Origin-Conditional Trajectory Encoding: Measuring Urban Configurational Asymmetries through Neural Decomposition

2025-12-03

AI Summary

Current urban trajectory analysis suffers from three fundamental disconnects: (1) decoupled spatiotemporal representation, (2) neglect of navigational directional asymmetry (A→B ≠ B→A), and (3) overreliance on external auxiliary data. To address these, we propose a *source-conditioned neural decomposition framework* that jointly learns spatial configuration—geometrically encoded via visibility ratio and curvature—and mobility patterns—modeled by bidirectional LSTMs. We further introduce learnable source embeddings and contrastive learning to explicitly capture cognitive asymmetry, disentangling shared cognitive priors from source-specific spatial narratives. Evaluated on six synthetic cities and real-world trajectories from Beijing’s Xicheng District, our method reveals systematic cognitive inequality rooted in urban morphology. This work establishes the first geometry-only, quantifiable assessment tool for urban planning, architectural design, and intelligent navigation—free from external data dependencies.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Urban analytics increasingly relies on AI-driven trajectory analysis, yet current approaches suffer from methodological fragmentation: trajectory learning captures movement patterns but ignores spatial context, while spatial embedding methods encode street networks but miss temporal dynamics. Three gaps persist: (1) lack of joint training that integrates spatial and temporal representations, (2) origin-agnostic treatment that ignores directional asymmetries in navigation ($A o B e B o A$), and (3) over-reliance on auxiliary data (POIs, imagery) rather than fundamental geometric properties of urban space. We introduce a conditional trajectory encoder that jointly learns spatial and movement representations while preserving origin-dependent asymmetries using geometric features. This framework decomposes urban navigation into shared cognitive patterns and origin-specific spatial narratives, enabling quantitative measurement of cognitive asymmetries across starting locations. Our bidirectional LSTM processes visibility ratio and curvature features conditioned on learnable origin embeddings, decomposing representations into shared urban patterns and origin-specific signatures through contrastive learning. Results from six synthetic cities and real-world validation on Beijing's Xicheng District demonstrate that urban morphology creates systematic cognitive inequalities. This provides urban planners quantitative tools for assessing experiential equity, offers architects insights into layout decisions'cognitive impacts, and enables origin-aware analytics for navigation systems.

Problem

• Integrates spatial and temporal representations in trajectory analysis
• Addresses origin-agnostic treatment ignoring directional navigation asymmetries
• Reduces reliance on auxiliary data using fundamental geometric properties

Innovation

• Jointly learns spatial and movement representations using geometric features
• Decomposes navigation into shared patterns and origin-specific signatures
• Uses bidirectional LSTM with contrastive learning for origin-aware analysis