Search papers, labs, and topics across Lattice.
This paper introduces AERR-Nav, a novel Zero-Shot Object Navigation framework designed to improve robot navigation in unknown multi-floor environments. AERR-Nav dynamically adjusts its state using an Adaptive Exploration-Recovery-Reminiscing Strategy, enabling specialized responses to diverse navigation scenarios through transitions between exploration, recovery, and reminiscing states. Experiments on HM3D and MP3D benchmarks demonstrate state-of-the-art performance compared to other zero-shot methods, validating the effectiveness of the proposed strategy.
Robots can now nimbly navigate complex, multi-floor environments without prior training, thanks to a new strategy that dynamically switches between exploration, recovery, and memory recall.
Zero-Shot Object Navigation (ZSON) in unknown multi-floor environments presents a significant challenge. Recent methods, mostly based on semantic value greedy waypoint selection, spatial topology-enhanced memory, and Multimodal Large Language Model (MLLM) as a decision-making framework, have led to improvements. However, these architectures struggle to balance exploration and exploitation for ZSON when encountering unseen environments, especially in multi-floor settings, such as robots getting stuck at narrow intersections, endlessly wandering, or failing to find stair entrances. To overcome these challenges, we propose AERR-Nav, a Zero-Shot Object Navigation framework that dynamically adjusts its state based on the robot's environment. Specifically, AERR-Nav has the following two key advantages: (1) An Adaptive Exploration-Recovery-Reminiscing Strategy, enables robots to dynamically transition between three states, facilitating specialized responses to diverse navigation scenarios. (2) An Adaptive Exploration State featuring Fast and Slow-Thinking modes helps robots better balance exploration, exploitation, and higher-level reasoning based on evolving environmental information. Extensive experiments on the HM3D and MP3D benchmarks demonstrate that our AERR-Nav achieves state-of-the-art performance among zero-shot methods. Comprehensive ablation studies further validate the efficacy of our proposed strategy and modules.
