Backscattered Visible Light Sensing of Retroreflective Foils Utilizing Random Forest Based Classification for Speed and Movement Direction Determination and Identification of an Indoor Moving Object

Making the Internet of Things “green” has become a major research focus in recent years. The anticipated massive increase in the numbers of sensor and communication devices makes this endeavor even more important, resulting in various solution approaches ranging from energy harvesting to energy efficient routing schemes. In this work, we propose a system that can perform some of the main tasks of the Internet of Things, namely identification and sensing of an indoor moving object, by the means of visible light sensing in combination with off-the-shelf retroreflective foils, without the necessity to place any actively powered components on the object itself. By utilizing the supervised machine learning approach of random forest, we show that these two tasks can be fulfilled with up to 99.96% accuracy. Based on our previous findings in this regard, we propose some advancements and improvements of the overall system, yielding better results in parallel with an increased complexity of the system. Furthermore, we expand the number of performable tasks toward additional movement direction determination. The achieved results demonstrate the applicability of visible light sensing and its potentials for a “green” Internet of Things.