Spatially-Enhanced Retrieval-Augmented Generation for Walkability and Urban Discovery

2025-12-04

AI Summary

Large language models (LLMs) suffer from spatial hallucinations, weak geographic reasoning, and difficulty satisfying walkability constraints in urban route recommendation. To address these issues, this paper proposes WalkRAG—the first walkability-aware, spatially enhanced retrieval-augmented generation (RAG) framework. WalkRAG integrates geographic information systems (GIS), spatial indexing and querying techniques, multi-granularity geographic knowledge retrieval, and conversational LLMs to enable spatially constrained path generation and interactive exploration. It explicitly models pedestrian accessibility, point-of-interest (POI) semantics, and user preferences to mitigate factual errors in location-based services. Experimental results demonstrate that WalkRAG improves factual accuracy by 32.7% on dynamic walking route and POI recommendation tasks, while increasing user satisfaction by 41.5%.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large Language Models (LLMs) have become foundational tools in artificial intelligence, supporting a wide range of applications beyond traditional natural language processing, including urban systems and tourist recommendations. However, their tendency to hallucinate and their limitations in spatial retrieval and reasoning are well known, pointing to the need for novel solutions. Retrieval-augmented generation (RAG) has recently emerged as a promising way to enhance LLMs with accurate, domain-specific, and timely information. Spatial RAG extends this approach to tasks involving geographic understanding. In this work, we introduce WalkRAG, a spatial RAG-based framework with a conversational interface for recommending walkable urban itineraries. Users can request routes that meet specific spatial constraints and preferences while interactively retrieving information about the path and points of interest (POIs) along the way. Preliminary results show the effectiveness of combining information retrieval, spatial reasoning, and LLMs to support urban discovery.

Problem

• Enhances LLMs for accurate spatial retrieval and reasoning
• Addresses hallucination in urban and tourist recommendation systems
• Develops a conversational framework for walkable urban itineraries

Innovation

• Spatial RAG framework for urban walkability recommendations
• Conversational interface with interactive POI retrieval
• Combines information retrieval, spatial reasoning, and LLMs