What’s new

    • Mar. 2022 Our paper “Step Attention: Sequential Pedestrian Trajectory Prediction” is published and available on IEEE Sensors Journal. The work addresses the problem of trajectory prediction for pedestrians, and shows promising performance on both benchmark dataset and data collected from real-world intersections.

    • June 2021. Our paper “A Learning-based Trajectory Prediction Approach for Heterogeneous Traffic Agents: Implications for Transfer Learning” is accepted by ITSC 2021. The paper proposes a deep learning algorithm for accurate trajectory prediction, and investigates using transfer learning to boost prediction accuracy on rarely-seen agents like cyclists.

    • May 2021. Our recent patent “System for Predicting Aggressive Driving” was published, and appeared on media.

    • Mar. 2021. I am joining Tusimple as a Research Engineer Intern this summer. I will be working on prediction and prediction-related fields for autonomous vehicles.

    • Feb. 2021. I gave a presentation at UMich ERS about my most recent work: Step attention- sequential trajectory prediction. The work is about accurate pedestrian trajectory prediction. Corresponding paper is coming out soon.

    • Dec. 2020. Our most recent work “V2XSim: a Simulator for Connected and Automated Vehicle Environment Simulation” is published. We present a vehicle-to-everything (V2X) simulator for connected vehicle environment simulation from robotics perspective. [Demo, PDF, Github]

    • Feb, 2020. Our paper “Increasing GPS localization accuracy with reinforcement learning” is published on IEEE transactions on Intelligent Transportation Systems. It reduces error by ~50% compared to extend kalman filter(EKF) on GPS localization.

    • Oct. 2019. Our paper “Parallel computing algorithm for real-time mapping between large-scale networks” is published on IEEE ITSC 2019. In the paper we propose a scalable massively-parallel algorithm to solve the general mapping problem in large-scale networks in real-time.