Military surveillance systems and civilian technologies are drowning in data from radars, sonars, phone signals, and social media, and existing processing methods cannot keep up. This project tackles a fundamental bottleneck: how to transform a chaotic flood of high-volume, high-speed, multi-source data into reliable, actionable information. Current signal processing techniques were designed for simpler, single-sensor inputs and cannot handle the scale, speed, or uncertainty of modern data streams—whether from defence sensors or everyday networks. The consortium of five universities and ten industry partners will develop new mathematical inference algorithms that are computationally efficient, scalable, and able to fuse information from both physical sensors (radar, lidar) and human-generated sources (analyst reports, social media). They will also apply machine learning to detect anomalies, threats, and opportunities in real time. If successful, the work could improve defence situational awareness and autonomy, but also directly affect civilian technologies: smarter automotive lidar and radar systems, more secure facial recognition, and more robust broadband networks. The research is fundamentally about the mathematics of information processing, but its outputs will underpin systems that quietly keep society running—from traffic management to public safety communications.
View original technical description
Persistent real-time, multi-sensor, multi-modal surveillance capabilities will be at the core of the future operating environment for the Ministry of Defence; such techniques will also be a core technology in modern society. In addition to traditional physics-based sensors, such as radar, sonar, and electro-optic, 'human sensors', e.g. from phones, analyst reports, social media, will provide new valuable signals and information that could advance situational awareness, information superiority, and autonomy. Transforming and processing this broad range of data into actionable information that meets these requirements presents many new challenges to existing sensor signal processing techniques. In a future where a large-scale deployment of multi-modal, multi-source sensors will be distributed across a range of environments, new signal processing techniques are required. It is therefore timely to consider the fundamental questions of scalability, adaptability, and resource management of multi-source data, when dealing with data that is high-volume, high-velocity, from non-traditional sources, and with high uncertainty. The UDRC Phase 3 project, Signal Processing in an Information Age is an ambitious initiative that brings together internationally leading experts from 5 leading centres for signal processing, data science and machine learning with 10 industry partners. Led by the Institute of Digital Communications at the University of Edinburgh, in collaboration with the School of Informatics at Edinburgh, Heriot-Watt University, University of Strathclyde and Queen's University Belfast. This multi-disciplinary consortium brings together unique expertise in sensing, processing and machine learning from across these research centres. The consortium has been involved in defence signal processing research through the UDRC phases 1 & 2, the MOD's Centre for Defence Enterprise, and the US Office of Naval Research. The team have significant experience in technology transfer, including: tracking and surveillance (Dstl), advanced radar processing (Leonardo, SEA); broadband beamforming (Thales); automotive Lidar and radar systems (ST Microelectronics, Jaguar Land Rover), and deep learning face recognition for security (AnyVision). This project will investigate fundamental mathematical signal and data processing techniques that will underpin future technologies required in the future operating environment. We will develop the underpinning inference algorithms to provide actionable information, that are computationally efficient, scalable, and multi-dimensional, and incorporate non-conventional and heterogeneous information sources. We will investigate multi-objective resource management of dynamic sensor networks that include both physical and human sensors. We will also use powerful machine learning techniques, including deep learning, to enable faster and robust learning of new tasks, anomalies, threats, and opportunities, relevant to operational security.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know