SoftSense: Collaborative Surface-Based Object Sensing and Tracking Using Networked Coils

Object sensing and tracking using electric and magnetic fields allow intelligent interaction, automation, and adaptation in cyber-physical systems. Our approach, called Softsense, uses a software-defined collaborative sensing technique for detection of the type of object and where it is placed on a large surface. Unlike RF based sensing approaches that are generally dependent on channel conditions, capacitive sensing that detects only low conductive objects, and Qi-based inductive sensing that is effective at few mm in range, Softsense is designed to operate without the above limitations. Softsense is cost effective, low power and scalable, which allows extension over large surfaces. First, we introduce a dual-coil inductive sensing architecture based on nested coils, i.e., passive (outer) and active (inner) coils, for low-power contact-less sensing. A data driven support vector machine-based approach helps to classify different materials using the voltage readings obtained at the passive coil. SoftSense combines sensed voltage information from multiple different coils spread over the surface for collaborative sensing. We validate our design on a real sensing prototype with customized coils, fabricated sensing circuit, and a network software controller. Experimental results show that each sensing coil only consumes few milliwatts, i.e., 18x less than the inductive sensing and 15x less than classical magnetic resonance sensing, extends sensing depth to 3 cm, and enables coverage of large surface sensing.