Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Holistic Semantic Representation for Navigational Trajectory Generation

Ji Cao, Tongya Zheng, Qinghong Guo, Yu Wang, Junshu Dai, Shunyu Liu, Jie Yang, Jie Song, Mingli Song

TL;DR

This work introduces HOSER, a holistic semantic representation framework for navigational trajectory generation. It jointly models road-network structure (road- and zone-level), multi-granularity trajectory semantics, and destination-guided generation to produce high-fidelity synthetic trajectories. Empirical results across three real-world datasets show that HOSER outperforms state-of-the-art baselines on global and local similarity metrics and demonstrates strong few-shot and zero-shot generalization. The approach yields trajectories that are effective for downstream tasks, offering a viable privacy-preserving alternative for spatio-temporal data analysis.

Abstract

Trajectory generation has garnered significant attention from researchers in the field of spatio-temporal analysis, as it can generate substantial synthesized human mobility trajectories that enhance user privacy and alleviate data scarcity. However, existing trajectory generation methods often focus on improving trajectory generation quality from a singular perspective, lacking a comprehensive semantic understanding across various scales. Consequently, we are inspired to develop a HOlistic SEmantic Representation (HOSER) framework for navigational trajectory generation. Given an origin-and-destination (OD) pair and the starting time point of a latent trajectory, we first propose a Road Network Encoder to expand the receptive field of road- and zone-level semantics. Second, we design a Multi-Granularity Trajectory Encoder to integrate the spatio-temporal semantics of the generated trajectory at both the point and trajectory levels. Finally, we employ a Destination-Oriented Navigator to seamlessly integrate destination-oriented guidance. Extensive experiments on three real-world datasets demonstrate that HOSER outperforms state-of-the-art baselines by a significant margin. Moreover, the model's performance in few-shot learning and zero-shot learning scenarios further verifies the effectiveness of our holistic semantic representation.

Holistic Semantic Representation for Navigational Trajectory Generation

TL;DR

This work introduces HOSER, a holistic semantic representation framework for navigational trajectory generation. It jointly models road-network structure (road- and zone-level), multi-granularity trajectory semantics, and destination-guided generation to produce high-fidelity synthetic trajectories. Empirical results across three real-world datasets show that HOSER outperforms state-of-the-art baselines on global and local similarity metrics and demonstrates strong few-shot and zero-shot generalization. The approach yields trajectories that are effective for downstream tasks, offering a viable privacy-preserving alternative for spatio-temporal data analysis.

Abstract

Trajectory generation has garnered significant attention from researchers in the field of spatio-temporal analysis, as it can generate substantial synthesized human mobility trajectories that enhance user privacy and alleviate data scarcity. However, existing trajectory generation methods often focus on improving trajectory generation quality from a singular perspective, lacking a comprehensive semantic understanding across various scales. Consequently, we are inspired to develop a HOlistic SEmantic Representation (HOSER) framework for navigational trajectory generation. Given an origin-and-destination (OD) pair and the starting time point of a latent trajectory, we first propose a Road Network Encoder to expand the receptive field of road- and zone-level semantics. Second, we design a Multi-Granularity Trajectory Encoder to integrate the spatio-temporal semantics of the generated trajectory at both the point and trajectory levels. Finally, we employ a Destination-Oriented Navigator to seamlessly integrate destination-oriented guidance. Extensive experiments on three real-world datasets demonstrate that HOSER outperforms state-of-the-art baselines by a significant margin. Moreover, the model's performance in few-shot learning and zero-shot learning scenarios further verifies the effectiveness of our holistic semantic representation.
Paper Structure (43 sections, 19 equations, 10 figures, 6 tables, 1 algorithm)

This paper contains 43 sections, 19 equations, 10 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Methodology
  4. Road Network Encoder
  5. Road-Level Semantic Representation.
  6. Zone-Level Semantic Representation.
  7. Multi-Granularity Trajectory Encoder
  8. Spatio-temporal Point Semantics.
  9. Trajectory Semantics.
  10. Destination-Oriented Navigator
  11. End-to-End Learning
  12. Optimization.
  13. Generation.
  14. Experiments
  15. Experimental Setups
  16. ...and 28 more sections

Figures (10)

  • Figure 1: The overview of the proposed HOSER framework. The Road Network Encoder is responsible for modeling the road network at different levels. The Trajectory Encoder is used to extract semantic information from the partial trajectory, which is then fed into the Navigator and combined with destination guidance to generate the next spatio-temporal point.
  • Figure 2: Visualization of metrics distributions.
  • Figure 3: Visualization of the trajectories in Beijing (a larger view for Beijing, as well as for the other two cities, can be found in Appendix C.1).
  • Figure 4: HOSER's performance with varying amounts of training data across three trajectory datasets. "Full" denotes the complete dataset, with sizes of 629380.0, 481359.0, and 205116.0 for Beijing, Porto, and San Francisco, respectively.
  • Figure 5: Visualization of the trajectories in Beijing.
  • ...and 5 more figures

Theorems & Definitions (4)

  • Definition 1: Road Network
  • Definition 2: Trajectory
  • Definition 3: Trajectory Generation
  • Definition 4: Human Movement Modeling