Autonomous underwater vehicle

FlatFish being recovered after a test run. (Source: Florian Cordes, Ground Truth Robotics GmbH)
FlatFish being recovered after a test run. (Source: Florian Cordes, Ground Truth Robotics GmbH)
Scientific Leader:
Project leader:
Contact person:

The FlatFish project is a venture undertaken by the German Research Center for Artificial Intelligence (DFKI), the Brazilian Institute of Robotics (BIR) and Shell. It aims at designing an autonomous underwater vehicle (AUV) for repeated inspections of oil & gas subsea structures whilst being submerged for extended periods of time.

Duration: 26.11.2013 till 30.06.2017
Donee: German Research Center for Artificial Intelligence GmbH
Partner: FIEB/SENAI, Shell Brazil
Application Field: Underwater Robotics
Related Projects: CUSLAM
Localization and mapping in confined underwater environments (09.2009- 07.2012)
Inspection of offshore pipelines (12.2010- 07.2011)
Europa-Explorer (12.2012- 04.2016)
Related Robots: DAGON
Autonomous Vehicle for Aquatic Learning, Operation and Navigation
Exploration AUV for long-distance-missions

Project details

FlatFish in the salt water basin above the pipeline mockup. (Source: Florian Cordes, Ground Truth Robotics GmbH)
FlatFish returns to the base vessel after a long day of tests in the Atlantic Ocean. (Source: Marc Ronthaler, Ground Truth Robotics GmbH)
FlatFish during open water testing in Salvador, Brazil. (Source: Marc Ronthaler, Ground Truth Robotics GmbH)

To achieve this, a variety of acoustic and optical sensors are employed and a subsea docking station will be designed. Two identical autonomous underwater vehicles (AUVs) will be built, enabling simultaneous software development in both Brazil and Germany.

Mission scenario: The vehicle is docked near an offshore facility and connected to a power supply and data uplink. The operator (either on the facility or at a remote location) commands the vehicle to perform an inspection task. Where the inspection target is remote from the facility the AUV undocks and follows a pipeline or flow line to a target area whilst collecting acoustic and visual data of the pipeline or flow line. Upon reaching the target area, acoustic and visual data from the inspection object is gathered. The vehicle then backtracks using acoustic and visual homing algorithms to return and connect to the docking station. The recorded data is then uploaded to be processed by automated systems and/or the human operator whilst the vehicle’s battery is being recharged. At the same time, the operator has the option of defining a follow-up mission.

DFKI’s contributions to the FlatFish project includes the mechanical and electrical design of the vehicle as well as manufacturing and integration of the AUVs. Also, the basic software modules will be implemented at DFKI to allow for sensor data processing, vehicle control and navigation solutions such as waypoint following.
Higher-level algorithms will be developed as a joint effort between DFKI in Bremen and BIR in Salvador, Bahia.

For the evaluation of energy and high-bandwidth data transfer while submerged, the docking station demonstrator was equipped with optical data transmitters and electrical contacts for battery charging. The data transfer system is based on blue LEDs and achieved data rates of 2.8 MBit/s during the final tests witnessed by representatives from Shell (Rio, Brazil and Houston, USA) and from SENAI-CIMATEC (Salvador, Brazil).

The energy transfer is based on electrical contacts made of Titanium, which are isolated from the surrounding salt water by a sealing system. During the test, successful underwater battery charging was shown several times.

As the industry partner, Shell plans to take the results from the FlatFish project (docking, autonomous inspection, 3D reconstruction, opto-acoustic homing, underwater data and energy transfer) in an effort to commercialize the FlatFish system.


FlatFish: Autonome SSIV Inspektion & Docking

Voll autonome FlatFish Mission: Das AUV koppelt sich ab, inspiziert sein Missionsziel (SSIV, subsea isolation valve) und kehrt zur Docking Station zurück.

Back to the list of projects
last updated 20.03.2023
to top