A General Framework for Control Systems Approach to Collaborative Human-Machine Systems

Abstract

This paper presents a new work for the effective management of human-machine collaboration by applying a model-based control systems approach. One fundamental idea is that contemporary machines are assumed to be equipped with human assistance systems (HAS) with varying levels of intelligences. Thus, machines are collaboratively managed by both the human and HAS. This view goes beyond the full automation paradigm which is considered as a special case and is more general than the human-machine interaction paradigm. Another fundamental idea is to take a control systems view of collaborative human-machine systems, especially the so-called model-based control systems paradigm. The HAS technology is developed based on these two views. This paper in particular describes details of these two views along with a preliminary development of a HAS system for management of human cognitive engagement through regulating a multi-modal interface in a car-following case study. HAS revoked drivers’ cognition engagement for decision making first and retained the right of action interception as last. This approach demonstrated reduced driver intrusion while maintaining similar driving safety. In general, this approach minimized the deviation between the required human cognitive interference with the human state and the actual level of the interference based on the machine state.