Machine learning and deep learning have become essential tools in many sectors. Ground transportation systems, for example, use predictive algorithms to enhance travel flows, anticipate congestion, and optimize fuel consumption. Now, these advancements are taking to the skies.
Significant progress in AI navigation is transitioning from research labs to practical applications, promising improved reliability, safety, and efficiency in aeronautics.
Inertial AI Navigation for UAVs in GPS-Compromised Environments
The reliance on GPS signals is a given for most UAV operations, yet these signals can be disrupted during natural disasters or military scenarios. Bavovna.ai, a participant in the US Air Force Labs Mass Challenge acceleration program, is creating an AI-powered Positioning, Navigation, and Timing (PNT) system suitable for aerial, surface, and subsurface vehicles.
Bavovna’s inertial navigation system leverages sensor fusion and pre-trained machine learning and deep learning algorithms for autonomous operation. Built with robust core electronics, it resists common electromagnetic warfare threats and is a low-SWaP (size, weight, and power) solution, ideal for various UAV models and Class II aerial vehicles.
In tests, the Aurelia X6 Max multicopter demonstrated the ability to fly autonomously without remote control, GPS, or any other communication, gather location data, and return home. Bavovna aims for a positioning error margin of just 0.5%, even on complex journeys up to 30 miles (48 km). The team is expanding the system’s capabilities to include signals intelligence, mine detection, automatic target engagement, and security surveillance.
AI Copilots for Commercial Aviation Flights
Modern aircraft are equipped with advanced autopilot systems that assist pilots in controlling altitude, course, thrust, and navigation. However, pilots often face an overwhelming number of alerts and system interfaces that demand their attention.
NASA has identified 34 different activities that can distract or preoccupy pilots, ranging from communication tasks to searching for visual meteorological conditions (VMC) traffic, which can lead to human errors and potentially hazardous situations.
Air-Guardian, a new initiative from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), seeks to enhance the human-machine interface (HMI) of existing autopilot systems to ensure safer flight operations. This AI system uses eye-tracking to detect pilot distractions and “saliency maps” to interpret aircraft behavior.
Employing a continuous-depth neural network model, the copiloting system can recognize early signs of risks and assume control when necessary. In trials, the Air-Guardian system has reduced flight risks and improved navigational success.
Advanced Airflow Traffic Management to Alleviate Congestion
Unexpected events, such as adverse weather conditions, can significantly impact air traffic, causing congestion in specific sectors of the air navigation space. This congestion can ripple through the entire network, leading to widespread flight delays.
The ASTRA project, co-funded by the European Union and led by the Universita Ta Malta, aims to mitigate these issues through AI-enabled tactical hotspot prediction and resolution. The project’s goal is to enhance the prediction of air traffic congestion areas an hour in advance and provide optimal resolution strategies to traffic controllers.
The ASTRA project will train its prediction algorithm on historical data from 2019 to the present, provided by the EUROCONTROL organization. The AI navigation system will offer Flow Management Positions (FMPs) prescriptive scenarios to optimize traffic flow, ensuring safety, efficiency, reduced fuel consumption, and minimized environmental impacts.
AI’s potential in aerial navigation is vast, and further advancements in sensor fusion, AI-powered PNT, and aerial traffic management are expected in the coming years.
