ChatPlanner: A Large Language Model Framework for Personalized Public Transit Routing

2026-06-13 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Public Transportation & Urban Mobility, Autonomous Vehicles & Smart Transportation · Depth: Expert, quick

Summary

ChatPlanner is a novel Large Language Model (LLM) framework designed for personalized public transit routing, addressing the challenge of integrating diverse user preferences. It employs fine-tuned LLMs with Retrieval-Augmented Generation (RAG) to extract routing parameters and interpret nuanced user preferences from natural language queries. These preferences are then integrated into the objective function of a public transit routing algorithm. The framework was validated using preference-aware datasets incorporating eight personas and five contexts. Experiments confirmed ChatPlanner reliably generates feasible solutions. Fine-tuning enforces output structure and learns general preference patterns, while RAG provides query-specific context. This combination achieves the highest accuracy in information extraction and preference interpretation, identifying valuable route alternatives overlooked by existing planners.

Key takeaway

For Machine Learning Engineers developing smart city or transportation solutions, ChatPlanner offers a robust method for integrating complex user preferences into routing algorithms. You should combine fine-tuned LLMs with RAG to accurately interpret natural language queries. This ensures your systems generate more valuable, personalized route alternatives. This approach can significantly enhance user satisfaction by moving beyond generic route planning.

Key insights

ChatPlanner integrates fine-tuned LLMs and RAG to enable personalized public transit routing by interpreting natural language preferences.

Principles

• LLMs can interpret nuanced user preferences.
• RAG enhances LLM accuracy with query-specific context.
• Fine-tuning enforces output structure and learns patterns.

Method

Fine-tune LLMs with RAG to extract routing parameters and interpret natural language preferences. Integrate these into a public transit routing algorithm's objective function for personalized results.

In practice

• Develop LLM-powered transit planning systems.
• Use RAG to refine conversational user queries.
• Design preference-aware datasets for LLM training.

Topics

• Large Language Models
• Public Transit Routing
• Personalized Routing
• Retrieval-Augmented Generation
• Natural Language Understanding
• Transportation Optimization

Best for: NLP Engineer, AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• TransitLM: A Large-Scale Dataset and Benchmark for Map-Free Transit Route Generation — TransitLM enables LLMs to learn map-free transit route generation directly from data, implicitly grounding GPS coordinates to stations.
• Interview with Sukanya Mandal: Synthesizing multi-modal knowledge graphs for smart city intelligence — LLMasMMKG uses LLMs to automate synthetic multi-modal knowledge graph creation for smart city cognitive digital twins.
• You’re Not Training ChatGPT By Pasting Data — Prompting, RAG, and training are distinct methods for interacting with LLMs, each with different impacts.
• Learning to Route LLMs from Implicit Cost-Performance Preferences via Meta-Learning — MetaRouter optimizes LLM routing by meta-learning implicit user cost-performance preferences via contextual bandits.
• Sparse Personalized Text Generation with Multi-Trajectory Reasoning — PaT enhances cold-start LLM personalization by integrating style and topic cues via iterative, differential-reward-based reasoning.
• Broadening Access to Transportation Safety Data with Generative AI: A Schema-Grounded Framework for Spatial Natural Language Queries — A schema-grounded natural language interface broadens access to transportation safety data while ensuring reproducible, auditable analysis.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.