Drones have come a long way in the last decade. What was once seen primarily as a military weapon is now considered a tool that can assist in countless commercial enterprises. As of 2021, the global commercial drone market was valued at $20.8 billion, with expectations to exceed $100 billion by 2023. Drones are being used in research and development, logistics, construction, public safety, inspections, real estate, agriculture, filmography, and more. Even the hobby drone market has grown, with an estimated market value of $3.26 billion by the year 2025. However, one factor holding back the full integration of drones in shared airspace is the development of affordable anti-collision software. Solving this issue is one of the primary objectives of the Federal Aviation Administration to ensure the full potential of drone use in today’s ever-evolving society.

A professor in the Department of Mechanical Engineering at the School of Engineering of Gipuzkoa – Donostia, University of Basque Country (UPV-EHU) in Spain, recently published a paper on how to prevent such collisions from happening. Professor Julián Estévez is an industrial engineer with a PhD in computer engineering. As a researcher with the Computational Intelligence Group of UPV-EHU, his focus has been on the development of autonomous navigation for drones and robotic systems.

In the July 2024 issue of Volume 150 of the online publication ScienceDirect, Professor Estévez published a paper titled A Low-Cost Vision System for Online Reciprocal Collision Avoidance with UAVs. In the paper’s introduction, Professor Estévez states, “Due to the expected increase in the number of UAVs in future services, sky-sharing aircrafts and infrastructure still remain a significant operational-safety concern because the UAVs will be required to operate avoiding obstacles and interacting close to each other.” The paper then goes on to highlight the research project Professor Estévez undertook to develop a collision avoidance program for autonomous drones.

Professor Estévez and his team of research assistants used an off-the-shelf AR.Drone 2.0 from French manufacturer Parrot. This is a great drone to test out collision scenarios because it is lightweight, has a foam frame, and costs less than $100. The drone, like nearly all drones, is equipped with a camera, a key component in Professor Estévez’s project. By further enhancing the drone with onboard sensors and machine learning, Professor Estévez intended to “teach” the drone to use the information viewed through its camera to prevent collisions.

While some anti-collision programs require drones to wirelessly share information with each other, Professor Estévez envisioned a far simpler and data-secure system using the drone’s camera. He explained that the camera’s lens is divided into two halves, right and left. If the drone’s camera perceives an object through its camera that is mostly to the left, it will adjust course to the right to accommodate. If an obstacle is mostly to the left, the drone will adjust to the right. He related it to how, when a person is walking down the street and another person is approaching them, each person’s eyes will perceive the “obstacle” as being more to their right or left and naturally adjust their course to avoid walking into them.

To do this, Professor Estévez placed red cards atop the drones. The cards are too light to interfere with any weight or flight scenarios. But, when two or more opposing drones see the red card of another drone through its camera lens, the computer algorithm prevents a collision from occurring. “When the percentage of the color red on the screen increases, it means that the drones are approaching each other head-on. So when a threshold is exceeded, the robot knows that it has to perform the avoidance maneuver,” said Professor Estévez. “All this happens autonomously, without the human operator intervening. It’s a simple way to prevent collisions and can be performed by low-cost sensors and equipment.”

“This work is a small step towards fully autonomous navigation, without any human intervention, so that drones can decide which maneuver to perform, which direction to take, thus preventing collisions with each other or with other airborne obstacles,” Professor Estévez goes on to say. “If we assume that, in the future, our airspace will be much more populated by commercial services performed by these drones, our work is a small contribution in this respect.” As the drone industry expands and evolves, such breakthroughs are essential for ensuring that autonomous systems can operate harmoniously within increasingly crowded skies. With ongoing research and development, we move closer to a future where drones can navigate complex environments autonomously and safely, unlocking their full potential for a wide range of applications.

Previous Drone News: