Technical Report for Argoverse Challenges on Unified Sensor-based Detection, Tracking, and Forecasting

TL;DR

This paper introduces a unified sensor-based detection, tracking, and forecasting network for autonomous driving, achieving top results in the Argoverse Challenges by integrating multiple tasks into a single model with a BEV encoder.

Contribution

A novel unified network architecture that combines detection, tracking, and forecasting tasks using a BEV encoder with spatial and temporal fusion.

Findings

Achieved 1st place in Detection, Tracking, and Forecasting at Argoverse Challenges.

Demonstrated effective multi-task learning on the Argoverse 2 dataset.

Unified approach outperforms separate task-specific models.

Abstract

This report presents our Le3DE2E solution for unified sensor-based detection, tracking, and forecasting in Argoverse Challenges at CVPR 2023 Workshop on Autonomous Driving (WAD). We propose a unified network that incorporates three tasks, including detection, tracking, and forecasting. This solution adopts a strong Bird's Eye View (BEV) encoder with spatial and temporal fusion and generates unified representations for multi-tasks. The solution was tested in the Argoverse 2 sensor dataset to evaluate the detection, tracking, and forecasting of 26 object categories. We achieved 1st place in Detection, Tracking, and Forecasting on the E2E Forecasting track in Argoverse Challenges at CVPR 2023 WAD.