Adaptive position following using only position measurement

摘要

A drawback of existing adaptive position tracking algorithms isthe requirement for velocity measurement or velocity calculation, whichin turn requires additional measurement hardware or undesirablenumerical differentiation. In this paper, we propose a new adaptivealgorithm for single-link robots which requires only positionmeasurement and does not need the measurement or computation ofvelocities. A novel technique called phase-lead adaptation is used inthe algorithm, so that the global stability and tracking convergence areestablished based on the Lyapunov method. We proposed an adaptivenonlinear controller which contains adaptive inertial and gravityfeedforwards and an adaptive phase-lead feedback. Adaptive compensationof friction based on an asymmetric Coulomb model is developed. Hardwareimplementation has been carried out in the laboratory, and theexperimental results are excellent. The same control technique can beapplied for general x-y-z type motion systems such as machine tools andpositioning tables