UAV path optimization with an integrated cost assessment model considering third-party risks in metropolitan environments

Various applications of unmanned aerial vehicles (UAVs) in urban environments facilitate our daily life and public services. However, UAV operations bring third party risk (TPR) issues, as UAV may crash to pedestrians and vehicles on the ground. It may also cause property damages to critical infrastructures and noise impacts to the public. Path planning is an effective method to mitigate these risks and impacts by avoiding high-risk areas before flight. However, most of the existing path planning methods focus on minimizing flight distance or energy cost, rarely considered risk cost. This paper develops a novel flight path optimization method that considers an integrated cost assessment model. The assessment model incorporates fatality risk, property damage risk, and noise impact, which is an extension of current TPR indicators at modeling and assessment levels. To solve the proposed integrated cost-based path optimization problem, a hybrid estimation of distribution algorithm (EDA) and CostA* (named as EDA-CostA*) algorithm is proposed, which provides both global and local heuristic information for path searching in cost-based environments. A downtown area in Singapore is selected for the case study. Simulation results demonstrate the effectiveness of the developed cost-based path optimization model in reducing the risk cost. The statistical analysis for 100 sampled environments also shows the reliability of the proposed method, which reduced the cost by [42.64%, 44.15%] at 95% confidence level.