Compliant Robotic Arm for Underwater Scenarios (CRAUS)

Developing an intelligent system for detecting, tracking, and collecting underwater litter

Compliant Robotic Arm for Underwater Scenarios (CRAUS)

Developing an intelligent system for detecting, tracking, and collecting underwater litter

Title: Compliant Robotic Arm for Underwater Scenarios (CRAUS)

Funded by: C-DORE Seedcorn Funding

Dates: 2023

University of Plymouth staff: Dr Pablo Borja Rosales, Dr Dena Bazazian

The Compliant Robotic Arm for Underwater Scenarios (CRAUS) project explores the integration of robotics and computer vision to develop an intelligent system for detecting, tracking, and collecting underwater litter. Developed at the University of Plymouth in 2023, the project combines an underwater robotic arm with computer vision techniques to enable the automated identification and targeted collection of waste in marine environments. The computer vision system analyses underwater imagery to detect potential litter, localise its position, and trigger the robotic arm to track and grasp the identified objects. By bringing together robotic manipulation, visual perception, and intelligent decision-making, CRAUS demonstrates the potential of AI-driven technologies to support underwater clean-up operations and contribute to healthier marine ecosystems.

The primary objective of CRAUS is to investigate how computer vision and compliant robotic manipulation can work together to enable the detection and collection of underwater litter. The project aims to develop an integrated system in which computer vision techniques identify and localise waste in underwater images, providing the visual information required to guide the robotic arm towards the target. The robotic system is designed to track the detected object and perform the grasping action, demonstrating a coordinated approach to underwater perception and manipulation.

Through this interdisciplinary approach, the project seeks to advance intelligent robotic solutions for marine environmental monitoring and litter removal, while establishing a foundation for future research into autonomous underwater clean-up systems and their applications in marine conservation.

By combining computer vision with underwater robotic manipulation, we can move beyond simply detecting marine litter towards developing intelligent systems that can actively help remove it. CRAUS brings together visual perception, AI, and robotics to explore how technology can support cleaner oceans, protect marine ecosystems, and open new possibilities for sustainable marine science.

Dr Dena Bazazian
Lecturer in Robotics and Machine Vision

Marine litter plastic bag on the seafloor

Marine litter is a growing environmental challenge that threatens marine biodiversity, damages habitats, and risks marine organisms through ingestion, entanglement, and habitat degradation. Identifying and removing underwater waste can be particularly challenging because of limited visibility, changing lighting conditions, water turbidity, and the difficulty of accessing submerged environments.

Conventional clean-up approaches can be labour-intensive, time-consuming, and hard to scale, especially when waste lies on the seabed or in sensitive marine habitats. These challenges highlight the need for innovative technological solutions that combine reliable underwater perception with precise manipulation.

Integrating computer vision and robotics offers a promising approach by enabling systems to identify potential litter, locate it, and perform targeted collection, reducing the need for direct human intervention in challenging underwater environments.

CRAUS addresses these challenges by combining computer vision with an underwater robotic arm to investigate an intelligent approach to marine litter detection and removal.

The computer vision component processes underwater imagery to identify potential waste objects and estimate their locations, enabling the system to initiate the robotic arm's tracking and grasping sequence. Using this visual information, the robotic arm can move toward the target and attempt to pick it up, bringing automated detection, target tracking, and robotic manipulation together in a coordinated framework.

This integration of computer vision and robotics represents an important step towards developing more capable underwater systems that can assist with marine clean-up activities.

By exploring how intelligent visual perception can guide physical interaction with underwater objects, CRAUS contributes to interdisciplinary research at the intersection of computer science, robotics, and marine science, with the longer-term potential to support ocean-cleaning technologies, reduce the environmental impact of marine litter, and help protect marine biodiversity.

Centre for DecarbOnisation, Resilience and Energy Systems

We bring together leading researchers from across the University of Plymouth to develop innovative solutions to the national and global challenges of decarbonising energy systems, strengthening resilience to climate change and security risks, and supporting sustainable economic growth.

Discover the Centre for DecarbOnisation, Resilience and Energy Systems 