Organizations: Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI, USA · Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, USA
Language-guided navigation enables mobile robots to interpret natural language commands and navigate toward user-specified targets in unstructured environments. However, integrating open-vocabulary visual perception with classical robotics pipelines for mapping and navigation remains a challenging problem. In this work, we present a complete ROS2-based system that combines frontier-based exploration, open-vocabulary object detection using Grounding DINO, depth-based 3D localization, and semantic mapping to enable end-to-end vision-language grounded navigation in simulation. The system accepts natural language commands such as ``go to the yellow box,'' detects candidate objects from RGB images using text prompts, estimates their positions in the world frame, and generates navigation goals executed by the Nav2 stack. In a Gazebo-based indoor environment with a TurtleBot3 robot, the system autonomously explored the room, detected three of the five target objects, and successfully executed both natural language navigation commands tested, without task-specific training. The results indicate that open-vocabulary grounding models can be integrated into classical navigation pipelines, though performance is limited by detection uncertainty, depth-based localization error, and incomplete exploration coverage. These findings highlight both the feasibility and current limitations of deploying vision-language grounded navigation systems using lightweight, CPU-only architectures.
Figures & tables
Fig. 1: System architecture of the vision-language grounded navigation pipeline, showing the five-layer structure from Gazebo simulation through sensor processing, perception, planning, and mission control.
Fig. 2: Gazebo simulation environment containing five geometric objects used for evaluation.
Fig. 3: Frontier detection on the Cartographer occupancy grid. White cells denote known free space, black cells denote walls/obstacles, and gray cells denote unknown regions. Frontier cells (the boundary between free and unknown space) are clustered, and the weighted strategy selects the navigation goal shown.
TABLE III: Natural Language Command Execution Results (“Reached” means Nav2 reported success at the stored map position)
Fig. 4: Nav2 executing a natural language navigation command. The green line is the global path from the NavFn planner, the red curve is the local trajectory from the DWB controller, and the arrow marker indicates the goal pose derived from the semantic map query.
Natural language interaction provides an intuitive way for non-expert users to communicate with robotic platforms. However, transforming user requests into executable navigation actions remains a challenging task, requiring the integration of language understanding, environment perception, and autonomous navigation. This work presents a language-driven navigation framework that enables mobile robots to interpret user requests in natural language to move the robot to a destination and autonomously navigate towards it. The framework is composed of modular ROS 2 components that cooperate to transform natural language instructions into navigation actions. Given a natural language request referring to a target in the environment (e.g., "go to the mail box"), the system identifies the referenced object, estimates its position using RGB-D data, and generates a navigation goal, which is then executed through the ROS 2 Nav2 navigation stack. The ROS 2-based implementation facilitates portability across different robotic platforms, requiring only the configuration of the corresponding topics and services. The system is evaluated in both simulation and real-world scenarios using a TurtleBot3 Waffle and a Unitree Go2 robot with a RealSense camera. Experimental results show that the framework successfully interprets both direct commands and contextual requests, generates meaningful natural-language feedback, and navigates towards the desired target. These results demonstrate the feasibility of combining semantic perception and autonomous navigation to provide an intuitive human-robot interaction paradigm. Code will be released as open source upon acceptance.
Jose Martínez-Fajardo, Pablo Pueyo, Fernando Caballero +1
Open-world navigation requires robots to make decisions in complex everyday environments while adapting to flexible task requirements. Conventional navigation approaches often rely on dense 3D reconstruction and hand-crafted goal metrics, which limits their generalization across tasks and environments. Recent advances in vision-language navigation (VLN) and vision-language-action (VLA) models enable end-to-end policies conditioned on natural language, but typically require interactive training, large-scale data collection, or task-specific fine-tuning with a mobile agent. We formulate navigation as a sparse subgoal identification and reaching problem and observe that providing visual anchoring targets for high-level semantic priors enables highly efficient goal-conditioned navigation. Based on this insight, we select visual frontiers as semantic anchors and propose OpenFrontier, a navigation framework that requires no task-specific training or fine-tuning and seamlessly integrates diverse vision-language prior models. OpenFrontier enables efficient navigation with a lightweight system design, without dense 3D semantic mapping, task-specific policy training, or model fine-tuning. We evaluate OpenFrontier across multiple navigation benchmarks and demonstrate strong zero-shot performance, as well as effective real-world deployment on a mobile robot.
Esteban Padilla-Cerdio, Boyang Sun, Marc Pollefeys +1
1ETH Zurich · 2Microsoft Spatial AI Lab · University of Bonn
Vision-language models (VLMs) provide a compelling foundation for reasoning-driven mobile navigation, offering rich contextual understanding and strong generalization from large-scale pretraining. Most existing navigation frameworks rely on imitation learning and therefore require substantial labeled trajectory data, limiting their scalability and robustness. In this work, we propose a parameter-efficient approach to fine-tune a pretrained VLM for autonomous navigation using an Imperative Learning paradigm. By optimizing against differentiable geometric cost fields rather than labeled trajectories, our model learns to generate collision-free paths exclusively from stereoscopic depth observations. We introduce a unified end-to-end navigation pipeline for natural-language-driven robotic control. This system leverages a shared VLM backbone with task-specific Low-Rank Adaptation (LoRA) modules, effectively bridging the gap from semantic target selection to low-level trajectory planning. Our approach achieves competitive Success weighted by Path Length (SPL) in unseen environments while updating less than 1% of the model's total parameters. Qualitative real-world experiments validate sim-to-real generalization and stable path planning without fine-tuning on real-world data. These results highlight a practical approach for deploying VLM-based agents on mobile robots, enabling high-level semantic navigation without the prohibitive requirement for large-scale, labeled trajectory data.
Sebastian Berger, Katharina Winter, Fabian B. Flohr
Munich University of Applied Sciences, Intelligent Vehicles Lab (IVL), 80335 Munich, Germany