Defence
SpaceAware Detect
SpaceAware Detect is an advanced solution based on an innovative Deep Learning technology for target identification and tracking of objects.
Field of Expertise
Aerospace, Space, Geospatial, AI/ML
Expertise
Dynamic intelligence Solutions (DIS) Limited is a Joint Venture between Riskaware Limited and Operational Solutions Limited (OSL), set up with the explicit aim of developing Artificial Intelligence and Machine Learning (AI/ML) based algorithms for the identification and classification of moving objects from track movement dynamic. DIS brings together the Enhanced Drone Identification and Target Tracking System (EDITTS), originally developed by Riskaware, with OSL's C-UAS and Unmanned Traffic Management (UTM) Command and Control (C2) system, FACE, to provide a world leading capability aimed at providing significant improvements in intelligence, security and situational awareness at ports and borders around the world. DIS's EDITTS is a machine learning (ML) solution for classifying targets based on their track dynamics and behaviour. Our approach is significantly more sophisticated than other systems in development with the ability to distinguish targets in a more nuanced way. EDITTS has are already demonstrated remarkable success in distinguishing between different types of UAS and in distinguishing these from other airborne objects such as birds in flight. The capability uses supervised deep learning in the form of a convolutional neural network, coupled with specialised routines to extract the underlying dynamics data in a processable form. Investigation of alternative methods such as sequential Bayesian formulations and various alternative neural network architectures have concluded that the proposed algorithm significantly outperforms these alternatives. A key advantage of using track dynamics as a means of classification is that it is generalisable to any type of sensor that can detect position and movement, and does not require imagery, which may be scarce, particularly if visibility is poor and/or targets are still far away. A key feature of EDITTS is a multi-sensor track fusion capability. This is based on a highly optimised particle filter algorithm. Particle filters have particularly strong capability for fusing data from multiple sensors. Importantly, they do not rely on gaussian assumptions about the positional noise, hence, can cope with the unusual noise behaviour from some types of sensors, including RF sensors. A highly optimised associator algorithm has also been developed to assign target data to particular tracked targets at very high scales. Under the Innovate UK Future Flight grant funding, the EDITTS algorithm will undergo significant enhancement with the aim of developing a robust, operationally deployed capability within 24 months of project start. In the first 12 months, work will focus on integration of the core algorithm within OSL's FACE C-UAS and UTM C2 system and to enhance EDITTS to utilize both raw and processed sensor data. This will enable the core algorithms to be trained and extensively validated using both live and historical data sets from the FACE sensor suite. Throughout the validation process the core algorithms will be refined and enhanced to provide optimal performance based on live and synthetic trails. During this period, EDITTS will also be extended to track and classify ground-based targets and to incorporate high resolution environmental context information, such as terrain and building data. Year two will initially focus on system productisation and developing and running an extensive test and evaluation framework utilising live field and synthetic data trials. This will result in a robust, site calibrated capability ready for operational trials at Heathrow Airport within 18 months from project start. The remaining 6 months of the project will focus supporting the operational trails and fine tuning the EDITTS algorithms for improved operational performance. This funding opportunity provides a unique and unmissable opportunity for Dynamic Intelligence Systems to work with OSL and partners to develop, test and evaluate a world leading counter UAS capability and to deploy this capability within one of the largest and busiest airports in the world, Heathrow. This would provide a springboard into the global ports and borders security market. For example, it is estimated that there are over 41,000 airports worldwide of which over 1,200 are large international operators. In January 2019, the Independent calculated that the Gatwick drone incident alons cost the airport in the region of £50M in financial damages. The BBC later reported that Gatwick had spend £5m on technologies designed to prevent further attacks. Based on a study by Philippe Wendt etal, using Frankfurt Airport as a case study, the authors estimated that a single drone incident can cost an airport in the region of €60,000 per hour. Their conclusions stated that "the investment cost of a counter-drone system worth approximately €2.9 million per unit can be justified from the perspective of the airport operator (Fraport) based on the costs of a single 48-h incident during a peak period of activity". These case studies demonstrate the scale of the issue and the potential size of the market. In addition to airport security, the wider defence and security market also has significant opportunities for C-UAS technologies. For example, Guardian reported that in October 2018, seven men were jailed after using drones to fly £550,000 worth of drugs into prisons in the Midlands and the north-west. They also reported that in Donbas, Russian-backed separatists weaponised consumer drones to drop grenades on Ukrainian government trenches. Furthermore, with a growing number of countries and commercial organisations getting involved in space activities comes an increase in the number of operational spacecrafts, defunct satellites, and space debris. Space situational awareness and object tracking, identification and intent analysis has become increasingly important for companies, institutions and governments operating satellites. With over 23,000 objects in orbit greater thank 10cm in size, and many hundreds of thousands of pieces of debris that are smaller, understanding what threat they pose to operational systems is vital. This issue has been recognised by the UK Defence Science Technology Laboratory in their latest Defence and Security Accelerator (DASA) Space to Innovate Call and is an area of direct applicability to the EDITTS technology.