Field Theory Based Navigation for Autonomous Mobile Machines

Abstract The paper discusses the role of field theory in the navigation of autonomous agents. Although the current forms of field theory based implementations have serious limitations the method is one of the most promising for providing a generic and robust methodology. The aim here is to remedy the shortcomings and provide a case study in which a biased random walk strategy is compared with a standard chemotaxis method in a chemical gradient field. Computer simulations are presented which demonstrate that the seemingly inefficient biased random walk is the better strategy under naturally unstableconditions.