Rethinking Trajectory Prediction via "Team Game"
This work addresses trajectory prediction in multi-agent settings like team games, offering a more nuanced approach to interaction modeling, though it appears incremental in building on existing methods.
The paper tackles the problem of multi-agent trajectory prediction by introducing a novel formulation that explicitly models interactive group consensus via a hierarchical latent space, allowing joint capture of group-level and individual-level interactions, and achieves superior performance on team sports and pedestrian datasets.
To accurately predict trajectories in multi-agent settings, e.g. team games, it is important to effectively model the interactions among agents. Whereas a number of methods have been developed for this purpose, existing methods implicitly model these interactions as part of the deep net architecture. However, in the real world, interactions often exist at multiple levels, e.g. individuals may form groups, where interactions among groups and those among the individuals in the same group often follow significantly different patterns. In this paper, we present a novel formulation for multi-agent trajectory prediction, which explicitly introduces the concept of interactive group consensus via an interactive hierarchical latent space. This formulation allows group-level and individual-level interactions to be captured jointly, thus substantially improving the capability of modeling complex dynamics. On two multi-agent settings, i.e. team sports and pedestrians, the proposed framework consistently achieves superior performance compared to existing methods.