Human trajectory forecasting
Since the past few decades, human trajectory forecasting has been a field of active research owing to its numerous real-world applications: evacuation situation analysis, deployment of intelligent transport systems, traffic operations, to name a few. We can cast the problem of human trajectory forecasting as learning a representation of human social interactions. However, social interactions in crowded environments are not only diverse but often subtle. Recently, deep learning methods have outperformed their handcrafted counterparts, as they learned about human-human interactions in a more generic data-driven fashion. However, modeling social interactions is an extremely challenging task as no fixed set of rules governs human motion.
A task closely related to learning human social interactions is forecasting the surrounding people's movement, which conforms to common social norms. We can refer to this task of forecasting human motion as human trajectory forecasting. One of the applications of human trajectory forecasting is to improve robot mobility operation in a socially compliant manner in crowded spaces. We want to propose a method for effectively capturing Human-Human social interactions occurring in dense crowds that can indirectly affect the robot's anticipation capability.
