Edge Robotics: Edge-Computing-Accelerated Multi-Robot Simultaneous Localization and Mapping

Nancy J. Delong

Simultaneous Localization and Mapping (SLAM) is an approach that aims to simultaneously build a graphic map and observe the agent’s locale in an unknown surroundings, typically making use of the edge robotics concept. A modern paper, published on arXiv.org, proposes to leverage the emerging edge computing paradigm to conduct multi-robotic laser SLAM in low latency.

Edge robotics is based on the principle of edge computing

Edge robotics is dependent on the principle of edge computing. Impression credit history: asawin by means of Pxhere, CC0 Public Area

Edge computing makes use of vicinal computing assets in actual physical proximity to finish units to shorten info conversation length, reduce offloading transmission delay, and allow for the highly developed high-quality of companies. The proposed layout demonstrates that migrating SLAM workloads from robots to edge servers can properly augment the robots’ processing functionality.

It is also revealed that merging a subset of regional maps at the edge shrinks info size and lowers conversation fees. The simulation of the approach demonstrates its usefulness, and a realistic prototype on a few robots verifies its feasibility and validity.

With the huge penetration of sensible robots in multifarious fields, Simultaneous Localization and Mapping (SLAM) strategy in robotics has captivated growing notice in the neighborhood. Still collaborating SLAM in excess of various robots still stays challenging owing to performance contradiction in between the intense graphics computation of SLAM and the restricted computing functionality of robots. Even though common options resort to the impressive cloud servers performing as an exterior computation supplier, we show by actual-earth measurements that the significant conversation overhead in info offloading prevents its practicability to actual deployment. To deal with these challenges, this paper promotes the emerging edge computing paradigm into multi-robotic SLAM and proposes RecSLAM, a multi-robotic laser SLAM process that focuses on accelerating map building system beneath the robotic-edge-cloud architecture. In distinction to typical multi-robotic SLAM that generates graphic maps on robots and totally merges them on the cloud, RecSLAM develops a hierarchical map fusion strategy that directs robots’ uncooked info to edge servers for actual-time fusion and then sends to the cloud for world merging. To optimize the all round pipeline, an successful multi-robotic SLAM collaborative processing framework is released to adaptively optimize robotic-to-edge offloading customized to heterogeneous edge resource disorders, in the meantime making sure the workload balancing among the edge servers. Extensive evaluations show RecSLAM can attain up to 39% processing latency reduction in excess of the condition-of-the-artwork. Besides, a evidence-of-concept prototype is formulated and deployed in actual scenes to exhibit its usefulness.

Exploration paper: Huang, P., Zeng, L., Chen, X., Luo, K., Zhou, Z., and Yu, S., “Edge Robotics: Edge-Computing-Accelerated Multi-Robotic Simultaneous Localization and Mapping”, 2021. Url: https://arxiv.org/abs/2112.13222


Next Post

Did you hear the one about the swimming worm?

Experts at Cold Spring Harbor Laboratory (CSHL) began the NeuroAI Scholar program to entice post-masters stage and postdoctoral scientists with an interest in combining artificial intelligence (AI) with neuroscience. Related Posts:10 Trends Accelerating Edge Computing AI design of a swimming worm. Image credit history: CSHL (still impression from the YouTube movie) A […]