Recently, the Shenyang Institute of automation of the Chinese Academy of Sciences and the Edinburgh robotics center of the United Kingdom have made new progress in cooperative research, proposed an autonomous operation method of mobile manipulator based on deep reinforcement learning in dynamic and unstructured environment, and successfully applied the latest artificial intelligence learning theory to the control of real complex mobile manipulator. Relevant research results are published in the journal sensors.
It is a complex task for robots to work in a large number of dynamic and unstructured environments such as space, land and underwater. Compared with traditional industrial robots, the operation requirements are higher. Generally, robots need to have a variety of functions such as perception, navigation, decision-making, operation and so on.
Shenyang Institute of automation and the research team of Edinburgh robot center jointly use neural network to construct an overall reinforcement learning control model of robot. The deep learning method is used to process the environment and target information obtained by the robot camera, and then take the perceptual information and the current state of the robot man as the system input to independently control the overall behavior of the robot. Through interactive learning and training in simulation and actual environment, the autonomous operation of mobile manipulator in real environment is finally realized, which lays a foundation for the application of deep reinforcement learning to more complex underwater floating base robot system.
Under the strategic cooperation framework (MOU) between the State Key Laboratory of robotics of Shenyang Institute of automation and Edinburgh robotics center, the two sides actively carry out personnel exchange, academic exchange and joint training of postgraduates, and cooperate in many research fields, including robot control based on deep reinforcement learning, underwater 3D scene reconstruction and target recognition based on vision Autonomous operation of underwater robot, etc. With the deepening and strengthening of cooperation, more scientific research achievements will be continuously produced in the follow-up to promote the common progress of relevant scientific research of both sides.