Cooperative Probabilistic Trajectory Forecasting under Occlusion
This work addresses safety-critical navigation for connected agents in occluded environments, such as GPS-denied or indoor settings, but is incremental as it builds on existing cooperative perception methods.
The paper tackles the problem of predicting pedestrian trajectories under occlusion by designing an end-to-end network that cooperatively estimates occluded pedestrian states and predicts trajectories with safety guarantees. The result shows that the uncertainty-aware trajectory prediction is almost similar to ground truth trajectories without occlusion.
Perception and planning under occlusion is essential for safety-critical tasks. Occlusion-aware planning often requires communicating the information of the occluded object to the ego agent for safe navigation. However, communicating rich sensor information under adverse conditions during communication loss and limited bandwidth may not be always feasible. Further, in GPS denied environments and indoor navigation, localizing and sharing of occluded objects can be challenging. To overcome this, relative pose estimation between connected agents sharing a common field of view can be a computationally effective way of communicating information about surrounding objects. In this paper, we design an end-to-end network that cooperatively estimates the current states of occluded pedestrian in the reference frame of ego agent and then predicts the trajectory with safety guarantees. Experimentally, we show that the uncertainty-aware trajectory prediction of occluded pedestrian by the ego agent is almost similar to the ground truth trajectory assuming no occlusion. The current research holds promise for uncertainty-aware navigation among multiple connected agents under occlusion.