COMPUTER-VISION2026-04-14

VULCAN: Vision-Language-Model Enhanced Multi-Agent Cooperative Navigation for Indoor Fire-Disaster Response

Shengding Liu, Qiben Yan

This paper addresses a critical gap in robotics: how to make teams of robots navigate and explore buildings during fires when visibility is poor and conditions are dangerous. Current multi-robot Navigation & LocomotionNavigationMoving through an environment toward a goal. systems work well in normal indoor settings but fail badly when smoke blocks cameras and heat warps sensors. VULCAN combines vision-language models (AI that understands both images and text descriptions) with multi-agent Control & PlanningPlanningFiguring out what the robot should do before or during movement. to help Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. teams explore fire disaster sites more effectively. For developers, this means: (1) you can use large language models to help robots understand hazardous environments, (2) coordinating multiple robots becomes more robust when you fuse Modern Robot LearningMultimodalUsing more than one type of input, like vision, language, touch, or proprioception. Perception & SensingSensorA device that provides information about the robot or its environment. data and language understanding, and (3) there's now a realistic fire-disaster simulator (based on Habitat-Matterport3D) to test your multi-robot systems under harsh conditions.

THE PROBLEM

This paper focuses on computer vision. This paper addresses a critical gap in robotics: how to make teams of robots navigate and explore buildings during fires when visibility is poor and conditions are dangerous. Current multi-robot Navigation & LocomotionNavigationMoving through an environment toward a goal. systems work well in normal indoor settings but fail badly when smoke blocks cameras and heat warps sensors. VULCAN combines vision-language models (AI that understands both images and text descriptions) with multi-agent Control & PlanningPlanningFiguring out what the robot should do before or during movement. to help Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. teams explore fire disaster sites more effectively. For developers, this means: (1) you can use large language models to help robots understand hazardous environments, (2) coordinating multiple robots becomes more robust when you fuse Modern Robot LearningMultimodalUsing more than one type of input, like vision, language, touch, or proprioception. Perception & SensingSensorA device that provides information about the robot or its environment. data and language understanding, and (3) there's now a realistic fire-disaster simulator (based on Habitat-Matterport3D) to test your multi-robot systems under harsh conditions. Read the paper by tracking the Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. definition, the Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or data assumptions, and the evidence that supports the claimed improvement.

HOW IT WORKS

1

Task framing

The paper frames the work as computer vision. Start here because it defines what success means and which assumptions the rest of the method inherits.

2

Core method

This paper addresses a critical gap in robotics: how to make teams of robots navigate and explore buildings during fires when visibility is poor and conditions are dangerous. Current multi-robot Navigation & LocomotionNavigationMoving through an environment toward a goal. systems work well in normal indoor settings but fail badly when smoke blocks cameras and heat warps sensors. VULCAN combines vision-language models (AI that understands both images and text descriptions) with multi-agent Control & PlanningPlanningFiguring out what the robot should do before or during movement. to help Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. teams explore fire disaster sites more effectively. For developers, this means: (1) you can use large language models to help robots understand hazardous environments, (2) coordinating multiple robots becomes more robust when you fuse Modern Robot LearningMultimodalUsing more than one type of input, like vision, language, touch, or proprioception. Perception & SensingSensorA device that provides information about the robot or its environment. data and language understanding, and (3) there's now a realistic fire-disaster simulator (based on Habitat-Matterport3D) to test your multi-robot systems under harsh conditions. When reading the method section, identify the inputs, the learned or engineered representation, and the Core ConceptsActionA command the robot sends to its motors, controller, or low-level system. or prediction produced by the system.

3

Data and supervision

For robotics work, the data story is part of the method: check whether the system depends on Imitation & Reinforcement LearningTeleoperation (teleop)A human remotely controlling the robot, often to collect demonstrations., Simulation & Sim-to-RealSimulationA virtual environment where robots can be trained or tested., internet video, human labels, or Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. rollouts.

4

Evaluation evidence

The paper should be judged through its Simulation & Sim-to-RealEvaluationMeasuring how well a robot system performs. protocol: what data is used, what Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or simulator is tested, and which Evaluation & ResearchBaselineA reference method used for comparison. comparisons support the claim. Look for the gap between the headline result and the Simulation & Sim-to-RealDeploymentPutting the trained system on a real robot. setting you would actually care about.

KEY RESULTS

Main contributionConceptual contribution

This paper addresses a critical gap in robotics: how to make teams of robots navigate and explore buildings during fires when visibility is poor and conditions are dangerous. Current multi-robot Navigation & LocomotionNavigationMoving through an environment toward a goal. systems work well in normal indoor settings but fail badly when smoke blocks cameras and heat warps sensors. VULCAN combines vision-language models (AI that understands both images and text descriptions) with multi-agent Control & PlanningPlanningFiguring out what the robot should do before or during movement. to help Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. teams explore fire disaster sites more effectively. For developers, this means: (1) you can use large language models to help robots understand hazardous environments, (2) coordinating multiple robots becomes more robust when you fuse Modern Robot LearningMultimodalUsing more than one type of input, like vision, language, touch, or proprioception. Perception & SensingSensorA device that provides information about the robot or its environment. data and language understanding, and (3) there's now a realistic fire-disaster simulator (based on Habitat-Matterport3D) to test your multi-robot systems under harsh conditions.

WHY DEVELOPERS SHOULD CARE

This paper addresses a critical gap in robotics: how to make teams of robots navigate and explore buildings during fires when visibility is poor and conditions are dangerous. Current multi-robot Navigation & LocomotionNavigationMoving through an environment toward a goal. systems work well in normal indoor settings but fail badly when smoke blocks cameras and heat warps sensors. VULCAN combines vision-language models (AI that understands both images and text descriptions) with multi-agent Control & PlanningPlanningFiguring out what the robot should do before or during movement. to help Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. teams explore fire disaster sites more effectively. For developers, this means: (1) you can use large language models to help robots understand hazardous environments, (2) coordinating multiple robots becomes more robust when you fuse Modern Robot LearningMultimodalUsing more than one type of input, like vision, language, touch, or proprioception. Perception & SensingSensorA device that provides information about the robot or its environment. data and language understanding, and (3) there's now a realistic fire-disaster simulator (based on Habitat-Matterport3D) to test your multi-robot systems under harsh conditions.

LIMITATIONS

The main limitation to check is whether the claimed behavior holds outside the paper's reported setup. That means testing across different Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. embodiments, scenes, objects, and data distributions.

WHAT COMES NEXT

The practical next step is independent reproduction with clear baselines, ablations, and stress tests. For a developer, the useful follow-up is to map the paper's computer vision assumptions onto a concrete Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. stack, then test the smallest version of the method that could run end to end.

RELATED PAPERS