Simulated Human Data for Driving Task

该数据集汇总了10位真实人类在多轮驾驶模拟中收集的数据，用于训练代理模型。此外，该数据集还用于训练模拟的人类代理，以评估辅助通信的有效性。规模上，数据涵盖了10位人类参与者，在24至36轮的模拟中收集。所涉及的任务是带有辅助通信的驾驶模拟。

This dataset aggregates data collected from 10 real human participants during multi-turn driving simulations, and is intended for training proxy models. Additionally, it is also utilized to train simulated human agents for evaluating the effectiveness of auxiliary communication. In terms of scale, the data was collected across 24 to 36 simulation rounds from these 10 participants. The tasks involved in this dataset are driving simulations with auxiliary communication.