OA-Bug: An Olfactory-Auditory Augmented Bug Algorithm for Swarm Robots in a Denied Environment
Swarm robots achieve 96.93% search coverage in denied environments by mimicking animal cooperation.
A research team led by Siqi Tan has developed a novel navigation algorithm called OA-Bug (Olfactory-Auditory augmented Bug algorithm) that enables swarms of autonomous robots to operate effectively in denied environments. These are scenarios where traditional aids like Global Navigation Satellite Systems (GNSS), pre-built maps, data sharing between units, and central command processing are unavailable or jammed. Inspired by animal cooperation, the system uses two simulated sensory channels: olfactory (smell) for leaving persistent chemical-like trails and auditory (sound) for broadcasting real-time alerts. This bio-inspired approach allows simple robots to coordinate complex search patterns without relying on sophisticated individual intelligence or constant communication with a base.
The algorithm's performance was rigorously tested in a custom simulation environment, where it demonstrated a remarkable 96.93% coverage rate in search tasks. This represents a significant improvement over a comparable existing algorithm, the Sensor-based Gradient Bug Algorithm (SGBA). The research moved beyond simulation, with the team conducting successful proof-of-concept experiments on physical swarm robot platforms, confirming the real-world viability of the approach. The work, documented in a paper that has undergone multiple revisions since its 2022 arXiv preprint, was formally presented at the prestigious 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).
The core innovation of OA-Bug lies in its hybrid communication strategy. The olfactory component creates a decentralized, persistent memory of explored areas, preventing redundant searches. Simultaneously, the auditory component allows for immediate, localized coordination—such as alerting nearby robots to discovered obstacles or targets—enabling dynamic adaptation. This makes the swarm robust to individual failures and highly scalable, as the coordination logic is distributed. The method provides a practical pathway for deploying robot swarms in challenging real-world applications like disaster response, hazardous material inspection, or military reconnaissance, where infrastructure cannot be relied upon.
- Achieves 96.93% search coverage in simulations, outperforming the similar SGBA algorithm.
- Uses bio-inspired olfactory (persistent trails) and auditory (real-time alerts) signals for robot coordination.
- Enables operation in denied environments with no GPS, maps, data sharing, or central processing.
Why It Matters
Enables reliable deployment of robot swarms for search & rescue or inspection in GPS-denied, unstructured real-world environments.