Debugging the Hype: AI and ASVs

Friday June 15th, 2018 TAGS:

Artificial Intelligence (AI) is often claimed to be the latest and greatest in robotics and the future of autonomous technology. While AI is certainly a big part of future progression for autonomous and unmanned technology, it is definitely not new.

AI covers a wide range of techniques, including information extraction, numerical optimisation, data analytics, predictions and, of course, machine learning. AI has been through Gartner’s ‘hype cycle’ several times, starting with innovation leading to inflated expectations, counteracted by disillusionment, rebalanced by enlightenment and finally flourishing into a plateau of productivity. This is where we start to see realised products emerging.

Artificial intelligence plotted onto Gartner’s hype chart.

The reason AI keeps popping up as a ‘new’ thing is that new computational possibilities are constantly being discovered and consequently the idea of what qualifies as intelligence is always being redefined.

The essence of AI is the process of a machine simulating human-like decision-making. New possibilities are continually opening up because of boosted computing power and our ability to store and process significantly larger amounts of data than ever before.

Uses of AI have been recorded as early as 1957 with the first implementation of the perceptron algorithm for image recognition by Frank Rosenblatt at the Cornell Aeronautical Laboratory. In machine learning, the perceptron is a binary classifier (a function that  maps inputs to 0 or 1). It is a linear classifier, i.e. it makes predictions based on a linear combination of its inputs.

After the initial hype, the limitations of the perceptron were highlighted, in particular by Minskey and Papert in the 1960s, and scepticism about the capability of the perceptron set in. Since then considerable research has been undertaken and perceptrons form the basis of what we now know as “deep learning”, which is at the heart of many successful image recognition systems today.

In fact, ASV Global (ASV) is currently undertaking a research project to train and validate  vision algorithms to detect and classify objects at sea.  While deep learning approaches are much more practical today because of the computing power and data volume available to us they still have their limitations, like any technology, and it is important to understand when it is and isn’t appropriate to use them or possibly combine them with other methods. Rigorous research and development is required to ensure the right combination of techniques to produce a reliable system.

Further to this, we have seen AI applied to various problems in different industries, such as trade and business analysis, machine prognostics and diagnostics, and most prominently, recently in advanced driver assistance systems for automatic parking, lane assistance and even traffic sign recognition.

Whilst we continue to enjoy the surge of research in AI that allows us to continue further developing our autonomy systems, ASV Global sees huge potential beyond the public hype, such as autonomous path planning, understanding water traffic, situational awareness, boat prognostics and diagnostics and much more. Research and development on our system increases the longevity of our technology and ensures our fundamental goal: to enhance the safety and reliability of autonomous navigation and ultimately save time, money and, potentially, even lives.

ASV Unmanned Systems Developer, Hannah Thomas said, “artifical intelligence is shaping the world around us and improving the way we work. To be at the heart of a such a driving force in today’s industries is a privilege and a delight, and the wealth of opportunity at ASV was a key factor in attracting me to join. Although AI is not new, it is certainly getting more and more exciting today.


Hannah Thomas

Hannah Thomas

Hannah has had experience in designing and building artificial intelligence solutions for a wide range of applications across industries. Before coming to ASV Global, she completed a MathCompSci degree at Oxford and worked in investment banking. She is the first data scientist in the Control Team at ASV and seeks to bring new insights through the power of AI and analytics. The Control Team is the largest team at ASV Global and is continually growing. These people are the brains behind the company’s proprietary autonomous vessel control system; the hardware and software that makes a boat autonomous. Without them there would be no ‘A’ in ASV.

Join the Autonomous Revolution

ASV Global is actively seeking graduate and experienced software engineers to join its control team. For more information click here.

Go Back

Get in touch

Please contact us using the form below or by emailing

I have read and agreed to ASV’s privacy policy.
I understand that my details will be used by ASV to contact me as per the privacy policy and will not be shared with third parties.



To stay up to date with what’s happening with L3 ASV enter your email address below.