Experiences with deploying Compressive Sensing and Matrix Completion techniques in IoT devices

The Internet of Things (IoT) presents itself as a promising set of technologies for providing innovative smart applications in a number of domains, spreading from agriculture to buildings and industrial control. For providing smart applications in the various domains, large numbers of devices must be deployed within large areas, communicating wirelessly with each other. These IoT devices are mainly resource constrained, with limited computing capabilities and battery life. Furthermore, it is well-known that the most energy consuming operations in wireless devices is the wireless transmission. Thus, there is a strict requirement that to prolong the lifetime of these devices, the transmissions must be controlled. Additionally, communication security is a key issue in many smart applications, because, according to recent studies, there is a lack of secure communication protocols in most existing IoT devices. This paper aims to address these two main issues in the IoT world, providing a framework for lightweight encryption and compression of data using the Compressive Sensing theory, and discusses the experiences of the implementation of the framework on real world devices. Moreover, it is also well-known that the transmissions of wireless sensors are very sensitive to packet loss due to concurrent transmissions of other wireless protocols (i.e. WiFi). This paper also presents the implementation of a framework for applying Matrix Completion techniques in real world IoT devices.