Robot fish work together to detect and analyse pollution

The robotic fish, built at Essex University for an EU ICT project dubbed SHOAL, have been designed to manoeuvre like fish in the natural world and work together to conduct real-time in-situ detection and analysis of pollution, which improves on the current method of sample collection and laboratory tests onshore.

SHOAL is managed by BMT Group and the consultancy’s Luke Speller believes this capability can cut the detection and analysis of pollutants from weeks to seconds.

The battery-operated fish, made with a wet hull design, contains an array of sensors designed to help them perform their mission.

Speller, SHOAL project leader and senior research scientist at BMT Group, told The Engineer that a sonar in the front of the fish aids navigation and obstacle avoidance.

He added that the sensors consist of four main electrodes for testing different pollutants including copper, lead and phenols plus general water quality parameters such as oxygen. Pressure sensors, battery monitors, a temperature sensor, and an AHRS (attitude heading reference system) are also installed.

Artificial intelligence (AI) created by BMT lets the fish operate independently and as a swarm to create 3D pollution maps. The AI also includes a battery monitoring system that can issue an alert when the fish’s battery needs recharging.

‘When the fish are low on battery they go back to the base station and inform the port they are ready to be recharged,’ said Speller via email. ‘In the future inductive charging underwater would be a great addition to the system.’

The 35kg robotic fish have a jointed tail and use an oscillatory fish-like motion to propel themselves through the water. There is also a weight inside the fish to allow them to dive and rise.

Speller explained that the hydrodynamic design allows the robot fish to move through the water effectively and that they are not likely to get caught up in weeds or debris.

Prof Huosheng Hu from Essex University’s School of Computer Science and Electronic Engineering added that the robotic fish make no noise, have good agility, and do not disturb the environment.

Speller conceded that the robotic fish have not yet fully achieved the efficiency of real fish, but they have shown that robotic fish in the real world environment are possible.

‘Further, the robotic fish inspire people and show people non-standard designs are possible,’ he said. ‘This will hopefully draw more people and provide opportunities for further advancements in robotics.’

Trials of the robotic fish are set to continue into June, and data collected during testing will be analysed thereafter.

Organisations such as port authorities and water companies are seen as potential customers, but the fish can also be put to use in other underwater domains such as security, diver monitoring or search and rescue at sea.

SHOAL is a consortium of six European organisations including the Tyndall National Institute, which supplied the chemical sensors; Strathclyde University, which provided hydrodynamic research; and Thales Safare, which supplied the communication network.