Probabilistic Distribution Modeling of Crosstalk in Multi-Conductor Transmission Lines via Maximum Entropy Method

A novel statistical method, referred to as the maximum entropy (MaxEnt) method is proposed in this paper for effective probabilistic modeling of crosstalk in multi-conductor transmission lines (MTLs). The principle of the MaxEnt method states that for the given information constraints, the probability distribution that has the MaxEnt is considered to be the most unbiased and reasonable one. As for low- and high-dimensional spaces composed of the input random variables in MTL, the statistical moments of crosstalk required by the MaxEnt method as its information constraints are obtained through full-factorial numerical integration and sparse grid numerical integration, respectively. Then, a probability distribution model of crosstalk in MTL is established based on the MaxEnt method. Compared with the conventional Monte Carlo method, the proposed method can not only accurately predict the probability distribution and statistics of crosstalk in MTL but also considerably increases the computational efficiency. Therefore, the MaxEnt method is effective in modeling the probability distribution of crosstalk in MTL.