DRACo-SLAM: Distributed Robust Acoustic Communication-efficient SLAM for Imaging Sonar Equipped Underwater Robot Teams (2210.00867v1)

Abstract: An essential task for a multi-robot system is generating a common understanding of the environment and relative poses between robots. Cooperative tasks can be executed only when a vehicle has knowledge of its own state and the states of the team members. However, this has primarily been achieved with direct rendezvous between underwater robots, via inter-robot ranging. We propose a novel distributed multi-robot simultaneous localization and mapping (SLAM) framework for underwater robots using imaging sonar-based perception. By passing only scene descriptors between robots, we do not need to pass raw sensor data unless there is a likelihood of inter-robot loop closure. We utilize pairwise consistent measurement set maximization (PCM), making our system robust to erroneous loop closures. The functionality of our system is demonstrated using two real-world datasets, one with three robots and another with two robots. We show that our system effectively estimates the trajectories of the multi-robot system and keeps the bandwidth requirements of inter-robot communication low. To our knowledge, this paper describes the first instance of multi-robot SLAM using real imaging sonar data (which we implement offline, using simulated communication). Code link: https://github.com/jake3991/DRACo-SLAM.

Distributed Robust Acoustic Communication-efficient SLAM for Underwater Robotics

The paper "DRACo-SLAM: Distributed Robust Acoustic Communication-efficient SLAM for Imaging Sonar Equipped Underwater Robot Teams" outlines an innovative approach to tackling the challenges faced by multi-robot systems operating underwater. It presents a novel distributed simultaneous localization and mapping (SLAM) framework, specifically designed for teams of underwater robots equipped with imaging sonar. The framework significantly advances navigation and mapping capability in scenarios where conventional methods relying on GPS or high-bandwidth communication are infeasible.

Overview and Methodology

The authors propose a system that leverages imaging sonar data to establish a collaborative mapping and localization framework where robust state estimation and low bandwidth communication are paramount. Traditional SLAM approaches often depend on direct encounters or inter-robot ranging for data fusion. In contrast, this research introduces an elegant solution that circumvents these requirements by using pairwise consistent measurement set maximization (PCM) and scene descriptors to ensure robustness to erroneous loop closures without relying on direct encounters.

The paper explores the challenges of maintaining accurate state estimates in GPS-denied underwater environments. These constraints necessitate reliance on sonar technology, which, while effective, presents issues such as low resolution and noise that complicate robust state estimation and mapping. DRACo-SLAM addresses these concerns by efficiently managing inter-robot communication to facilitate multi-agent coherence, where only essential data is transmitted to infer the relative states.

Results and Implementation

The proposed framework was validated using two real-world datasets, featuring different robot team configurations. The evaluations evidenced that DRACo-SLAM effectively reduced the communication overhead while maintaining high precision in trajectory estimation. The authors meticulously measured inter-robot error metrics, noting mean absolute errors in the range of 1.17 to 7.25 meters, demonstrating significant improvements in state estimation accuracy. Notably, the network utilization was strategically reduced to an average of 161-337 bits per second, emphasizing the system's communication efficiency, which is crucial in low-bandwidth underwater environments.

Implications for Future Work

The success of DRACo-SLAM paves the way for a new class of underwater robotic systems capable of cooperative operations without the need for pre-determined configurations or explicit inter-agent synchronization. The ability to achieve reliable SLAM under significant environmental and communication constraints represents a substantial forward step for underwater robotics.

The practical implications extend to various applications such as environmental monitoring, infrastructure inspection, and search and rescue missions. Theoretically, the integration of robust indirect loop closure mechanisms could inform advancements in multi-agent systems beyond the aquatic domain, fostering more sophisticated collaboration strategies in terrestrial and aerial robotics.

Conclusion

In conclusion, DRACo-SLAM offers a robust, efficient, and viable solution for distributed SLAM in challenging underwater environments, setting a precedent for future developments in collaborative robotic systems. This research underscores the importance of optimizing communication and leveraging novel data association strategies to enhance the functionality and autonomy of multi-robot teams, signaling promising future avenues in AI and robotics research.