Performance of Bust-Trapping Codes

An adaptive error control scheme called burst-trapping has been proposed as a means to combat errors in a channel where both random and burst disturbances occur. 1 Such a channel is called a compound channel. Error statistics derived from field trial data over telephone channels indicate that the switched telephone network is a prime example of a compound channel. It is of interest to see how the proposed technique performs on telephone channels and to compare its performance with other known techniques. In this report, a computer simulation is used to determine the performance of such codes. A program has been written which simulates any burst-trapping code of rate (b -- 1 )/b, b an integer. The program was used on the three well-known sets of telephone error data: the 477 478 T H E BELL SYSTEM TECHNICAL J O U R N A L , APRIL 1070 Alexander-Gryb-Nast (AGN) data, 2 the Townsend-Watts (TW) data 3 and the Vestigial-Sideband (VSB) data. 4 The output of the simulation program gives both the bit and block error rate before and after the application of burst-trapping error control. The simulated block error rate with a large interleaving degree shows an excellent agreement with the result of an analytical approach presented in the companion paper. 5 Since two independent approaches are used, the agreement seems to indicate both techniques are accurate. The analytical technique gives the tail of distribution which cannot be produced by the simulation without a large amount of raw data.