Pedestrian trajectory prediction improves with adaptive domain learning
Trajectory Prediction Method for person under WSN with Hierarchical Envelope Adversarial Domain Adaptation
Computer Vision and Pattern Recognition
Summary
Predicting where people will walk is useful for things like self-driving cars and smart city cameras, but it’s hard because people’s movements can be very different in each place and situation. The authors noticed current methods only look at one type of social interaction and fail to handle different environments well. They created a new way to model multiple interaction types and adapt to different settings automatically. Their method predicts pedestrian paths more accurately across varied real-world data and runs fast enough for practical use.
What this means in practice
- •For autonomous vehicle developers: Improve pedestrian path prediction models to better adapt to different traffic environments for safer autonomous driving.
- •For surveillance system engineers: Enhance pedestrian movement prediction to boost accuracy and robustness in intelligent surveillance across varied scenes.
Authors
Wenqian Jiang, Qifeng Sun, Xiaosong Gao, Xin Zou, Xingrui Ni, Yuan Hu, Yongming Li
Abstract
Pedestrian trajectory prediction holds significant applications across autonomous driving, intelligent surveillance, and smart cities. However, the complexity and unpredictability of pedestrian interactions make this task a persistent challenge. Current models face two common limitations: 1) Consider only a single type of social coupling. This introduces redundant interactions. More critically, it fails to account for the varying nature of trajectory coupled across different scenarios and between different pedestrians. Consequently, models cannot deeply adapt to diverse scenes. 2) Inadequate domain shift handling. Although very few methods address domain shift, they rely on statistical criterion-based domain distribution alignment. Such approaches exhibit strong dependency on predefined statistical metrics. This leads to significant limitations in complex and dynamic environments. To address these problems, a Hierarchical Envelope Adversarial Domain Adaptation with Loosely-Tightly coupled model (HEADA-LTCTPM) is proposed. First, an envelope sample construction mechanism was proposed. It constructs envelope samples and extends them into graph structures. Secondly, an adversarial domain adaptation module was proposed. This integrates both local and global domain adaptation strategies. It dynamically adapts to diverse coupled relationships across scenarios. These innovations collectively enhance prediction accuracy and robustness. In the experimental section, two representative public datasets were employed for validation and comprehensive comparisons with six relevant baseline algorithms were conducted. Results demonstrate that our model achieves significantly higher accuracy compared to existing methods, with the ADE and FDE metrics reduced by 17.6% and 19.1% respectively. The time overhead meets practical requirements, which verifies the effectiveness of our key innovations.