Deep Learning in Mobile Device Sensors
Surabhi Lingwal, Puneet Kumar Garg, Banit Negi, Jitendra Singh Rauthan · Apple Academic Press eBooks · 2025
Deep learning, a subset of artificial intelligence with large computational power, is a rapidly emerging technique for human activity recognition through mobile device sensors. Human activity recognition encompasses the study of human behavior, the interaction of humans with computers, and ubiquitous computing. Today, wireless network sensors are gaining importance due to low-power integrated circuits and are thus increasingly employed in the development of efficient, robust, and affordable wearable devices that can capture data in real time and transmit it for longer durations. Mobile sensor devices are employed in human activity recognition considering human physiological sensing along with other environmental factors. Human activity recognition using mobile and wearable sensor devices is a growing research area that involves feature extraction and image classification using deep learning. Mobile phones and wearable-based sensors are gaining importance for human activity identification as they are affordable, ubiquitous, unpretentious, portable, and allow ease of use and installation. Mobile sensor devices are location-independent, making them cost-effective compared to other methods based on wireless signals, allowing for easy deployment without causing any health hazards from radiation. This chapter presents 70 various deep learning techniques and their applications for mobile device sensors, along with the challenges encountered when applying mobile sensors to different types of datasets.