A decoding scheme calledw-distance decoding (or weighted erasure decoding) has been studied for decoding binary block codes onQ-ary output channels by computer simulation and analytic derivation of the probabiliiy of error bound. Optimum distribution ofw-weights and the optimum threshold level of quantization are obtained by both simulation and minimization of the probability of error bound. The asymptotic behavior (signal-to-noise ratio\rightarrow \infty)of the error bound is determined by numerical methods with the help of a digital computer.

A Markov gap model is proposed for real digital communication channels. Analytic properties of the model are investigated and compared with data from VHF and troposcatter channels. The model is also simulated, and the satistics of the simulated data are compared favorably with those of the real data.

The simple partitioned Markov chain model is used to evaluate the effectiveness of burst error correcting codes. A Massey diffuse convolutional code is analyzed as an example to illustrate the method. Calculated results and simulation results are presented.

A simple partitioned Markov chain model is proposed for characterization of the HF channel. It is demonstrated that from the model, the error-burst distribution can be derived with good agreement with simulation results. In this paper the transition probability matrixPof the model is raised to theNth power PN. PNis used to represent the channel after interleaving of degreeN. The error-burst distribution after interleaving is derived from PN. The results show again that there is good agreement between calculated and simulated error-burst distribution. It is concluded that the model can be employed to study the effect of interleaving on burst-error channels.