Enhancing Constraint-Based Robot Task Specification with Dual Controller and Estimator Synthesis

Task specification for robot manipulators continues to be a time consuming and expensive process, limiting their rapid deployment for industrial applications. This is especially true for the automation of sensor-based tasks, where the robot should adapt to uncertainties and disturbances in the environment. Previous work showed how robot task specification can be eased by using the task function approach to synthesize a robot controller. The main novelty of this paper is to further ease task specification by also using the task function approach to synthesize an estimator, which can be used to estimate geometric position uncertainties and the disturbances acting on them. Furthermore, as a result of the duality between control and estimation, an existing constraint-based robot task specification framework can be reused to perform both control and estimation without modification. The approach is validated on the well-known force-controlled contour following task. It is shown that, as expected, when disturbances are estimated and compensated for, the force tracking errors can be substantially decreased, while the contour following speed can be substantially increased. The presented approach can be used to enable rapid deployment of sensor-based applications while significantly improving the control performance compared to current industrial practice.