Efficient Location-Based Tracking for IoT Devices Using Compressive Sensing and Machine Learning Techniques

Abstract

AbstractIn this chapter, a scheme based on compressive sensing (CS) for the sparse reconstruction of down-sampled location data is presented for the first time. The underlying sparsity properties of the location data are explored and two algorithms based on LASSO regression and neural networks are shown to be able to efficiently reconstruct paths with only ∼20% sampling of the GPS receiver. An implementation for iOS devices is discussed and results from it are shown as proof of concept of the applicability of CS in location-based tracking for Internet of Things (IoT) devices.