Investigating factors influencing UAS task automation in dynamic environments

MSc (Computer Science), North-West University, Potchefstroom Campus

UAS (Unmanned Aerial System) technology and capabilities have greatly improved which means that there are increased application possibilities and various industries in diverse fields are starting to show interest in UAS implementations. The problem with widespread implementations is that not all industries have a skilled UAS operator or pilot with a license. More often than not, these industries are not necessarily interested in flying the UAS, but only in the data collected or specific task undertaken by the UAS. Thus, it is more important for UASs to perform automated tasks without placing a huge burden on the operator or pilot, to be able to focus on the task at hand. The problem is that most commonly available UASs that support autonomous tasks are influenced by external factors that are not always compensated for. In this study, external factors influencing the UAS task environment are investigated in literature to determine the most significant factor. The most significant factor that influences UAS task automation and performance is identified as wind therefore experiments were conducted to illustrate its effects. The UAS performed automated flight experiments to predetermined waypoints along a chosen flight path. The data collected from the UAS flight logs, as well as the weather station was analysed to illustrate and determine the influence of wind on the UAS flight, and how it affected the accuracy of the automated flight in terms of deviation from the planned flight path between the waypoints, the time it took to reach the waypoints, as well as the drift sustained when hovering. These results are used to make recommendations for conducting autonomous UAS flights in dynamic environments with wind, to improve the accuracy and applicability in diverse applications.

Masters