Dynamics based and dynamics level noise mitigation
Noise is ubiquitous and it limits the performance of many engineering applications. When it is probable that a single system might fail to operate properly and produce a correct output in the presence of a strong noise, the common practice is to use a redundant set of these systems, let them operate separately and independently, and at the end choose the output that the majority of these systems produce. For example, rather than having one single system, use three systems, and choose the output that at least two systems produce. The probability that the majority of redundant systems produce a wrong output is less than the probability that a single system produce a wrong output.
We have showed that coupling dynamical systems mitigate local noise. We have demonstrate that coupled dynamics attenuates the incoherent local noise, whereas it preserves the coherent signal. We have utilized this feature to systematically apply coupled dynamics to reduce noise. The technique we introduce is simple: arrange a series of identical systems and dynamically coupled them together. The coupled dynamics across the systems mitigates the local noise.
Furthermore, we have demonstrated that dynamically coupling redundant systems results in better noise robustness than the case where majority wins technique is applied. The reason is simple. Coupled dynamics actively and dynamically mitigates noise across the coupled systems, whereas the majority wins has no actual noise mitigation capability, rather it allows the noise grows in systems, and then at the end it chooses the majority result.
We are now introducing straightforward and cheap to fabricate circuit designs for dynamical couplings. Such a dynamics based approach for noise mitigation can be an cable method to reduce device level noise and to implement more robust to noise systems.