CARPAL: Confidence-Aware Intent Recognition for Parallel Autonomy
This work addresses the challenge of safe decision-making in parallel autonomy for drivers, though it is incremental as it builds on existing intent recognition methods.
The paper tackled the problem of predicting driver intentions for advanced driver assistance systems by proposing a multi-task neural network that predicts probabilistic driver trajectories and utility statistics for downstream tasks, resulting in improved recall and fall-out metrics on a realistic urban driving dataset.
Predicting driver intentions is a difficult and crucial task for advanced driver assistance systems. Traditional confidence measures on predictions often ignore the way predicted trajectories affect downstream decisions for safe driving. In this paper, we propose a novel multi-task intent recognition neural network that predicts not only probabilistic driver trajectories, but also utility statistics associated with the predictions for a given downstream task. We establish a decision criterion for parallel autonomy that takes into account the role of driver trajectory prediction in real-time decision making by reasoning about estimated task-specific utility statistics. We further improve the robustness of our system by considering uncertainties in downstream planning tasks that may lead to unsafe decisions. We test our online system on a realistic urban driving dataset, and demonstrate its advantage in terms of recall and fall-out metrics compared to baseline methods, and demonstrate its effectiveness in intervention and warning use cases.