A Robust Human-Autonomy Collaboration Framework With Experimental Validation

Abstract

In this letter, we introduce a robust human-autonomy collaboration framework focusing on flight control applications. The objective is to optimize performance by always keeping the human operator in control of the vehicle while compensating for human limitations. A significant aspect of this framework is its robustness to human intent estimation errors. This is achieved by precisely modulating the automation assistance to prevent undesired interference with the human operator. We provide human-in-the-loop experimental results, demonstrating significant performance improvements when intent estimation is accurate. Experiments also validate that the pilots maintain vehicle control even when the estimation is faulty.