Overview

Object-goal navigation asks a robot to locate and reach a specified object, often from a short language instruction and with only partial views of its surroundings. Approaches include reinforcement-learned policies, semantic maps and frontier exploration, and vision-language or language-model agents that plan hierarchically; many also use spatial memory to avoid repeated searches. Open-vocabulary targets, persistent maps, and real-robot tests are broadening the task beyond fixed object categories and static simulations. A central challenge is balancing reliable target grounding and safe, efficient movement with limited compute, especially when objects move, scenes change, or the robot must verify that it has truly reached the target.

Latest papers

CardsList

93 papers11 papers in the last four weeks, down 15% on the four weeks before. 0.1% of all new papers.

  1. Learning Category-level Last-meter Navigation from RGB Demonstrations of a Single-instance

    Dec 11, 2025Tzu-Hsien Lee, Fidan Mahmudova, Karthik DesinghRobot Skill LearningRobot Navigation

  2. What Matters in RL-Based Methods for Object-Goal Navigation? An Empirical Study and A Unified Framework

    Oct 2, 2025Hongze Wang, Boyang Sun, Jiaxu Xing +5Robotic RLRobotic Perception

  3. Open-Vocabulary Object-Goal Navigation by Generalizing Semantic Mapping with Dense CLIP

    Jul 12, 2024Meng Wei, Chenyang Wan, Tai Wang +6Embodied NavigationVision-Language Navigation