An Exploration of Embodied Visual Exploration

Embodied computer vision considers perception for robots in novel, unstructured environments. Of particular importance is the embodied visual exploration problem: how might a robot equipped with camera scope out new environment? Despite progress thus far, many basic questions pertinent to this problem remain unanswered: (i) What does it mean an agent explore its environment well? (ii) Which methods work well, and under which assumptions environmental settings? (iii) Where do current approaches fall short, where future seek improve? Seeking answers these questions, we first present taxonomy existing algorithms create standard framework benchmarking them. We then perform thorough empirical study of four state-of-the-art paradigms using proposed two photorealistic simulated 3D environments, architecture, diverse evaluation metrics. Our experimental results offer insights suggest performance metrics baselines exploration. Code, models data are publicly available.