Classification of Adversities in Urban Deployment of Internet-of-Things

Communication in urban deployment of low-power Internet-of-Things (IoT) systems faces a wide variety of adversities. A significant communication hindrance is caused by Cross Technology Interference (CTI). CTI originates from quite a few different types of wireless communication technologies such as BLE, Microwave Oven, LoRa etc. Mitigation strategies to overcome the issues due to an adversity depends on its exact source/cause. Thus, to achieve an appropriate mitigation the first important step is to classify and detect the adversities being faced by a link. Recent research works have explored that communication links in urban deployment of IoT-systems face problems also due to large moving vehicles passing over the roads. Fundamentally, presence of obstacles due to road traffic makes a low-power communication link weak. Thus, a traffic affected link can be easily confused as a weak link. In this work we first identify that, instead of considering a traffic-affected link to be a purely weak link, it should be considered to belong to a separate class. We propose a light-weight strategy to automate the process of differentiating between traffic-affected link and weak link in low-power IoT-systems. Through empirical analysis of outdoor link measurement data we show that our strategy can distinguish between weak link and traffic affected link with 95% of mean classification accuracy within 2 min of time under a quite busy vehicular traffic.