New statistical model for burst error distribution

Due to the adoption in high capacity Digital Radio Relay Systems of transmission techniques such as low redundancy error-correcting codes, trellis coded modulations, Viterbi receivers, etc., the errors tend to be clustered in bursts. Up to now, a model used to describe the burst error distribution was the Neyman-A contagious model. This model was conceived to be applied in entomology and bacteriology, but it is also used in telecommunications. So, according to this model, the bursts and the number of errors in bursts are both supposed to be Poisson distributed. However, this distribution does not represent accurately the real situation, so, in order to derive a new probability distribution function, that improves the accuracy of the description of a burst error distribution a new model is required. In this paper a new model based on Erlang distribution for the number of errors in bursts, is described by using the Neyman approach.