Underwater robot can search for missing black boxes

European researchers have demonstrated a new underwater robot that can retrieve objects such as aircraft black boxes from unexplored seabeds.

The I-AUV (autonomous underwater vehicle for intervention) developed by project Trident, which includes researchers from Spain, Italy, Portugal and the UK, combines an ability to map and navigate the ocean floor with a complex robotic arm.

‘It is based on a multipurpose system that incorporates a two-stage strategy,’ said project coordinator Prof Pedro Sanz of the Universitat Jaume I of Castellón in Spain.

‘In the first stage, the system is released into water. It builds a visual map of the seabed by scanning the area of interest where the goal of the intervention is.

‘After this, the system rises to the surface and has to be told, before launching it back into the water system, which is the goal of the intervention and what do we want to do with it. In a second stage, the operation is executed.’

The key challenges for the researchers, building on a previous project called Reconfigurable AUV for Intervention, were engineering a more complex and multipurpose robot and developing more robust and efficient control algorithms in order to overcome adverse underwater conditions.

Trident I-AUV was most recently tested in Port of Sóller in Mallorca in October 2012. Sanz said: ‘Before an unexpected problem of loss of a degree of freedom in the arm, thanks to the visual free floating manipulation and using other degrees of freedom of the vehicle itself, there was no problem in supplying the deficiency and in successfully solving the recovery of the object that we had planned,’

Partners involved in the project included a spin-off from the University of Genoa called Graal Tech that developed the arm mechatronics, a team from the University of Bologna that worked on the robotic right hand, and a group from the University of Girona that built the vehicle itself.

‘In addition, we develop all aspects of the simulation tool that has to  be generated to previously test in virtual reality all the actions that we want to transfer to the real scenario,’ said Sanz.