CS 326 A: Motion Planning Target Tracking and Virtual Cameras
Papers First paper: H.H. Gonzalez-Banos, C.Y. Lee, and J.C. Latombe. Real-Time Combinatorial Tracking of a Target Moving Unpredictably Among Obstacles. Proc. IEEE Int. Conf. on Robotics and Automation, Second paper: D. Nieuwenhuisen, M.H. Overmars. Motion planning for Camera Movements in Virtual Environments
General Problem Placement of camera / selection of viewpoint Make “important” things visible, e.g, keep moving object (target) in view Natural motion of camera
What is Known in Advance? Environment? Target trajectory? No distance constraint Constant distance target robot
What Goal is Possible? Always keep the target in sight Keep the target in sight as long as possible Minimize time when target is not visible
Techniques Known environment and target trajectory Back-chaining of visibility regions (open-loop strategy) Known environment, but unknown trajectory Real-time dynamic programming, with horizon h
Target Tracking target robot
States Are Indexed by Time State = (robot-position, target-position, time) Action = (stop, up, down, right, left) Outcome of an action = 5 possible states, each with probability 0.2 ([i,j], [u,v], t) ([i+1,j], [u,v], t+1) ([i+1,j], [u-1,v], t+1) ([i+1,j], [u+1,v], t+1) ([i+1,j], [u,v-1], t+1) ([i+1,j], [u,v+1], t+1) right Each state has 25 successors
h-Step Planning Process Planning horizon h “Terminal” states: States where the target is not visible States at depth h Reward function: Target visible +1 Target not visible 0 Maximizing the sum of rewards ~ maximizing escape time R (state) = t where 0 < < 1 discounting
h-Step Planning Process Planning horizon h The planner computes the optimal choice for the first step. This step is executed. And everything is repeated again … (sliding horizon)
h-Step Planning Process Planning horizon h h is chosen such that the computation over the tree can be done in one increment of time (real-time constraint)
Techniques Known environment and target trajectory Back-chaining of visibility regions (open-loop strategy) Known environment, but unknown trajectory Real-time dynamic programming, with horizon h One advantage: Flexibility Two difficulties: - Computation of visibility regions - Large branching factor
Techniques Known environment and target trajectory Back-chaining of visibility regions (open-loop strategy) Known environment, but unknown trajectory Real-time dynamic programming, with horizon h Unknown environment and target trajectory Risk-based approach (1 st paper)