Located in Flagstaff, Arizona, Northern Arizona University (NAU) has been at the forefront of education for over 120 years. The university’s first graduating class in 1901 was of 4 women who received the necessary credentials to become educators themselves in the Arizona Territory. During the Great Depression, when education was seen as a luxury, NAU found a way to increase enrollment while other institutions struggled. Today NAU has the distinction of being ranked a High Research Carnegie R2 Doctoral University. Recently, the NAU’s Wireless Networking and Smart Health (WiNeSH) Lab run by Assistant Professor of Electrical Engineering and Computer Science Dr. Abolfazl Razi, has been gaining some attention due to the groundbreaking research being conducted.
In the WiNeSH Lab, Dr. Razi is working on developing predictive modeling structures through the development of Artificial Intelligence (AI) wireless protocols. These AI systems are being applied to biomedical and molecular biology systems and aerial systems. Dr. Razi’s lab has received millions in funding from many scientific back initiatives over the last 5 years. This past year he was granted $480,000 from the National Science Foundation for a project called “Proactive Inverse Learning of Network Topology for Predictive Communication among Unmanned Vehicles”. As described by Dr. Razi, “We are currently developing a new generation of AI-enabled wireless networking protocols for unmanned aerial system (UAS) by incorporating the predicted network topology into decision making at different layers of communication protocols. The goal is [sic] design new machine learning tools that models and predicts network status change, user behavioral trends, and traffic mobility in order to accommodate predictable events by taking early decisions.”
Dr. Razi is building a system in which a drone can be aware of its topology, the mathematical space it takes up in an environment. Drones that can take in their surroundings to avoid collisions are becoming the next big thing in drone development. Dr. Razi’s concept goes one step beyond sensing an environment. The code he is working on is an AI system that allows a drone to be aware of itself and the environment it encounters. As our society becomes more reliant on drones, they need to be made safer for mass use. Currently, drone progression is being held back by the fact that for them to operate autonomously, they need to have total situational awareness.
For example, when a police officer or firefighter uses a drone in an emergency, the drone must be able to help the officers rather than cause a greater emergency. There is always the concern that a drone can malfunction and crash. This is why emergency teams need to have officers dedicated to the sole purpose of piloting the drone and will only have 1 drone in services at a time. If officers were able to use multiple drones, without fear of them malfunctioning, the drones could make their jobs far more efficient and safe. If these same drones could operate autonomously, emergency workers could work even more efficiently. The same concept applies to companies like Amazon that wish to enable mass swarms of drones as delivery vehicles.
“When we have hundreds of drones with limited communication ranges flying together, we need to keep connectivity and information flow uninterrupted,” Dr. Razi said. “The focus of this project is to enable UAVs to monitor themselves and each other, taking into account different scenarios.” The research that Dr. Razi is conducting will enable drones to perform as a human would, aware of its spatial parameters and other objects it may encounter. These other encounters could be buildings, wires, wildlife, or other drones. Dr. Razi’s code will give drones AI capabilities to do all of this autonomously, which is of particular importance to the use of drone swarms.
Drones being used in a swarm have to be part of a shared network. This network relates all mission objectives to the swarm. Each drone has a place within the swarm that relates to the others around it. If a single drone in the swarm has a problem, the rest of the drones in the swarm will be able to react accordingly. There is no need to abort a mission, the remaining drones simply compensate for the anomaly and carry out their programmed objectives. As Dr. Razi explained, “We want other drones to be able to analyze the trajectory and identify misbehavior or misconduct, or even interference of an outside drone and diagnose the problems within a network.”
Ultimately, drones with AI predictive capabilities will be more independent of human operators. If a drone can predict what will happen as it enters an environment, it can react accordingly. For example, if firefighters are using drones to build a plan of action in a forest fire, the drones can operate autonomously while the firefighters tend to other critical tasks. If a tree branch starts to drop, the drone can sense and avoid it without interrupting its mission of collecting data. This data can then be wirelessly transmitted to a mission commander so that firefighters can safely do their job. Drones have the power to help people in countless ways. Once drones can think and react with AI, much like how a human would, they will become even more beneficial to those who use them.