Autonomous sensor systems are delivering an operational advantage to defence forces.
Understanding that reliable, specialist datasets are fundamental to developing this technology, Overview recently took part in one of the largest artificial intelligence (AI) data capture trials ever conducted for UK defence research.
Collaboration with defence industry leaders
The 5-day event was hosted by the Defence Science and Technology Laboratory(Dstl) and included industry and international partners, supported by elements of the UK armed forces.
As well as Overview, industry participants included companies such as:
Thales, Leonardo, Frazer-Nash Consultancy, Faculty-ai, Lockheed Martin, Vizguard, Living Optics, Chess Dynamics, MSI, Sesanti and BAE Systems.
International teams comprised:
Airforce Research Lab(AFRL)), Munitions Directorate, US Department of Energy Pacific Northwest National Laboratory (PNNL) and US Army Combat Capabilities Development Command Armaments Centre (DEVCOM AC) supported by Opto-Knowledge Systems Inc. (OKSI).
Trial logistics
The trial, which took place along the Hampshire coastline and in the Solent, included almost 130 personnel, 13 vessels, a light aircraft and multiple uncrewed air vehicles (UAVs).
The trial featured beach landings with passengers disembarking / embarking to generate data representative of a range of human behavioural traits.
The two principal scenarios were ‘synchronised landings’ involving volunteers acting as a trained military unit, and ‘chaotic landings’ where participants exited the boats in a deliberately random manner.
Podview autonomous sensors were utilised to record the activity, alongside over 50 cameras, radars and other sensors.
PodView sets the standard for agility
Over the 5 day trial PodView autonomous sensors were deployed in a variety of locations to capture the movement of vessels, UAVs and personnel. The lightweight plug and play systems are easy to transport and were deployed in under 5 minutes so moving between locations was quick and easy.
Once deployed, full motion video and metadata were collected in real-time, and even when the wind picked up to 40 knots, PodView continued to deliver!
Two Podviews were used in the trial; PodView2000 captured optical and LWIR video, whilst PodView XL captured long range optical and MWIR video.
The video captured will be used alongside other information to build datasets to train and validate AI algorithms for object detection and recognition.
In summary
Charlie Maslen, Dstl’s trial technical authority, sums up the trial perfectly, saying:
“Data generated during this ambitious and challenging trial will enable MOD and industry partners to develop new AI products for Defence, helping keep UK forces safe and delivering operational advantage. Being able to guarantee the integrity of the data underlines MOD’s commitment to the ethical, safe and responsible use of AI.”
For further information about PodView get in touch.
