Multi-inspection-robot collaborative path planning method and system based on DQN model

• Main Authors: ZHANG RUI, QI MANSHUN, CHEN HAO, ZHANG HAIHUA, FANG GUOQUAN, CAI BIAO, HAN XIANGZHENG, QIAN QILONG
• Format: Patent
• Language: Chinese; English
• Published: 04.07.2023
• Subjects: CONTROLLING; PHYSICS; REGULATING; SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES

The invention discloses a multi-inspection-robot cooperative path planning method and system based on a DQN model. The method comprises the steps that position coordinates of all inspection robots and arrival states of all task points are obtained, and a multi-inspection-robot cooperative state space is constructed; acquiring moving directions selected by all the inspection robots to construct a cooperative action space of the multiple inspection robots; classifying the states by taking an anti-collision mechanism triggered between the inspection robots and the obstacles and an anti-collision mechanism triggered between the inspection robots as constraint conditions, and defining reward values corresponding to various states; according to the state, the action and the reward value, the DQN model calculates the expectation of a return value obtained after the set action is executed in the set state, the expectation is optimized through deep neural network parameter training, and the state and the action corres