Scaling the dynamical systems approach to path planning

As robotic agents are called upon to perform in increasingly complex and naturalistic physical environments, it will be necessary to develop design strategies that enable the specification of complex tasks that can be carried out in flexible ways. This requires the ability to react flexibly to dynamically changing environments as well as the ability to perform decision-making and sequencing of behavior. We describe an approach to decision-making and sequencing that utilizes the qualitative theory of dynamical systems to model stable decision-making in changing environments. The task dynamics methodology is considered as an approach to scaling the ´dynamical systems´ approach to autonomous path planning