We present a hybrid-driven trajectory prediction method based on group emotion. The data driven and model driven methods are combined to make a compromise between the controllability, generality, and efficiency of the method on the basis of simulating more real crowd movements. A hybrid driven method is proposed to improve the reliability of the calculation results based on real crowd data, and ensure the controllability of the model. It reduces the dependence of our model on real data and realizes the complementary advantages of these two kinds of methods. In addition, we divide crowd into groups based on human relations in society. So our method can calculate the movements in different scales. We predict individual movement trajectories according to the trajectories of group and fully consider the influence of the group movement state on the individual movements. Besides we also propose a group emotion calculation method and our method also considers the effect of group emotion on crowd movements.

中文翻译：

---

基于群体情感的混合驱动弹道预测

我们提出了一种基于群体情感的混合驱动轨迹预测方法。在模拟更多实际人群运动的基础上，将数据驱动方法和模型驱动方法结合起来以在该方法的可控制性，通用性和效率之间做出折衷。提出了一种混合驱动的方法，以提高基于真实人群数据的计算结果的可靠性，并确保模型的可控制性。它减少了我们的模型对真实数据的依赖性，并实现了这两种方法的互补优势。此外，我们根据社会中的人际关系将人群分为几类。因此我们的方法可以计算不同比例的运动。我们根据群体的轨迹来预测个体的运动轨迹，并充分考虑群体运动状态对个体运动的影响。此外，我们还提出了一种群体情感计算方法，该方法还考虑了群体情感对人群运动的影响。