Defense

Adaptive perception: trailblazer for cognitive radar

In the loop of a cognitive radar, perception allows follow-up actions such as the self-adjustment of the parameters, task allocation or the acquisition and utilization of »knowledge«.
© Photo Fraunhofer FHR

In the loop of a cognitive radar, perception allows follow-up actions such as the self-adjustment of the parameters, task allocation or the acquisition and utilization of »knowledge«.

Instead of the 400 carrier frequencies of the conventional methods of inverse fast fourier transformation (IFFT, above) compressive sensing (CS, below in red) can create equally good or better distance profiles with just 25 percent of the measurements.
© Photo Fraunhofer FHR

Instead of the 400 carrier frequencies of the conventional methods of inverse fast fourier transformation (IFFT, above) compressive sensing (CS, below in red) can create equally good or better distance profiles with just 25 percent of the measurements.

MIMO radar systems can, for example, exploit waveform diversity by simultaneously transmitting several pulses at different frequencies at any time.
© Photo Fraunhofer FHR

MIMO radar systems can, for example, exploit waveform diversity by simultaneously transmitting several pulses at different frequencies at any time.

Optimized utilization of resources, enhanced performance and high adaptability – cognitive radar relieves the burden on the operator and opens up new application areas. Prerequisite: the radar system must be able to perceive correctly its surrounding environment.

Cognitive radar systems can intelligently adjust their operational parameters and control to the situation and task at hand. This is attributable to their adaptive perception capability: they provide an optimal image of the radar scene by generating and processing radar signals in line with the specific situation and adapting these in an ongoing process. The means of doing this are varied and radar systems that are already seen as being up-to-date can be offered an additional degree of freedom.

However, the combination of these different technologies – from waveform selection and design to MIMO radar and compressive sensing – and innovations in the hardware area are necessary to move one step closer to highly automated, cognitive radar systems. This is the task of the group »Adaptive Perception«, which was established Fraunhofer FHR at the beginning of 2015.

Key technology MIMO

The waveform emitted by the radar is responsible for the resolution, precision and evaluability of the measurement of target distance and target speed. To achieve optimal adaptive perception, the radar system must be capable of emitting different waveforms and adapting these to the radar scene in a dynamic manner. This can be accomplished with various adjustment screws, by changing the waveform parameters or through continuous adaptation of the beam pattern. The type and extent depend on the task at hand such as detect, locate or classify and on the image criteria that has to be optimized, e.g. improved signal-to-noise ratio or the avoidance of potential electronic interference.

MIMO (Multiple Input-Multiple Output) radar systems currently offer the best possibility for large waveform diversity. In contrast to phased array radar systems, each individual antenna has a waveform generator with the result that each element can transmit at a different frequency and/or create different waveforms at each pulse.

The team is therefore investigating various MIMO topologies and orthogonal transmission concepts. One of the focal points lies on time multiplexing, a technique that is frequently used in modern radar systems and relatively easy to implement. As this alone does not fully exploit the potential of simultaneous transmission of the MIMO concept, the team is currently investigating coded waveforms with good correlation properties. Here, the modulation of the amplitude using a digital code allows a clear separation of the virtual MIMO channels and creates a good ambiguity function. Another focal point is the development of criteria on estimate accuracy for the adaptive selection of the optimal waveform parameters and antenna topologies.

Efficient signal processing through compressive sensing

The data volume generated for adaptive perception when the MIMO capacities are used to the full is enormous and signal processing is complex. It is important, however, that cognitive radar produces good results during real time operations. One promising approach is the compressive sensing (CS) theory: it uses the purely mathematical "sparsity" of most signals, and hence their compressibility in certain domains for subsequent signal processing. It is therefore capable of carrying out a good or even better reconstruction of the original signal using comparatively fewer measurements.

The FHR engineers are using the CS approach to develop algorithms for adaptive perception. Applied to the design of the array antenna topology as well as signal processing and beamforming, the resolution (super resolution), localization precision and target detection have already been approved significantly. The dynamic adaptation and, where appropriate, the compression of the transmit and receive signal, is much easier to implement with the CS framework. This brings additional advantages for resource optimization, particularly when used in modern multi-function radar systems. With their arbitrary beam control and programmable functionalities, they form ideal platforms for the application and demonstration of these concepts and algorithms.