Neural-network-based control esp. for multi-degree of freedom robot three-arm manipulator - represents geometric characteristics of arms by model, and calculates full set of state variables e.g. arm angles and both end-point coordinates of manipulator to obtain best solution

摘要

Control of the system is based upon a neural network processing structure. The network is defined in terms of neurons, which for a specific robot structure will be assigned to each variable. For example, the angular position of one movable element of the robot will define a parameter set against a spatial coordinate system. The outputs of each neuron (1-17) are fed back as inputs to form a 'backwards' propagation process. Various operating strategies can be evaluated to determine the redundancies in the system. This allows the most efficient operational cycle to be implemented. USE/ADVANTAGE - Provides efficient control and use of multi-degree of freedom robot.