Feasibility Analysis of UAV Operations for Monitoring of Infrastructure Networks: a Risk-based Approach

Abstract

This paper proposes a method to analyze the feasibility of monitoring missions of infrastructure networks (eg. railways, roads, power lines, etc.) by Unmanned Aerial Vehicles (UAVs). The method consists in evaluating the risk at ground for people due to an UAV flying over each part of the network. Based on a risk categorization defined by probability and exposition time, several metrics are developed to analyze the feasibility of flights in terms of risk. More specifically, these metrics are related to the cumulative length of feasible flight segments and their connectedness. A decision making process is then defined to automatically classify all the lines of the network in a comprehensive and synthetic way. The whole method takes into account flight regulation and flight operation constraints as well as considerations relative to business interests in asset management of the infrastructure network. An illustration is proposed through a benchmark example on the French railway network.

DROSERA: A DROne Simulation Environment for Risk Assessment

“…Effect of wind is also clearly visible, which is very important for long range outdoor missions where wind conditions can be non negligible and may be very changing during the flight. The main problem of this approach is that Monte Carlo simulations are time prohibitive to be used in an integrated tool such as the one presented in this paper (DROSERA), which aims at evaluating the risk at each point of the flight trajectory, for long range missions, and possibly for a huge number of missions when considering inspection of infrastructure at a country scale Bertrand et al (2019). Therefore, a full data basis of ground impact probability maps has been generated "offline", considering a large set of possible values for parame-ters of influence (altitude, speed, wind direction and speed, etc.).…”

“…deployment costs) are accounted for in this process. More explanation on the definition of these bar graphs can be found in Bertrand et al (2019) as well as an example of their use in the analysis of flight trajectories for inspection of the French railways network by UAVs.…”