Improving the performance of Learned Controllers in Behavior Trees using Value Function Estimates at Switching Boundaries

Open in new window