ModeSeq Team achieved the second position in the 2024 Waymo Open Dataset Motion Prediction Challenge

In the 2024 Waymo Open Dataset Motion Prediction Challenge, the ModeSeq Team supervised by Prof. Jianping Wang (team members: Mr Zikang Zhou from CityU, Dr Yung-Hui Li from Hon Hai Research Institute and Mr Yu-Kai Huang from CMU) achieved Second Place in the Motion Prediction Challenge. The prizes for second place is $5,000 in Google Cloud credits.

Waymo, a subsidiary of Alphabet Inc, is an American autonomous driving technology company headquartered in California.  In 2019, Waymo publicly released the Waymo Open Dataset to support the AI research community in exploring various aspects of machine perception and autonomous driving technology.  In the Sim Agents Challenge, the teams were given the agent tracks for the past 1 second on a corresponding map, and optionally the associated lidar for this time interval, then they were tasked with simulating 32 realistic joint futures for all the agents in the scene. 

The team's primary goal in the challenge was to anticipate the future intentions of traffic participants while acknowledging and addressing uncertainties ahead. By structuring multiple potential future scenarios for each participant as a sequence, the team significantly increased the range of anticipated future behaviors. These diverse motion predictions empower autonomous vehicles to navigate through unpredictable driving conditions, minimizing the risk of collisions with other traffic participants and thus boosting the overall safety of autonomous driving.