There are two sides to the coin. Cognitive radar adapts its own waveform in real time to the environment and to observed jamming — it changes frequency, pulse shape or search pattern as soon as it notices that it is being jammed. The result is an arms race at machine speed: your cognitive jammer against their cognitive radar, a fight that plays out in milliseconds and in which the human mainly supervises.
That also brings new challenges:
- Training data: representative signals from real threats are scarce and sensitive.
- Trust and explainability: can the operator trust a decision that a neural network made “internally”?
- Misclassification: a wrong assessment leads to ineffective jamming — or worse, to fratricide against friendly systems (Doctrine & EMSO).
Cognitive EW is therefore above all a software race: the winner is not whoever has the most powerful transmitter, but whoever can update their algorithms fastest.