Approach behaviors into small groups of people

Time: Thu 2019-08-29 11.15 - 12.15

Lecturer: Fankgai Yang, CST, EECS

Location: Room 4423, Lindstedtsvägen 5, KTH, Stockholm

Contact:

While many works involving human-agent interactions have focused on individuals or crowds, modelling interactions on the group scale has not been considered in depth. It is important to model this group level as in real situations crowds are often composed of an intermediate level consisting of clusters of small groups. In this seminar, the research question is narrowed down in generating approach behaviors into small groups of people. My ongoing PhD research is presented in three progressive parts: a) a social-aware navigation method for generating socially-acceptable and realistic approach group trajectories and creating a synthetic dataset with a mixture of real data. b) data-driven methods trained on the generated dataset for generating approach group trajectories. c) collection of high quality group interaction data. Moving forward, we aim to retrain the data-driven methods with collected data and develop new methods with less but high quality training data.

Page responsible:Web editors at EECS
Belongs to: Computational Science and Technology
Last changed: Aug 05, 2019