Performance robustness of a noise-assisted transmission line (Q1038450)

Consider a neural, storage or information transmission line of a nonlinear communication channel modeled by a system of coupled oscillators with \(n\)-th oscillator motion governed by \[ dX_n(t) = \left(-\frac{\partial U(X_n(t))}{\partial x}+\varepsilon X_{n-1}(t)\right) dt + \sigma dW(t) \] driven by standard Wiener process \(W\), where \(U\) is the potential given by \[ U(x) = U_0 \left(\frac{x}{x_c}\right)^2 \left[ \left(\frac{x}{x_c}\right)^2 - 2 \right] \] and a certain input sequence \(\varepsilon X_0\). \(\varepsilon>0\) is the coupling strength of adjacent oscillators in this system. One is interested in transmission properties of these systems characterizing its performance by means of certain metrics in communication such as output Bit Error rate (BER) and Signal-to-Noise ratio (SNR). Note that the BER measures the probability of receiving erroneous bits and, for linear channels with additive Gaussian noise, the increasing SNR decreases the minimium allowable BER. There is a critical coupling strength \[ \varepsilon \approx 1.54 \frac{U_0}{x_c^2} \] which separates different transmission regimes, namely that of noise-supported (subcritical) and coupling-supported (supercritical) transmission lines. The supercritical regime allows transmissions without noise, whereas the subcritical regime requires noise in order to sustain transmission. Using the Euler-Maruyama method, the authors compute approximations of BER and SNR for a double-well forward-coupled SR-driven information transmission line. The authors show that an enhanced transmission performance characterized by both SNR and BER can be achieved in a certain region of parameters, compared to linear systems with additive noise (e.g. where output BER remains flat for a broad range of noise for a supercritical coupling strength).