Indoor Autonomous Navigation with EdgeSLAM

Contemporary approaches to indoor autonomous navigation heavily depend on pre-established extensive location systems equipped with detailed maps and require a significant infrastructure setup. This chapter introduces Pair-Navi, an innovative system that bypasses these necessities by leveraging the path information shared by an initial user (i.e., the leader) to guide subsequent navigators (i.e., followers) using Peer-to-Peer (P2P) technology. Utilizing the capabilities of visual SLAM accessible on today’s smartphones, our system addresses the challenges posed by environmental dynamics and the computational demands that usually limit real-time execution. We enhance system resilience against variable environmental conditions by filtering out transient elements, focusing solely on stable, unchanging features. For seamless real-time operation on handheld devices, Pair-Navi strategically segregates and rearranges the intertwined components of SLAM for both leaders and followers. Demonstrated on widely available smartphones across varied architectural environments and tested against two renowned benchmarks (TUM and KITTI), Pair-Navi showcases a near-instant navigation success rate of 98.6%, with effectiveness slightly reduced to 83.4% after 2 weeks, significantly surpassing existing alternatives by over 50%. As a genuinely infrastructure-independent solution, Pair-Navi illuminates the path forward for efficient and practical indoor navigation for mobile consumers.