NaviFormer: A Deep Reinforcement Learning Transformer-like Model to Holistically Solve the Navigation Problem

2026-04-18

AI Summary

This work proposes NaviFormer, a novel end-to-end navigation framework that integrates high-level path planning and low-level trajectory generation within a unified architecture, addressing the limitations of traditional methods that treat these components in isolation and thus fail to achieve global cooperative optimization. NaviFormer is the first to adapt the Transformer architecture to end-to-end navigation, leveraging a single deep reinforcement learning model to jointly predict routes and trajectories while simultaneously modeling sequential decision-making and environmental perception. Experimental results demonstrate that NaviFormer not only maintains high navigation accuracy but also achieves substantially superior computational efficiency compared to existing approaches, making it well-suited for real-time global navigation scenarios.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Intelligent Robots: Localization, Mapping, and Navigation
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Path planning is usually solved by addressing either the (high-level) route planning problem (waypoint sequencing to achieve the final goal) or the (low-level) path planning problem (trajectory prediction between two waypoints avoiding collisions). However, real-world problems usually require simultaneous solutions to the route and path planning subproblems with a holistic and efficient approach. In this paper, we introduce NaviFormer, a deep reinforcement learning model based on a Transformer architecture that solves the global navigation problem by predicting both high-level routes and low-level trajectories. To evaluate NaviFormer, several experiments have been conducted, including comparisons with other algorithms. Results show competitive accuracy from NaviFormer since it can understand the constraints and difficulties of each subproblem and act consequently to improve performance. Moreover, its superior computation speed proves its suitability for real-time missions.

Problem

• route planning
• path planning
• navigation
• holistic approach
• real-time navigation

Innovation

• NaviFormer
• Transformer architecture
• deep reinforcement learning
• holistic navigation
• real-time path planning