Identifying Dominance Traits of Pedestrians based on Walking Gaits
Tanmay Randhavane
COMP 790-058: Multi-Agent Simulation for Crowds and Autonomous Driving

With applications like autonomous driving and surveillance, understanding the behavior of pedestrians has become an important problem. Attempts have been made to learn pedestrian behavior using data driven techniques. In this work, I plan to learn dominance traits of pedestrians' based on their walking gaits.

Previous Work

Previous works have learned pedestrian behaviors based on their:

Out of these sources of information, walking gaits have been used to identify pedestrians': Pedestrians' personality traits like dominance have been shown to be related to their walking gaits but a mathematical model to map their gait features and dominance characteristics has not been developed.

Semester Goals

Over the course of this semester, my goal is to develop a mathematical model to map walking gaits of pedestrians to their dominance characteristics. This mapping can then be used to identify pedestrians' dominance characteristics and to generate walking gaits with desired dominance characteristics.