Divide-and-Conquer for Lane-Aware Diverse Trajectory Prediction
CVPR 2021 | Our work addresses two key challenges in trajectory prediction: (i) learning multimodal outputs and (ii) improving predictions by imposing constraints using driving knowledge. Recent methods have achieved strong performance using multi-choice learning objectives like winner-takes-all (WTA), but they highly depend on their initialization to provide diverse outputs. Our first contribution proposes a novel divide-and-conquer (DAC) approach. As a better initialization technique than WTA objective, it results in diverse outputs without any spurious modes. We also introduce a novel trajectory prediction framework called ALAN, which uses existing lane centerlines as anchors to provide trajectories constrained to the input lanes.