Nonlinear Observers for Human-in-the-loop Control Systems

The development of models for a human-in-the-loop with hardware is an area of ongoing research. The ability to simulate a human-in-the-loop with hardware provides a platform for better understanding the dynamics of human and machine cognition. A human-in-the-loop model provides information that can be used to design more efficient human interfaces and smarter autonomous assistant controllers. This can make a complex task such as flying an aircraft safer and more accessible. This thesis explores different possibilities for human operator models to be modeled in the loop with a vehicle. A human is modeled as a linear state feedback controller in the loop with the task of controlling a simple solid ball. The human arm is modeled controlling a joystick as the human is considered to control the ball with a joystick. Nonlinear sliding mode observers are developed to estimate the gains of a feedback control law and nonlinear sliding mode observers are developed to estimate the torques on the shoulder, elbow, and joystick joints. The nonlinear observers are simulated on a human-in-the-loop system to show the accuracy of the observers