Eyes on orbit: Students use embedded AI to detect space debris in real time

As the number of satellites and debris in low Earth orbit rapidly grows, ensuring the safety and sustainability of space operations has never been more critical. A team of electrical and computer engineering students at the University of Manitoba tackled this challenge with their 2026 capstone project, combining off-the-shelf sensors, ingenious engineering and AI-driven software to detect space objects in real time. 

Capstone projects are a key part of the final year of an engineering degree and serve as a culminating academic experience. Students work in teams to solve complex, real world problems, applying the technical knowledge and practical skills they have developed throughout their studies. These projects often involve industry or research partners and require students to move through the full engineering process, from problem definition and design to testing, iteration and implementation.

An autonomous onboard detection system

Among the capstone teams, one group led by Professor Peng Hu stood out for focusing on a growing concern in modern aerospace systems. As low Earth orbit becomes increasingly crowded, collisions between satellites and debris are becoming more frequent. Each collision creates more fragments, compounding the risk and putting billions of dollars’ worth of space assets in danger.

There is also a measurable risk to air travel. A 2025 study by researchers at the University of British Columbia estimates a 26 percent chance that space debris could pass through some of the world’s busiest airspace in the coming year. While the likelihood of debris striking an aircraft remains low, by 2030 any given commercial flight could face roughly a 1 in 1,000 chance of encountering falling debris.

To address this challenge, the team developed an autonomous onboard detection system that uses an internal database and simulation to identify nearby objects, helping reduce the risk of collisions and improve overall space safety. 

The project team includes Che Carby, Isindu Muhandiram, Jaden Lomibao, Malcolm Gutierrez, Martin Atanasov and Mina Abdalmasih, with Isindu and Mina graduating this year. Each student contributed across hardware design, prototyping, software development and system integration. The team was supervised by Dr. Peng Hu, whose research in embedded AI for space object detection helped guide key design decisions, particularly in integrating AI with optical sensing.

Formally titled Integrated Optical Sensing System for Real Time Space Object Detection, with the sub title Onboard LiDAR Vision Fusion Object Detection System, the project centres on building a proof of concept system that combines cameras and LiDAR sensors. One challenge the students encountered was the limitation of off the shelf two dimensional LiDAR sensors.

They addressed this by designing a custom tilting mechanism that effectively converts the sensor into a three dimensional scanner. Another key challenge was achieving real time performance through the combined use of computer vision and LiDAR sensing, which required careful system integration and optimization. To support testing, the team also developed a conveyor based testbed that simulates the motion of space objects. These solutions highlight the team’s ability to adapt and innovate under constraints while also creating a foundation for further research and testing.

The team plans to publish their results and continue refining the testbed, contributing to broader efforts in space safety and sustainability. With further development, this approach could help reduce costs by offering a more affordable alternative to traditional high cost sensing systems, making advanced onboard detection more accessible across the space industry.

Beyond the technical outcomes, the project provided significant hands-on experience and professional growth. Students developed skills in AI, software, electronics and system integration while strengthening their abilities in teamwork, problem-solving and project management. 

Overall, this capstone project highlights the kind of practical, forward-thinking work being carried out by students at the Price Faculty of Engineering at the University of Manitoba. By combining off-the-shelf hardware with AI-driven software, the team shows that impactful solutions do not always require large scale or expensive systems. In many cases, constraints drive creativity and lead to stronger, more thoughtful designs.

More broadly, it reflects the purpose of the capstone experience itself. Students must carefully consider each design choice and its consequences, clearly define the problem, understand their stakeholders and balance competing priorities. In doing so, they demonstrate their ability to think independently and make informed engineering decisions. 

In essence, capstone represents one of the purest forms of engineering in the undergraduate experience. There is no single correct path, only the challenge of defining and building a solution from the ground up. It is this balance of creativity, responsibility and technical rigour that makes the experience a defining moment in a student’s education and strong preparation for professional practice.