TrajEvo: Designing Trajectory Prediction Heuristics via LLM-driven Evolution

2025-05-07

AI Summary

Trajectory prediction is critical for social robotics and autonomous driving, yet conventional heuristic methods suffer from low accuracy, while deep learning approaches incur high computational overhead, lack interpretability, and exhibit poor generalization. This paper proposes a novel LLM-driven paradigm for automatic heuristic design, integrating evolutionary algorithms with large language models (LLMs). We introduce a cross-generational elite sampling strategy and a statistical feedback loop to jointly preserve population diversity and enhance LLM reasoning quality. The resulting heuristics achieve millisecond-level inference latency, explicit behavioral logic for full interpretability, and strong out-of-distribution generalization. On the ETH-UCY benchmark, our method significantly outperforms traditional heuristics; on the unseen Stanford Drone Dataset (SDD), it surpasses both existing heuristics and state-of-the-art deep learning models. Our approach establishes a reliable, transparent, and deployment-friendly trajectory prediction framework for human–robot collaborative environments.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Search and Optimization: Heuristic Search
• Intelligent Robots: Behavior Learning & Control

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Trajectory prediction is a crucial task in modeling human behavior, especially in fields as social robotics and autonomous vehicle navigation. Traditional heuristics based on handcrafted rules often lack accuracy, while recently proposed deep learning approaches suffer from computational cost, lack of explainability, and generalization issues that limit their practical adoption. In this paper, we introduce TrajEvo, a framework that leverages Large Language Models (LLMs) to automatically design trajectory prediction heuristics. TrajEvo employs an evolutionary algorithm to generate and refine prediction heuristics from past trajectory data. We introduce a Cross-Generation Elite Sampling to promote population diversity and a Statistics Feedback Loop allowing the LLM to analyze alternative predictions. Our evaluations show TrajEvo outperforms previous heuristic methods on the ETH-UCY datasets, and remarkably outperforms both heuristics and deep learning methods when generalizing to the unseen SDD dataset. TrajEvo represents a first step toward automated design of fast, explainable, and generalizable trajectory prediction heuristics. We make our source code publicly available to foster future research at https://github.com/ai4co/trajevo.

Problem

• Improving trajectory prediction accuracy in human behavior modeling
• Reducing computational cost and explainability issues in deep learning methods
• Enhancing generalization of trajectory prediction heuristics across datasets

Innovation

• LLM-driven evolution for trajectory prediction
• Cross-Generation Elite Sampling enhances diversity
• Statistics Feedback Loop improves prediction analysis