Artificial Intelligence promises to answer some EW preoccupations, provided that we ask the right questions.
Embedded Tech Trends conference
The 2020 Embedded Tech Trends conference held in Atlanta, Georgia, between 27 and 28 January, hosted by the VITA open systems advocacy organisation was an excellent opportunity to discuss some of the emerging trends in electronic component, board and systems technologies. The theme of this year’s conference was “inventing, connecting, inspiring (and) thriving,” and the event witnessed lively discussion over two days of presentations and meetings with some of the most important names in these industries.
Artificial Intelligence (AI) was an oft revisited theme. The ability of machines to intelligently crunch ‘big data’ and learn from their experiences and environments will have clear implications for the defence industry and armed forces around the world.
“Today we have big data on the battlefield,” observed Devlon Yablonski, principal product manager for sensor processing for Mercury Systems. The gathering of huge quantities of data was the direct result, Mr. Yablonski argued, of battlefield sensors be they optronic, acoustic or electromagnetic, “getting more and more capable in capturing more and more data.”
In particular, he cited the proliferation of hypersonic weapons. Such threats can typically travel at speeds in excess of 3,333 knots (1,799 kilometres-per-hour). This is already being witnessed in the anti-ship missile domain.
The People’s Republic of China’s new DF-100 anti-ship cruise missile, unveiled in early October 2019, can reportedly hit such velocities. The result of this is that the human brain has less time to respond effectively once such a threat is detected and inbound, thus increasing the reliance on machines to aid, and in some cases perform, decision-making.
“Hypersonics are super fast, and perhaps only computers are quick enough to deal with such threats?” posited Mr. Yablonski.
AI at the Sensor
The potential contribution that AI can make to decision-making also fell under the purview of Nigel Forrester, director of business development at Concurrent Technologies. For Mr. Forrester one of the crucial factors to be considered regarding AI is how to apply the unique capabilities this brings to the analysis of sensor data. He argued that we need to be asking “how are we using AI to create actionable intelligence at the point of data collection?”
Mr Forrester argued that it was vital for AI capabilities to be placed as close to point of data collection as possible. This means ensuring that such capabilities are collocated with the sensor so that data processing can be done in as short amount of time as possible.
This chimes with Mr Yablonski’s belief that AI offers distinct advantages at the tactical and operational levels regarding the speed at which it can make process big data. Having AI capabilities located near the point of data collection greatly increases this.
Moreover using AI to process the data at the sensor reduces the amount of raw data that has to be shared on a tactical network or on a cloud, thus reducing the bandwidth burden for tactical communications networks.
In the electronic warfare context, Mr. Forrester stresses that AI has a hugely important role to play: Communications and radar signals are using increasingly complex waveforms: “We have seen a lot of Signals Intelligence (SIGINT) applications where you are trying to understand parameters and characteristics, and now things are becoming even more complex with things like frequency agility.”
Future of AI in Electronic Warfare
AI is fast becoming a reality in the electronic warfare domain. As Mr. Forrester’s and Mr. Yablonski’s presentations noted, asking the difficult questions now will ease the large scale introduction of AI into electronic warfare, particularly SIGINT analysis, in the near future.