New Convolutional Codes--Part III

This paper is the last part of a three-part series on threshold decoding new convolutional codes. The first part of this paper concludes the list with new rate 1/2 codes. For these codes a characteristic equation, in terms of the number of correctable errors and code constraint length, is derived by least square approximations. The second part of the paper is concerned with the usefulness of the codes derived including those in Parts I and II. Based upon the uncorrectable error statistics of the decoder, the concatenation of such codes is suggested. The characteristic of this class of codes with the property of not producing additional and bursty errors at the output of the decoder when the capability of the decoder is exceeded, is not shared by most powerful decoders such as Reed-Solomon, BCH, and Viterbi.