Exploring Model Compression for Deep Machine Learning-Based Indoor Localization

Abstract

GPS technology has revolutionized the way we localize and navigate outdoors. However, the poor reception of GPS signals in buildings makes it unsuitable for indoor localization. Wi-Fi fingerprinting-based indoor localization is one of the most promising ways to meet this demand. Unfortunately, the vast majority of work in this field fails to address the challenges associated with deployability on resource-constrained embedded systems. In this chapter, we explore deep learning model compression and its relationship to accuracy in the context of indoor localization. Our analysis shows that our proposed framework, CHISEL, surpasses the best-known efforts in the field while maintaining embedded device localization resilience.