MultiCrowdNav

该数据集是一个用于社交导航的多代理模拟器，它对CrowdNav模拟器进行了增强，支持多个代理，并整合了ORCA和社交力模型以模拟人类互动。此外，该模拟器还促进了从PPO到MAPPO的迁移，并支持多种人类策略，以实现更为真实的人际互动模拟。该数据集的规模涵盖多个代理和行人，其任务是进行多代理社交导航。

This dataset is a multi-agent simulator designed for social navigation. It enhances the CrowdNav simulator by supporting multiple agents and incorporating ORCA and social force models to simulate human interactions. Additionally, this simulator facilitates the transfer from PPO to MAPPO and supports a variety of human-like strategies to achieve more realistic simulations of interpersonal interactions. The dataset covers scenarios with multiple agents and pedestrians, with its core task being multi-agent social navigation.