SUMMARY

Larger workspaces, faster cycle times, and more efficient operation will be the benchmark for the next generation of robots. Unfortunately, these requirements will demand fundamental changes in the way these manipulators are designed and controlled. While reductions in material costs, motor performance requirements, and supporting structures would seem to indicate a significant reduction in cost for lightweight vs. traditional manipulators, achieving the same or better performance metrics requires accurate pose reconstruction and estimation of flexible system dynamics.

The proposed research concerns the development of more robust estimation methods for reconstruction of state variables for feedback control, which can be applied to flexible manipulator applications. More robust estimation methods will permit the use of inexpensive sensors, and make flexible manipulators more attractive for industrial applications.

This requires expanding the theoretical basis for this approach, including a thorough stability analysis of the robust estimators and resulting closed loop control systems. In addition, detailed experiments will be conducted for validation of the approach on physical systems with practical limitations. These experiments include a thorough robustness examination of the sliding mode estimator for a single link manipulator, estimation under combined loading, investigation of sensing strategies, and consideration of multi-body systems with distributed flexibility.

The proposed experiments will utilize both research and commercial hardware testbeds and are sufficiently general to adequately assess the validity of the approach for realistic industrial applications. Expected contributions include the derivation, analysis, and experimental validation of a robust estimation scheme for the control of flexible manipulators.