Optimal contact decisions for ergodic exploration

This paper presents a method of determining optimal contact transition times for a surface exploration task. A transition time optimization algorithm is developed which uses a measure of ergodicity as the objective function, requiring the time-averaged dynamics of the system to optimally approximate the spatial average of a distribution on a finite time interval. This approach is demonstrated to be effective for determining a finite number of contact transition times in a computationally efficient way. The approach is generalized to include a measure of energy loss during sampling modes. The algorithm is shown to result in solutions that are qualitatively similar to the observed behavior of human subjects performing feature localization tasks.