Abstract
For small unmanned aircraft systems (UAS) entering the very low-level (VLL) airspace, there is a wide range of VAS-external threats impacting energy consumption. This poses a great challenge for missions covering long distances or requiring extended flight times. This paper introduces a framework for estimating the energy consumption of UAS depending on operational and environmental factors. It supports in-flight decision-making processes by considering the energy demand and remaining energy capacity levels. The proposed framework follows a model-based approach that is based on vehicle performance and non-linear electric-based battery models. These are then integrated into a system for contingency management. For contingencies impacting the energy limits, the system assesses the repercussions and initiates landing mitigation procedures. The novel capabilities of the framework are implemented as Robot Operating System (ROS) modules within a Software in the Loop (SITL) simulation environment, which incorporates the open-source PX4 flight stack. With this, it is possible to assess the impact of operational (different velocities, increased payloads) and external factors (varying wind conditions) on the energy consumption. In particular, use cases representing operations over urban areas with off-nominal situations due to increased adverse wind conditions were tested. The simulation results show that using the presented framework, the UAS resolves Loss of Energy hazards much earlier compared with the built-in energy failsafe mechanisms and chooses a suitable landing zone based on the energy cost of reaching the landing location. The earlier activation of mitigation procedures leads to higher remaining energy reserves compared with the benchmark solution, increasing the safety of the mission. In order to validate the performance and accuracy of the developed energy consumption framework, real world test flights were carried out. The comparison of simulation and flight tests results shows that the model provides an accurate approximation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have