Improving Smart Cities Safety using Sanity-Check Deep Neural Network Algorithm

Nowadays, in a smart city, unmanned aerial vehicles (UAVs) have widespread usage to manage citizens' requirements. Due to the widespread usage of UAVs, the safety of citizens has become a great concern as UAVs are also used for several malicious activities to spread vulnerabilities or carry explosives. Detection of UAVs is the most important step in keeping citizens safe. The hardware-level fault has a critical impact on deep neural network (DNN) model accuracy. The purpose of our study is to detect hardware-level faults in the DNN model and improve model accuracy. In our work, we have used the sanity-check DNN. Our implemented function improves the DNN accuracy for UAV detection. Our study improves the safety of citizens.