Robot Manipulator Adaptive Motion Control With Uncertain Model And Payload

Robot plays more and more important role in our industrial production and daily life. As an important part of the robot actuator system, the robot manipulator is a highly non-linear, coupling, and time-varying complicated system with expanding of the robot application area. So there are many uncertainties in the robot dynamics model. In order to improve the intelligence of the robot manipulator when facing these uncertainties, in this paper we use function approximation techniques(FAT) to estimate the unknown matrices in the dynamic model and the “one-step guess” idea to approximate the unknown payload, then design the impedance adaptive controller and discrete-time adaptive controller when there are some uncertainties in the model or payload.The main contributions are highlighted as below:(1)For the uncertainties in the dynamic model of robot manipulator,we use the FAT techniques to estimate the unknown matrices, then design the impedance adaptive controller and demonstrated the stability of the system. The tracking experiments of iCub robot manipulators show the practical feasibility of this algorithm.(2)For the payload uncertainties of the robot manipulator,we approximat the unknown payload based on the “one-step guess” idea and design the discrete-time adaptive controller. In the simulation experiments of two degree-of-freedom robot and six degree-offreedom robot PUMA560, the results indicate that this algorithm have a good estimate effect and tracking performance.The strategies proposed in this paper may provide some guidelines to improve the intelligence of the robot manipulator in practice.