The robot uses an AI system to analyze objects with ultrasound and cameras. It demonstrated its ability to pick up waste and bring it to the surface for the first time in the port of Marseille, France.
Ports worldwide are heavily impacted by marine litter, with items like e-scooters, bicycles, and fishing nets regularly found in harbor basins. For example, researchers in Dubrovnik found over 1,000 pieces of trash in just 100 square meters.
Autonomous waste collection is expected to offer a solution soon. The system includes an unmanned service boat, a drone, an underwater search robot, and the TUM diving robot. According to Dr. Stefan Sosnowski, the Chair of Information Technology Control at TUM, a cost-benefit analysis shows that this autonomous method is profitable for waste disposal at depths of 16 m or more.
TUM Diving Robot as Part of an Overall Concept
The process functions as follows: the service boat delivers power and data links to the underwater robots through a cable. It also transmits ultrasonic signals into the water to produce a basic ocean floor map. A specialized search robot roughly 50 cm long scans the seabed rapidly and effectively.
Using the search robot's data, the TUM submarine, powered by eight mini turbines, dives to the specific locations. It grabs the objects and uses a winch to load them onto an autonomous dinghy that serves as a floating waste container.
Among the features of the TUM underwater robot:
1. System Identifies Rubbish Objects and Displays Them in 3D
Since we first have to identify the rubbish and grasping objects requires a high degree of precision, we have a camera and sonar on board that enable orientation even in murky water.
Dr. Stefan Sosnowski, Chair, Information Technology Control, Technical University of Munich
Detecting underwater waste is challenging, mainly because there is very little image data of submerged objects available for training neural networks.
“That’s why the project partners have so far labeled over 7,000 images as objects that don't belong on the seabed,” said Sosnowski.
After the waste is recognized, the AI system transforms the images into 3D.
This is important for deciding where the object can be gripped securely.
Dr. Stefan Sosnowski, Chair, Information Technology Control, Technical University of Munich
2. Strong and Sensitive Gripper
The autonomous gripper developed by TUM for the diving robot is a giant, four-fingered hand with a volume of approximately 1 m3. It can apply a squeezing force of 4,000 newtons and grasp objects weighing up to 250 kg.
However, special sensors enable the gripper to gauge the exact amount of force to apply without causing damage. This prevents it from breaking delicate items like plastic buckets or shattering glass bottles.
3. A Cable Connects the Robot to the Power Supply and Data Network
Even though the TUM boat operates autonomously, researchers keep it connected by a cable for several practical reasons. An onboard battery would only power the boat for about two hours, so the cable provides a continuous power supply. It also slightly enhances the performance of the on-board AI system. Lastly, the cable serves a dual purpose as a rope, allowing the team to pull heavy objects from the seabed to the surface.
4. Buoyancy Foam Keeps the Diving Robot in Place
The 120 kg submarine is surrounded by buoyancy foam that keeps it neutrally buoyant, suspended when its mini turbines are unused. This design allows the underwater robot to move freely and maintain a precise course.
“This is important for approaching objects precisely,” adds Dr. Stefan Sosnowski.