Field swappable underwater storage unit for an AUV

A leader in Smart Defence, CMRE enables Nations to work more effectively and efficiently together by prioritising national needs, and focusing on research and technology challenges, both in and out of the maritime environment, through the collective power of its world-class scientists, engineers, and specialised laboratories in collaboration with many partners in and out of the scientific domain.

'Collaborative Autonomous Mine Countermeasures CA-MCM' is one of the main projects at CMRE, aiming at increasing the capabilities of Autonomous Underwater Vehicles (AUVs) by using Synthetic Aperture Sonar (SAS) to quickly and reliably detect, classify and localise mines. This implies the use of high performance, efficient processing systems installed in underwater robots working in real-time.

CMRE has been using CPU boards produced by SECO s.r.l. since 2013. The PicoITX boards (Single Board Computers) are suitable for integration in underwater vehicles where reliability, reduced footprint, low power consumption and computational power are key factors. As of today these boards have been installed on a wide range of vehicles that are part of the CMRE fleet.

The CMRE 'Collaborative Autonomous Mine Countermeasures CA-MCM' project focuses on increasing the efficiency and effectiveness of the full range of mine search missions. Advancing the use of AUVs for mine countermeasures has been achieved in a three-pronged approach:

1. The research and development of data processing algorithms for through-the-sensor underwater perception and autonomous behaviours for optimal data acquisition,
2. The development and implementation of the algorithms on-board the vehicle in real-time,
3. The maintenance and upgrade of the vehicle, its on-board processing capabilities and its sensors.

The research axis has gone from basic way-point following to the development of a machine decision-making capability. Incremental steps taken along the way include the development and implementation of:

• operator interrupt and re-direct,
• real-time on-board SAS processing at 3cm along-track resolution up to 150m range,
• world-leading on-board target detection and classification,
• novel sensing and adapting to sand ripples algorithms,
• sensing and adapting to adverse currents algorithms,
• sensing and analysing environmental information,
• extracting from different types of data environmental attributes such as complexity in terms of mine-hunting of an area,
• assessing the quality of the data,
• adapting track spacing to ensure optimal area coverage,
• autonomous re-acquisition of multiple views on detected targets,
• task arbitration and machine decision-making for behaviours de-confliction.

The next-gen of Autonomous Underwater Vehicles (AUVs) currently developed at CMRE is equipped with wide suite of sensors with on-board processing capabilities as well as a very high level of autonomy and adaptation to the environment in order to achieve high level mission tasks without human intervention. High computational power, large data acquisition and storage are needed to enable these features. Furthermore, the continuous development on these systems requires even larger storage capacity and very high reconfigurability.

One of the best example of these vehicles equipped with the PicoITX boards is the MUSCLE, CMRE major MCM AUV. With a high-resolution, high-frequency synthetic aperture sonar (SAS) installed, MUSCLE provides superb image quality of objects on the seafloor, and it has a high level of autonomy, thanks to its real-time processing software running on a dedicated high-end GPU-based system and advanced decision-making capabilities. Despite its relatively low TRL (Technology Readiness Level), the MUSCLE unmanned vehicle is regularly tested and successfully utilised in operational Mine Countermeasure (MCM) scenarios in which a very quick turnaround time from mission execution to post mission analysis and very short mission to mission downtimes are key factors.

CMRE began developing concepts for allowing a quick data transfer on autonomous vehicles in 2014. The challenges in developing an underwater field swappable storage unit are in the use of unconventional underwater connectors (typically not suited for standard data buses) and the miniaturization imposed by the portability of the system despite its ability to sustain high pressures.

The SECO Pico ITX boards with two Gigabit Ethernet ports and 2 SATA3 channels combined with their small form factor seemed a possible solution for such applications. The full integration of one of these boards into a small water-tight 300m rated case developed and built at CMRE was successfully accomplished in 2015. The system has then passed multiple bench tests and it is now approaching the field testing stage. The high computational power and the possibility of installing up to 8GB of RAM will allow for some data processing contribution to the overall reduction of the post-mission-analysis time.

The recently built underwater field-swappable storage unit is expected to significantly reduce the mission-to-mission downtime due to the data transfer from hours to minutes. This will make multiple missions in the same day possible and it will bring the operational readiness of the CMRE underwater robots, in particular the MUSCLE AUV, one step further. The quick storage unit swap will also allow the operator to have access to the results produced by the advanced internal processing unit of the MUSCLE AUV during the mission execution.

Thanks to the processing power of the SECO Pico ITX boards, the new system has been designed by CMRE to run completely autonomously from the other vehicle’s subsystem and it is capable of retrieving and storing the mission data fully unsupervised. Its use as a secondary storage unit allows for higher data integrity. The SECO Pico ITX boards have also been successfully tested with NATO approved storage units used in applications with strong requirements in terms of Information Assurance.

The availability on the market of different CPUs/SOCs, from ARM to x86, gives a wide range of options and power consumption vs processing power ratios. As a result, CMRE already envisages using this storage unit on multiple vehicles part of its fleet.