ASV Global to Develop AI-Based Safe Autonomous Navigation for USVs

By Mike Ball / 10 Apr 2018
Follow UST

BMT ship simulator

ASV Global (ASV) has announced that it is leading a new £1.2million research project in partnership with offshore consultancy firm BMT to enhance the safety and reliability of autonomous navigation. The project team will use deep learning machine vision systems trained with a unique combination of simulated and real world data. Part funded by Innovate UK, the UK’s innovation agency, this project will enhance situational awareness enabling unmanned surface vehicles (USVs) to operate in extreme and congested marine environments.

The Synthetic Imagery training for Machine Vision in Extreme Environments (SIMVEE) project will build upon ASV’s existing, COLREG cognisant, autonomous collision avoidance and path planning capability. The project will use BMT’s REMBRANDT simulator to train and validate ASV Global’s vision algorithms to detect and classify objects at sea.

A key project output will be improved situational awareness for both the autonomy onboard and the remote human supervisor. The unique combination of real world and simulated data to train deep learning algorithms will improve the reliability of the existing system extending safe operations into complex environments with a wide range of objects to detect, classify and avoid.

Richard Daltry, R&D Director at ASV Global said: “This work will provide a significant step in the capability of ASV Global’s ASView autonomous control and navigation system. Today we use a remote human supervisor and AIS to classify objects and ensure safe operations. The addition of machine vision that detects and classifies objects extends our COLREG compliant autonomous navigation, enabling operations in limited bandwidth with reduced supervisor workload.”

Phil Thompson, Managing Director at BMT, comments: “We’re delighted to be partnering with ASV on this project. This research will play a pivotal role in helping to accelerate the wider adoption of unmanned systems and increase trust in their feasibility by mariners around the world.”

Utilising data gathered by BMT’s REMBRANDT ship manoeuvring simulator as well as real world onboard camera data will allow the team to train the autonomy system with large quantities of data. This method provides a cost-effective solution to generating the data and accelerates the machine learning process. The project will not only allow ASVs to operate in the same way as traditional manned vessels at sea but will also open up new use cases and applications with the added use of BMT’s Search and Rescue Information System, (SARIS).

Posted by Mike Ball Mike Ball is our resident technical editor here at Unmanned Systems Technology. Combining his passion for teaching, advanced engineering and all things unmanned, Mike keeps a watchful eye over everything related to the unmanned technical sector. With over 10 years’ experience in the unmanned field and a degree in engineering, Mike’s been heading up our technical team here for the last 8 years. Connect & Contact