Artificial Intelligence-based Functionality for Unmanned Aerial Systems

Artificial Intelligence-based Functionality for Unmanned Aerial Systems

In an article released in late 2016, the topic of Artificial Intelligence (AI), and the level at which AI will be or should be incorporated into UAV use, is discussed heavily with reference to drive and functionality for the autonomous system to exist (Karpowicz, 2016). The drive for a technologically superior product in the world of unmanned aerial systems could be to minimize costs and human-machine interface (HMI) requirements across a multitude of users and operators. Developmental and maintenance costs are relatively unknown as the finite position or task in which a system may use an autonomous decision making framework changes based on other variables such as operating environment, design mission, and end-user requirements (Karpowicz, 2016). 

Parallel circumstances can be drawn at a basic level to the requirements set in place to maneuver a driverless car, where a basic operating requirement may be solely moving from A to B given all the external influences/factors. Variables and factors that would hinder performance could be other drivers on the road, communication links and constraints to communications channels between unmanned systems, and the specific road conditions in which the vehicle is designated for travel (Karpowicz, 2016).. Similarly, this stance can be taken into the frame for an aerial vehicle, where other unmanned systems and manned systems alike are required to share the designated airspace for continuous and safe operation. Exploring concepts such as Sense and Avoid (SAA) which can be meant to serve as a secondary or tertiary measure for collision avoidance, is also an important part of the AI concerns for advancing UAV technology (Knight, 2017). Ensuring variables are adequately accounted for and researched to not only find risk areas, but preprogram risk mitigation measures into the AI logic is extremely important for AI integration into UAS.

The level at which a vehicle is designed to perform AI-based functions is certainly at the discretion of the user/operator, and should be clearly annotated in the end-user requirements and expressed in the preliminary design review. Whether a UAV is required to execute an entire electronic intelligence (ELINT) mission from takeoff, to data collection/relay, to landing and shutdown or simply have automated landing and terminal area procedures is up to the user. Design and engineering level efforts will be directly affected by these requirements, but as the skies become more congested with multitudes of unmanned vehicles, the need to effectively monitor, as opposed to control, these vehicles may rely heavily on AI-based advancements.

References:

Karpowicz, J. (2016, September 8). How Will Advances in AI Impact the Approach to UAV Technology?. Retrieved from http://www.expouav.com/news/latest/will-advances-ai-impact-approach-uav-technology/

Knight, W. (2017, January 4). 5 Big Predictions for Artificial Intelligence in 2017. Retrieved from https://www.technologyreview.com/s/603216/5-big-predictions-for-artificial-intelligence-in-2017/