Artificial Intelligence (AI) and System Automation (SA) is combining with Digital Signal Processing and RF to create radar/sonar, Electronic Warfare (EW) and next-generation persistent ISR systems more accurate than ever, because of the advances in modern High Performance Embedded Computing.
Since the turn of the century, it appears a new form of radar/sonar that is neither all-active nor all-passive. It’s called adaptive radar/sonar, which can automatically select RF waveforms and processing filters in real time, based on the mission, its changing environmental conditions and target countermeasures. With the move to Active Electronically Scanned Array, like RBE2-AESA radar on the French RAFALE omnirole fighter-, wideband spectrum with very low energy level, back-end software control and processing, many radars are becoming multifunction combining detection, imaging, identification, tracking, targeting, communications and electronic warfare. Multifunction responsive systems use software-defined adaptive waveforms, to offer wideband RF front-end with real-time control while the back-ends define the mission of radar, communication and EW. Focusing on frequencies of interest, they present better defined and cleared picture to the console operators or to the algorithms using that information, eliminating noise and false alarms.
Functions that would have required several digital signal processing embedded computing boards in years 2000 (14x Quad-Core Power PC for example) are being done now in real-time onboard with one single Xeon D 12-Core SBC, thanks to the incredible improvement of Embedded computers in SWaP-C constraints. The explosion of advanced and HD sensors (like EO/IR, AESA, SAR…) data stream, the increasing types of distributed sensors embedded on aircrafts, UAV’s or ships data flow are fused in a cross-cueing/superimpose processing and delivered to CMS –Combat Management System- and come up with a clear synoptic picture shared by the combat management team to analyze the global situation to take the right decision at the right time by the decision-makers.
Next step, is to digest better and easier this waterfall of data coming out of sensors, integrating greater automation and in-system processing to reduce analysts overload. The natural path is to make the signal, display, and data processing more automatic, to reduce drastically the number of radar/sonar operators (cf. CRYSTAL 23 multifonctions C2 Navy console). The system reports tracks to the CMS, and the CMS operator, in natural language, can ask the radar why it thinks the threat is real. This is exactly where the Artificial Intelligence (IA) enters the picture: when we replace the human intelligence analysts in the decision chain. This is what DARPA – Defense Advanced Research Projects Agency-, in USA, call cognitive radar that will allow more efficient control. It comprises the processes of feedback, learning, information preservation, and adaptability in transmission. It adapts its transmission waveforms, radar beams, and other electronic data processing capabilities in order to achieve superior performance in detection, estimation and tracking by exploiting various knowledge sources. That will automatically generate more advanced countermeasures.
What is needed is an EW system that can detect and analyze new signals and come up with ways to jam or spoof those signals in real-time. Machine learning and cognitive EW scan the radio spectrum in real-time to determine what the adversary’s radar is doing and then, right there in the spot, create a jamming profile to protect the aircraft. The key challenge is keeping pace and being agile in a rapidly evolving environment.
On other hand, adaptive and cognitive radar allows to optimize resources. It will use less power while increasing the probability of detection by directing more RF energy on targets, increasing update rates. Advanced field-programmable gate arrays (FPGAs) that excel at parallel throughput and determinism, adaptive beamforming techniques, multi-core Server Class system-on-chip (SoC) with DSP capabilities, and GPU massive parallel computing are aiding in the development and fielding of more advanced adaptive radars and sonars.
- The two main enabling technologies have been the RF front-ends and the back-end processors:
Streaming Computation cutting edge FPGAs (like Xilinx UltraScale or Intel/Altera Stratix) with Analog-to-Digital converters (ADCs) and Digital-to-Analog converters (DACs) (see Vadatech FMC’s and FPGA VPX-3U carriers choice’s guide attached to this September ECRIN’s News) are getting faster and more powerful on the front-ends to offer excellent Digital Radio Frequency Memory (DRFM) functions with system latency in nanoseconds (up to Dual ADC 12-bit @ 6.4 GSPS plus Dual DAC 16-bit @ 12 GSPS on UltraScale XCKU115 FPGA carrier in VPX-3U form factor) for radar spoofing and other Electronic Countermeasures (ECM). Common examples of EW include detecting adversary’s communications at a certain frequency and then jamming that frequency or detecting an incoming radar signal and responding with a simulated return signal indicating a bogus location;
- Artificial Intelligence, with complex branching decision tree General Purpose Processor and deep learning algorithms Graphics Processors (Intel 12-Core Xeon D SoC with NVIDIA Pascal 6.2 TFLOPS GPGPU Engine) will play a big role to integrate multiple functions on the back-ends.
Cost Electronic Warfare System Architecture
Software-defined systems, with more RF components replaced by software back-ends, will fit changing mission and environment requirements using software modifications. Advanced signal processing techniques will allow the back-ends to analyze data in a more thorough fashion and more efficiently adapt the radar to the needs at hand. Adaptive systems will continue to converge, performing multiple RF functions. The technologies of radar, persistent ISR, and EW are common across those different classes of sensors, so in the future it is really the software back-ends and efficient management of the RF front-ends that will provide advantages in the military, to be flexible during the mission and over the life cycle to keep pace with a very evolving threat environment.
COTS technologies are pushing and will continue to push the frontiers of the possible in EW, ISR and Radar/Sonar applications. With its new TOPAZE D series in development, ECRIN Systems will offer lightweight, extremely flexible and powerful rugged system, thanks its adaptive VPX interconnect backplane system and COTS Open VPX modules. That will allow the flexibility to configure the board set that best suits the requirements of the application. At the difference with board vendors that push their own modules into a system solution, ECRIN Systems with its huge experience of 40 years as System Integrator, will select the best of building blocks coming from worldwide board vendors and will integrate them into its own system infrastructure, under its Quality Management Program control with dedicated Project Managers and R&D resources and support.