Autonomous Flight of High-Endurance UAVs to Monitor Powerlines

2.1 Previous Hardware and Software Weaknesses

Previous methodologies for tackling this issue had some key hardware and software weaknesses. In terms of hardware, there were a few crucial issues with the architecture of the system. UAVs in general are constrained to the amount of energy they can store with onboard batteries, or other energy storage mediums. Therefore, minimal power consumption is key for long endurance flights. In previous architectures, all computation was done on-board during the flight. This was unfavorable because the computation required for maneuvering and calculating paths for the UAV was substantial. Since the calculations were substantial, on-board computers needed extensive hardware, which would usually result in payload penalties for the UAV. Though embedded or lightweight computers would be a better choice in weight reduction, they are not powerful enough to support real time calculations and updates to help maneuver the vehicle.

In terms of software, there is a difficult and sensitive balance that must be struck between speed and computation intensity. The speed of path and obstacle avoidance calculations need to be rapid to reduce the chance of crashing the vehicle, while the computation intensity needs to be low to increase flight endurance. Newer architectures implement a ground control station (GCS), which can reduce on-board computation, but adds complexities of latency and security. If a UAV were to communicate with a GCS over the air, cyber attackers would be free to collect data between the two devices using a SDR (Software Defined Radio). Also, the communication protocol used cannot be too complex, as it needs to be exceptionally reliable. If the communication protocol is too computationally intensive, latency between the GCS and UAV increases which inherently increases the risk of crashing. If the communication protocol is simple, but not reliable, packet dropping yet again increases the risk of crashing. Thus, real time data transfer between the GCS and UAV not only requires a robust and reliable protocol, but one that is also secure.

2.2 Related Work

The study of (Motlagh et al., 2017) presents a case study on the power consumption reduction of video offloading from a drone to an MEC node for crowd surveillance. Although the techniques demonstrated are useful, they don’t provide performance gains or flight time benefits from this experiment. The research will instead focus on the flight time benefits of offloading image processing data to an MEC. The research will also be oriented towards powerline monitoring via a network of drones, which will thus develop a large sensor network.