Conditional Generative Adversarial Networks for Speed Control in Trajectory Simulation

TL;DR

This paper introduces Conditional Speed GAN (CSG), a novel method for controllable trajectory prediction that generates diverse, realistic paths based on user-specified speed, useful for simulation and data augmentation.

Contribution

The paper proposes CSG, a simple yet effective GAN-based model that explicitly controls trajectory speed, addressing the lack of controllability in existing multimodal trajectory prediction methods.

Findings

CSG achieves comparable benchmark metrics to state-of-the-art methods.

Naive concatenation aggregation performs similarly to attention and pooling mechanisms.

CSG is effective for simulating different pacing environments.

Abstract

Motion behaviour is driven by several factors -- goals, presence and actions of neighbouring agents, social relations, physical and social norms, the environment with its variable characteristics, and further. Most factors are not directly observable and must be modelled from context. Trajectory prediction, is thus a hard problem, and has seen increasing attention from researchers in the recent years. Prediction of motion, in application, must be realistic, diverse and controllable. In spite of increasing focus on multimodal trajectory generation, most methods still lack means for explicitly controlling different modes of the data generation. Further, most endeavours invest heavily in designing special mechanisms to learn the interactions in latent space. We present Conditional Speed GAN (CSG), that allows controlled generation of diverse and socially acceptable trajectories, based on user controlled speed. During prediction, CSG forecasts future speed from latent space and conditions its generation based on it. CSG is comparable to state-of-the-art GAN methods in terms of the benchmark distance metrics, while being simple and useful for simulation and data augmentation for different contexts such as fast or slow paced environments. Additionally, we compare the effect of different aggregation mechanisms and show that a naive approach of concatenation works comparable to its attention and pooling alternatives.