Outcome Using machine learning and artificial intelligence to improve the tracking of vessels in sonar spectrograms

Hydrophone sonar data can be used to determine the current and previous locations of nearby vessels, as well as to estimate where they may be heading. Trained human operators currently analyse the data in the form of broadband waterfall displays to detect and track the vessels. Although this is time-consuming and expensive, these human operators currently outperform traditional automated passive contact follower algorithms, such as the Kalman and Alpha-Beta filters: these filters are susceptible to the abundant underwater noise and struggle with crossing tracks and quiet contacts, while humans can use their experience to learn how to mitigate the challenging aspects of the task. An automatic detection and tracking model that is more accurate and robust than traditional methods would reduce the human operator's workload so that the human operators only need to investigate contacts of interest. It is also crucial that this model can detect the quieter contacts, which are of most interest to the navy, but do so without increasing the false alarm rate.

Although machine learning and artificial intelligence approaches seem ideal for this task, there are challenges. For example, the model must perform in real-time: the waterfall display is updated line-by-line each second. Whilst the model can use the previous rows of data, it must solely rely on the time vs. bearing (waterfall) data. The model may also need to be trained on large volumes of synthetic data due to the lack of real and unclassified data, and thus distributed training and techniques such as transfer learning may be required to increase the accuracy on what little real data exists.

The development of automatic detection and tracking algorithms are currently a hot topic in the data science and machine learning sectors due to the challenges involved in developing techniques that can outperform existing methods and the wide range of applications of the techniques in industry. This project would enable you to become an expert in a highly relevant and sought-after discipline, as well as engaging and gaining experience with an industry partner in the fascinating field of defence.

The expected outcome of this project is an automatic detection and tracking algorithm for broadband waterfall displays that will enable the human operators to only need to focus on contacts of interest thus reducing their workload. This model must perform in real-time as each new instance of data is recorded. The approaches developed will need to outperform techniques (such as those based on Kalman filters and Alpha-Beta filters) that have been configured, over decades, to work well in these contexts.

A previous feasibility study undertaken in collaboration with the University of Liverpool has shown promising results for Long Short-Term Memory (LSTM) networks. LSTMs can retain useful information about the previous time steps and exploit this when processing the current data. These models need to be adapted to work effectively with the broadband sonar data to efficiently produce tracks that are less susceptible to transient noise, more stable and can, potentially in combination with post-processing, correctly resolve crossing contacts.

This CDT's focus on using "Future Computing Systems" to move "Towards a Data-driven Future" resonates strongly with two themes of non-academic organisation. In both themes, albeit for slightly different reasons, commodity data science is insufficient and there is a hunger both for the future leaders that this CDT will produce and the high-performance solutions that the students will develop.

The first theme is associated with defence and security. In this context, operational performance is of paramount importance. Government organisations (e.g., Dstl, GCHQ and the NCA) will benefit from our graduates' ability to configure many-core hardware to maximise the ability to extract value from the available data. The CDT's projects and graduates will achieve societal impact by enabling these government organisations to better protect the world's population from threats posed by, for example, international terrorism and organised crime.

There is then a supply chain of industrial organisations that deliver to government organisations (both in the UK and overseas). These industrial organisations (e.g., Cubica, Denbridge Marine, FeatureSpace, Leonardo, MBDA, Ordnance Survey, QinetiQ, RiskAware, Sintela, THALES (Aveillant) and Vision4ce) operate in a globally competitive marketplace where operational performance is a key driver. The skilled graduates that this CDT will provide (and the projects that will comprise the students' PhDs) are critical to these organisations' ability to develop and deliver high-performance products and services. We therefore anticipate economic impact to result from this CDT.

The second theme is associated with high-value and high-volume manufacturing. In these contexts, profit margins are very sensitive to operational costs. For example, a change to the configuration of a production line for an aerosol manufactured by Unilever might "only" cut costs by 1p for each aerosol, but when multiplied by half a billion aerosols each year, the impact on profit can be significant. In this context, industry (e.g., Renishaw, Rolls Royce, Schlumberger, ShopDirect and Unilever) is therefore motivated to optimise operational costs by learning from historic data. This CDT's graduates (and their projects) will help these organisations to perform such data-driven optimisation and thereby enable the CDT to achieve further economic impact.

Other organisations (e.g., IBM) provide hardware, software and advice to those operating in these themes. The CDT's graduates will ensure these organisations can be globally competitive.

The specific organisations mentioned above are the CDT's current partners. These organisations have all agreed to co-fund studentships. That commitment indicates that, in the short term, they are likely to be the focus for the CDT's impact. However, other organisations are likely to benefit in the future. While two (Lockheed Martin and Arup) have articulated their support in letters that are attached to this proposal, we anticipate impact via a larger portfolio of organisations (e.g., via studentships but also via those organisations recruiting the CDT's graduates either immediately after the CDT or later in the students' careers). Those organisations are likely to include those inhabiting the two themes described above, but also others. For example, an entrepreneurial CDT student might identify a niche in another market sector where Distributed Algorithms can deliver substantial commercial or societal gains. Predicting where such niches might be is challenging, though it seems likely that sectors that are yet to fully embrace Data Science while also involving significant turn-over are those that will have the most to gain: we hypothesise that niches might be identified in health and actuarial science, for example.

As well as training the CDT students to be the leaders of tomorrow in Distributed Algorithms, we will also achieve impact by training the CDT's industrial supervisors.