Conditional Vehicle Trajectories Prediction in CARLA Urban Environment

TL;DR

This paper presents a hybrid approach for predicting vehicle trajectories in complex urban environments using raw signals and mid-level representations, achieving state-of-the-art results in CARLA benchmarks.

Contribution

The work introduces an original architecture inspired by social pooling LSTM for trajectory prediction, with a novel label augmentation mechanism for better generalization.

Findings

Significant improvements over previous state-of-the-art in CARLA benchmark

Effective handling of complex urban scenarios with raw and mid-level data

Proven capability to predict trajectories for both ego and neighboring vehicles

Abstract

Imitation learning is becoming more and more successful for autonomous driving. End-to-end (raw signal to command) performs well on relatively simple tasks (lane keeping and navigation). Mid-to-mid (environment abstraction to mid-level trajectory representation) or direct perception (raw signal to performance) approaches strive to handle more complex, real life environment and tasks (e.g. complex intersection). In this work, we show that complex urban situations can be handled with raw signal input and mid-level representation. We build a hybrid end-to-mid approach predicting trajectories for neighbor vehicles and for the ego vehicle with a conditional navigation goal. We propose an original architecture inspired from social pooling LSTM taking low and mid level data as input and producing trajectories as polynomials of time. We introduce a label augmentation mechanism to get the level of generalization that is required to control a vehicle. The performance is evaluated on CARLA 0.8 benchmark, showing significant improvements over previously published state of the art.