the inspection of substations.Intelligent mobile robots can complete the inspection of substations through internal algorithms and pre-entered knowledge maps.A few days ago,more and more robots have begun to work in substation inspection positions.Therefore,it will be an inevitable trend of smart grid for robots to replace manpower to complete the inspection of substations.Among the many aspects of the application of robots in substations,it is a crucial part to quickly respond to emergencies and arrive at the designated location for maintenance.If the accident cannot be handled quickly,it will cause immeasurable losses to the society and economy.To achieve this goal,the robot's path planning has become the top priority.In the face of the complex environment of the substation,the traditional path planning algorithm cannot meet the working requirements of the substation.Therefore,based on the existing path planning algorithms,a path planning algorithm for mobile robots in substations based on deep reinforcement learning is proposed.First of all,considering the complexity of the substation environment,it adopts a 3D modeling strategy based on UE5,and through the analysis of influencing factors,the dynamic components of the substation environment are integrated into the 3D model,which further improves the accuracy of environmental modeling.Then,combined with the characteristics of environmental modeling,a set of triple prediction correction positioning system based on CNN-BNN deep network,Kalman filter and PID control was designed and studied.Then,through the cooperation of the positioning system and the environment modeling,it is integrated into the deep reinforcement learning algorithm,and the parameters such as the Policy Network,reward function,loss function,etc.are designed to make it better to complete the robot's path planning work.Finally,through the simulation test based on UE5,the validity,robustness and accuracy of the robot path planning algorithm in the actual engineering environment are verified.