Dynamic Path Planning of Robot Based on Depth Learning

Multi joint serial industrial robot is widely used in industrial production because of its convenient operation, accurate positioning, flexible execution and other advantages. The working principle of industrial robot is mainly to remember the running track through manual teaching, and use control to make it reach the specified position and pose. Generally, industrial robots only follow one or more fixed routes in actual pipeline work. Dynamic path planning is a method types of robots, such as wheeled mobile robots or underwater robots (UUV). The key idea behind our method is to use the depth map generated by a stereo camera, rather than simply measuring the line of sight distance between two points as traditional methods. The depth map provides information about the relative position between two objects, which is crucial for finding a free path.