An Investigation of Control Strategies for Friction Compensation

Abstract

Control strategies are investigated for friction compensation in servomechanisms. As part of the investigation, several different models of friction are reviewed and analyzed for their relevancy to the control problem. Models of friction at zero and near-zero velocities are of particular concern since in this regime friction can introduce oscillations. These different models are considered in friction-compensating adaptive control design. Three friction-compensating adaptive controllers are designed based on strategies proposed in the literature. Adaptive controllers are well-suited to the friction compensation problem since they are nonlinear and have the additional advantage of providing system identification and tracking of slowly-varying parameters such as friction parameters. Stability analyses are performed for the controllers and yield asymptotic stability results for the system error. An original stability proof employing passivity theory is provided for one of the controllers. To test the effectiveness of the adaptive controllers, an experimental program is designed and implemented on a direct drive dc motor. Comparative position trajectory tracking experiments are performed with the three adaptive controllers, a controller with dither ( a commonly-used heuristic friction-compensating controller), and a traditional linear controller used as a benchmark. The results show that the adaptive controllers outperform the more traditional heuristic and linear controllers. Additionally, the experiments yield insight into the appropriateness of the different friction models under the tested operating conditions. In particular, the less popular Dailey model is observed to provide a reliable representation of friction behavior near zero velocity.