No abstract

There is a growing variety of manned and unmanned aerial vehicles that utilize batteries as their primary power source. These vehicles are composed of a large variety of interacting components and sensors that are needed for safe operation and to carry out their respective missions. As their interactions, complexity, and numbers increase, the risk for anomalies such as degradation of components, sensor faults, and erroneous controls also increase. These anomalies pose significant risks for vehicles flying over densely populated areas or conducting critical missions. It is, therefore, crucial to detect and mitigate these anomalies. There exist several approaches for anomaly detection such as traditional rule or threshold-based methods, model-based approaches, supervised machine learning-based methods, and even unsupervised methods to detect different types of abnormal behaviors. These methods have inherent drawbacks such as lack of sensitivity, inability to detect previously unknown faults, not being robust to compromised innetwork information, or requiring sophisticated system models. To this end, we propose BDAV, a Battery-based Diagnosis for Aerial Vehicles. Assuming a vehicle’s battery is nominal, BDAV utilizes a system’s battery as root of trust to diagnose other vehicle subsystems. Simple machine learning models learn physical dependencies between battery measurements and other vehicle operational variables and an unsupervised algorithm to detect and identify anomalies.