Capability Iteration Network for Robot Path Planning

Abstract

Path planning is an important topic in robotics. Recently, value iteration based deep learning models have achieved good performance such as Value Iteration Network(VIN). However, previous methods suffer from slow convergence and low accuracy on large maps, hence restricted in path planning for agents with complex kinematics such as legged robots. Therefore, we propose a new value iteration based path planning method called Capability Iteration Network(CIN). CIN utilizes sparse reward maps and encodes the capability of the agent with state-action transition probability, rather than a convolution kernel in previous models. Furthermore, two training methods including end-toend training and training capability module alone are proposed, both of which speed up convergence greatly. Several path planning experiments in various scenarios, including on 2D, 3D grid world and real robot with different map sizes are conducted. The results demonstrate that CIN has higher accuracy, faster convergence and lower sensitivity to random seed compared to previous VI-based models, hence more applicable for real robot path planning.

Research on automatic path planning for obstacle avoidance of intelligent substation patrol robot

“…It should be noted that the size, shape and scale of obstacles need not be used. The greater the difference, the more credible the effect of obstacle avoidance path planning [15].Next, on this basis, we will summarize and integrate the associated data and path planning information of the patrol robots in each area, classify and process the data information, convert the data into the form of data packets through a unified format, and transmit it to the corresponding location according to the sending end of the obstacle notification signal, so as to complete the collection of basic data information for subsequent use and comparative analysis [16].Then, the obstacle avoidance operator needs to be set according to the path range of the patrol robot. When avoiding obstacles, the multi-functional intelligent mobile patrol robot needs to identify the obstacles in front of the face in order and determine the safe distance to bypass.…”