DRoPE: Directional Rotary Position Embedding for Efficient Agent Interaction Modeling

2025-03-19

AI Summary

Autonomous driving trajectory generation faces the “impossible triangle” trade-off among accuracy, computational efficiency, and memory overhead. To address this, we propose Directional Rotational Position Encoding (DRoPE), the first method to align RoPE’s 2D rotational encoding with agent heading—enabling zero-overhead modeling of relative angular information via an identity scalar correction. We theoretically prove DRoPE achieves O(n) time complexity while introducing no additional memory footprint. Extensive evaluation across multiple state-of-the-art models demonstrates: (i) maintained or improved trajectory prediction accuracy; (ii) up to 47% reduction in GPU memory consumption; and (iii) inference latency scaling linearly with sequence length. Our approach bridges theoretical rigor and engineering practicality, establishing a new paradigm for efficient end-to-end trajectory prediction.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search
• Machine Learning: Mixture of Experts (MoE)

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate and efficient modeling of agent interactions is essential for trajectory generation, the core of autonomous driving systems. Existing methods, scene-centric, agent-centric, and query-centric frameworks, each present distinct advantages and drawbacks, creating an impossible triangle among accuracy, computational time, and memory efficiency. To break this limitation, we propose Directional Rotary Position Embedding (DRoPE), a novel adaptation of Rotary Position Embedding (RoPE), originally developed in natural language processing. Unlike traditional relative position embedding (RPE), which introduces significant space complexity, RoPE efficiently encodes relative positions without explicitly increasing complexity but faces inherent limitations in handling angular information due to periodicity. DRoPE overcomes this limitation by introducing a uniform identity scalar into RoPE's 2D rotary transformation, aligning rotation angles with realistic agent headings to naturally encode relative angular information. We theoretically analyze DRoPE's correctness and efficiency, demonstrating its capability to simultaneously optimize trajectory generation accuracy, time complexity, and space complexity. Empirical evaluations compared with various state-of-the-art trajectory generation models, confirm DRoPE's good performance and significantly reduced space complexity, indicating both theoretical soundness and practical effectiveness. The video documentation is available at https://drope-traj.github.io/.

Problem

• Efficient modeling of agent interactions for autonomous driving.
• Overcoming limitations in accuracy, computational time, and memory efficiency.
• Improving trajectory generation with directional rotary position embedding.

Innovation

• DRoPE enhances RoPE for agent interaction modeling.
• Introduces uniform identity scalar for angular alignment.
• Optimizes accuracy, time, and space complexity simultaneously.