From Prediction to Planning With Goal Conditioned Lane Graph Traversals

The field of motion prediction for automated driving has seen tremendous progress recently, bearing ever-more mighty neural network architectures. Leveraging these powerful models bears great potential for the closely related planning task. In this work, we show that state-of-the-art prediction models can be converted into goal-directed planners. To this end, we propose a novel goal-conditioning method. Our key insight is that conditioning prediction on a navigation goal at the behaviour level outperforms other widely adopted methods, with the additional benefit of increased model interpretability. Moreover, our Method can be applied at inference time only. Hence, no ground-truth navigation command is required during training. We evaluate our method on a large open-source dataset and show promising performance in a comprehensive benchmark. Code is available under https://mh0797.github.io/gc-pgp/.