Formal Verification of Autonomous UAV Behavior for Inspection Tasks Using the Knowledge Base System IDP

Abstract

Unmanned Aerial Vehicles (UAVs) have become useful tools in industries. In this paper, we verify the behavior of an autonomous UAV executing an inspection task. More specifically, we look into the use of the knowledge base system IDP as a verification tool. We propose an approach for the modeling and verification of the safety-critical UAV and its environment in IDP. The methodology and modeling choices that are beneficial for the performance of the verification task and the readability of the model are denoted. We identify the need for discrete domains and investigate the consequences. Verification is successfully achieved using both Bounded Model Checking (BMC) and Invariant Checking (IC).