ForagerSim tests map-based reasoning for robots in complex worlds
ForagerSim: Evaluating Embodied Map-Based Reasoning in Large-Scale Dynamic Worlds
RoboticsMachine Learning
Summary
It's hard for robots to think and navigate using maps like animals do, especially in big or changing places. The authors created ForagerSim, a computer platform that makes robots do tasks needing map skills in different settings, from mazes to open fields with changing conditions. They tried teaching robots using current methods and compared them to simple map-based approaches and humans. The results showed robots learned some simple tasks but struggled in bigger or trickier maps, where map-based thinking helped the most.
What this means in practice
- •For robotics engineers: Test robotic navigation strategies in simulated complex and dynamic environments that require map use.
- •For game developers: Design and evaluate AI agents that perform realistic search and exploration tasks in large virtual worlds.
Authors
Tomáš Musil, Martin Kotík, Michal Vavrečka, Karla Štěpánová, Sebastian Scherer, Martin Saska
Abstract
Animals build cognitive maps of vast environments, use them to perform map-based navigation, and might even be reusing the same circuitry used in spatial reasoning for creating and using maps of arbitrary task-spaces. In contrast, current embodied AI systems either rely on hand-engineered mapping pipelines, which break in the presence of unmodeled dynamics and are restricted to metric space, or on learned policies that rarely form map-like representations at all. Developing architectures that could learn to build and use map-like representations as flexibly as animals is, however, difficult because few embodied AI benchmarks actually require map-based reasoning. To address this, we present ForagerSim --- an interdisciplinary research platform and simulation environment focused on embodied tasks that require map-based reasoning, with scenarios ranging from mobile-robot search and rescue to animal foraging in both complex mazes and vast, potentially infinite open areas. ForagerSim is highly modular and allows modifying independent axes of difficulty related to perception (abstract 2D sensing to full RGB or LiDAR), scale (room-scale to multi-km scale), topology (open-fields to complex mazes), dynamics (smoke, debris, sensorimotor failures and others) and tasks (from simple 2D coverage to foraging-inspired well-alternation tasks). We compare reinforcement learning baselines with a simple engineered robotics stack and a human player on 8 example worlds which show how different difficulty axes affect existing methods. Our results show that, when sensing is simplified to 2D, learned policies can perform multi-object search and well-alternation tasks surprisingly well in simple topologies, but they fall short of human performance in larger or more complex-topology environments where map-based baselines perform well, suggesting that map-based reasoning is the bottleneck in these scenarios.