Across many parts of the developing world, overstretched waste management systems mean that plastics frequently end up in drains, rivers, and lagoons. Over time, these materials pollute ecosystems, threaten marine life, and disrupt the livelihoods of communities that depend on these water bodies.
At Ashesi University, the Aquabot research project, being undertaken by Dr. Stephen K. Armah, Joseph Kwabena Fosu Okyere, Gabriel M. Owusu and Peter Lawerh Kwao is exploring how technology might offer a response. The initiative focuses on designing and developing a low-cost Autonomous Surface Vehicle (ASV) capable of detecting and collecting floating plastic waste from water bodies.
The vehicle is powered by a compact computer, the Raspberry Pi 4, which serves as the system’s central processing unit. Chosen for its processing capacity and ability to run lightweight object-detection models; the device also supports connections to several external components required for the system operation. In effect, it coordinates communication between the sensors and other hardware installed on the vehicle. These components include a camera, a Global Positioning System (GPS), and ultrasonic sensors that help the vehicle observe the water surface, identify plastic waste, detect obstacles, and navigate safely.
Before constructing the prototype, a simulation exercise was conducted to test the design. The results showed that the vehicle displaced only a small volume of water, suggesting it would remain buoyant and stable during operation. The simulations also indicated that the combined weight of the vehicle and its electronic components would not compromise the system’s structural integrity.
Guided by these findings, a prototype was built using a catamaran frame- a configuration with two parallel hulls that improves stability on water. The hulls, which form the vehicle’s main structure, were designed to balance weight, buoyancy, and acceleration. After being smoothed and epoxy-coated to protect the wood, they were joined using an aluminum frame. A conveyor belt system was then installed between the hulls to collect floating waste. Four ultrasonic sensors were also added. Three help the vehicle detect obstacles and avoid collisions, while the fourth monitors plastic waste levels in the collection container, using green and red LEDs to indicate whether the container is within capacity or full.
Power for the system was supplied through two sources. A 10 Ah power bank powered the Raspberry Pi, while lithium-ion batteries run the remaining components. The Raspberry Pi also operated a computer vision model that analyzed live camera feeds to detect floating plastic and guided the vehicle toward it for collection. A web-based controller interface was created, enabling users to remotely control the ASV, access live camera feeds, and track sensor readings. The interface supports two modes of operation: a manual mode that allows users to steer the vehicle directly, and an autonomous mode that uses computer vision and control algorithms to locate and collect plastic waste independently.
Subsequent tests of the prototype, conducted in a swimming pool and the Korle Lagoon, revealed that connectivity issues occasionally disrupted remote control, and the vehicle did not always accurately identify waste. As a result, the vehicle sometimes remained idle during operation. To address this, a roaming feature was introduced. It allowed the vehicle to move randomly across the water surface if no plastic waste was detected and no user input was received after 5 minutes. Once plastic is detected, the system automatically switches back to autonomous mode and begins collecting the waste.
Overall, the project demonstrated that a low-cost ASV can collect plastic waste from water bodies. However, recommendations for further improvements were made. These include incorporating waypoint navigation so users can define specific routes for the vehicle, as well as adding pH and gas sensors to monitor water quality and provide insights into water pollution levels.
By prioritizing affordability and adaptability, the project points toward more scalable technological solutions for tackling plastic pollution in water bodies.













