Human-in-the-loop pose control of a non-holonomic vehicle

This paper describes the development of a receding-horizon based controller for human-in-the-loop pose control of a car-like vehicle. The target application is to provide assistance to drivers of mining dump trucks when parking their vehicles with respect to loading machinery such as swing loaders and rope shovels - this differs from the traditional parking problem due to visibility constraints and due to the relatively open space surrounding the ´parking spot´. The controller takes account of the steering and velocity loop dynamics, as well as dealing with the motion constraints of the vehicle. Experiments using a 1:5 scale model of a dump-truck/swing-loader system demonstrate the feasibility of the approach