Research in man-machine system is focusing on understanding the collaboration between the control system and the human operator, and adopting a systematic design approach is crucial for optimum system performance. In the same time, human operator is considered the most vulnerable part in the system. Therefore, modeling and understanding human behaviors is a central role of human-machine system.
Modeling human behavior is a challenging task. Unlike the modeling of physical systems and processes based on physics are very well understood and founded, the models of the human behavior are not advanced as these models of physical systems. The main reason behind this fact, is that human behavior are highly complex nonlinear, time varying systems.
Modeling human behavior has been approached from several domains and theories such as optimal control theory, general control theory, expert systems and cognitive science. The resulting models in most cases are not realistic, simplified not robust, and do not support the full man- machine systems modeling requirements for analysis and design of complex human-in- the-loop systems. In addition, these models are application orientated, where the parameter values of these models must be carefully tuned for each application.
In particular, the currant engineering models of the human operator considers the human as a black box with constrained input processing and simple types of dynamics. These types of models leak the most important aspect of the human, which is the ability to learn from and adapt to the environment. Also, using the the existing system identification algorithms which relies on the repeatable features of the experimental to give best fit models results in stationary models. Clearly, modeling of such highly complex systems i.e., human behavior, requires combining several domains of knowledge including mathematical techniques and intelligent systems.
This project presents some ideas of applying dynamical system theory, control theory, machine learning techniques and data driven models to develop more realistic model for human behavior. Moreover, the project investigates the concept of human-human in the loop, where there are several humans in one loop.