Zero Shot Object Navigation

Zero-shot object navigation (ZSON) focuses on enabling robots to locate and navigate to unseen objects—objects not present during the model's training—within unfamiliar environments. Current research emphasizes leveraging vision-language models (VLMs), often incorporating large language models (LLMs) for reasoning and instruction understanding, to achieve this, with approaches ranging from creating semantic maps to employing object-centric representations and pixel-based navigation. This field is significant for advancing embodied AI, pushing the boundaries of robot autonomy and potentially impacting applications in robotics, assistive technologies, and search-and-rescue operations.