Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Context-Aware Trajectory Anomaly Detection

Haoji Hu, Jina Kim, Jinwei Zhou, Sofia Kirsanova, JangHyeon Lee, Yao-Yi Chiang

TL;DR

This paper proposes a context-aware anomaly detection approach that models contextual information related to trajectories based on a trajectory reconstruction framework guided by contextual factors such as agent ID and contextual POI embedding and significantly outperformed existing methods by effectively modeling contextual information.

Abstract

Trajectory anomaly detection is crucial for effective decision-making in urban and human mobility management. Existing methods of trajectory anomaly detection generally focus on training a trajectory generative model and evaluating the likelihood of reconstructing a given trajectory. However, previous work often lacks important contextual information on the trajectory, such as the agent's information (e.g., agent ID) or geographic information (e.g., Points of Interest (POI)), which could provide additional information on accurately capturing anomalous behaviors. To fill this gap, we propose a context-aware anomaly detection approach that models contextual information related to trajectories. The proposed method is based on a trajectory reconstruction framework guided by contextual factors such as agent ID and contextual POI embedding. The injection of contextual information aims to improve the performance of anomaly detection. We conducted experiments in two cities and demonstrated that the proposed approach significantly outperformed existing methods by effectively modeling contextual information. Overall, this paper paves a new direction for advancing trajectory anomaly detection.

Context-Aware Trajectory Anomaly Detection

TL;DR

This paper proposes a context-aware anomaly detection approach that models contextual information related to trajectories based on a trajectory reconstruction framework guided by contextual factors such as agent ID and contextual POI embedding and significantly outperformed existing methods by effectively modeling contextual information.

Abstract

Trajectory anomaly detection is crucial for effective decision-making in urban and human mobility management. Existing methods of trajectory anomaly detection generally focus on training a trajectory generative model and evaluating the likelihood of reconstructing a given trajectory. However, previous work often lacks important contextual information on the trajectory, such as the agent's information (e.g., agent ID) or geographic information (e.g., Points of Interest (POI)), which could provide additional information on accurately capturing anomalous behaviors. To fill this gap, we propose a context-aware anomaly detection approach that models contextual information related to trajectories. The proposed method is based on a trajectory reconstruction framework guided by contextual factors such as agent ID and contextual POI embedding. The injection of contextual information aims to improve the performance of anomaly detection. We conducted experiments in two cities and demonstrated that the proposed approach significantly outperformed existing methods by effectively modeling contextual information. Overall, this paper paves a new direction for advancing trajectory anomaly detection.

Paper Structure

This paper contains 12 sections, 4 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Problem Statement
  3. Method
  4. Trajectory Preprocessing
  5. Context-Aware Variational Autoencoder
  6. Contextual Representation Generation
  7. Context-Aware Token Sequence Reconstruction
  8. Agent-Level Anomaly Inference
  9. Experiments
  10. Experiment Settings.
  11. Experiment Results.
  12. Case Study of POI Contextual Embeddings.

Figures (2)

  • Figure 1: An overview of the proposed method. Section \ref{['sec:preprocess']} shows preprocessing steps to process Global Positioning System (GPS) trajectory into stay points and stay points into grid token sequence. Section \ref{['sec:vae']} presents how we encode trajectory grid tokens into context-aware Variational Autoencoder (VAE) conditioning on contextual representations, which incorporates agent ID embedding and Points of Interest (POI) contextual embedding. Lastly, Section \ref{['sec:infer']} shows the retrieval of an agent-level anomaly score from the subtrajectory anomaly scores.
  • Figure 2: A visualization of grids from a specific cluster label generated by clustering the learned contextual embeddings of Points of Interest (POI).