Synthesis With Guaranteed Cost and Less Human Intervention for Human-in-the-Loop Control Systems

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Stochastic Systems

Background:

• Human-in-the-loop (HiTL) control systems require modeling complex human behaviors.
• Integrating human and machine control while minimizing human intervention is challenging.
• Existing models often struggle with human state uncertainty and control input dynamics.

Purpose of the Study:

• To develop a novel framework for synthesis with guaranteed cost and reduced human intervention in linear HiTL systems.
• To accurately model human behavior using a hidden controlled Markov process.
• To construct a hidden controlled Markov jump system (HCMJS) integrating human and machine models.

Main Methods:

• Modeling human behavior with a hidden controlled Markov process, accounting for stochasticity and uncertainty.
• Constructing a hidden controlled Markov jump system (HCMJS) for integrated human-machine interaction.
• Utilizing stochastic Lyapunov functional and bilinear matrix inequality for controller existence conditions.
• Employing a hybrid particle swarm optimization and linear matrix inequality algorithm for control law synthesis.

Main Results:

• A sufficient condition for the existence of human-assistance controllers is derived.
• The proposed method guarantees stochastic stability of the closed-loop HiTL system.
• A prescribed upper bound for the quadratic cost function is achieved.
• An algorithm for seeking optimal control laws for both human and machine is presented.

Conclusions:

• The developed HCMJS model effectively integrates human and machine dynamics for HiTL systems.
• The proposed control synthesis ensures system stability and cost performance with reduced human intervention.
• The method's effectiveness is validated through application to a driver-assistance system.