Multi-agent Adaptive Shared Control

This chapter addresses the second component in the final framework: adaptive authority allocation in shared control. First, a system-driven adaptive shared control is introduced, which uses unity-sum condition for the gains of teleoperation and autonomous controller. Unlike existing methods, this approach smoothly distributes control authority based on degradation of the real sensor measurements. The results of a user-study are presented for this approach along with the benefits and limitations of the method where the unity-sum condition arbitration was enforced. The next two sections in this chapter present two methods for a general multi-agent robot control system with varying authority using force and stiffness scaling. The factors that destroy passivity are identified and then, model-independent methods are developed to ensure passivity.