Soft output decoder for convolutional codes

A method of decoding a received convolutionally coded sequence of signals represented by a trellis of block length N, comprising the steps of:

• a) dividing the trellis into sections of length L;
• b) decoding using a generalized Viterbi algorithm a portion of the trellis using forward recursion from a point P1 that is before the beginning of each section up to the beginning of that section, wherein P1 is a function of a constraint length of the convolutional code and is chosen at a sufficient distance from the beginning of the section such that forward recursion determines a known state metric at the beginning of the section;
• c) decoding using a generalized Viterbi algorithm a portion of the trellis using backward recursion from a point P2 that is after the end of each section back to the end of that section, wherein P2 is a function of a constraint length of the convolutional code and is chosen at a sufficient distance from the end of the section ection such that backward recursion determines a known state metric at the end of the section;
• d) decoding using a generalized Viterbi algorithm the portion of the trellis within each section using a forward recursion starting from the known state at the beginning of the section defined in step b) and a backward recursion starting from the known state at the end of the section defined in step c);
• e) calculating a soft output at each stage of the section using the forward recursion state metrics defined in step d), the backward recursion state metrics defined in step d), and branch metrics at each stage, and outputting the soft output for that stage;
• and
• f) performing the forward and backward recursions of steps b) and c) and the soft output calculations of step e) with a sliding window decoding method within the section, wherein the sliding window decoding method is characterized by a window length A, wherein A is less than L.

A radiotelephone with a receiver and demodulator for processing a convolutionally coded sequence of signals represented by a trellis of block length N divided into sections of length L in a frame buffer by a soft-decision output decoder, the output decoder comprising:

• a memory;
• a forward recursion processor decodes a portion of the trellis using a learning forward recursion from a point P1 that is before the beginning of the section up to the beginning of the section, wherein P1 is chosen at a sufficient distance from the beginning of the section such that forward recursion determines a known state metric at the beginning of the section;
• a backward recursion processor decodes a portion of the trellis using a learning backward recursion from a point P2 that is after the end of the section back to the end of the section, wherein P2 is chosen at a sufficient distance from the end of the section such that backward recursion determines a known state metric at the end of the section,
• the backward recursion processor subsequently decodes the portion of the trellis within the section using backward recursion starting from the known state at the end of the section back to the beginning of the section to define a set of known backward recursion state metrics within the section;
• the forward recursion processor subsequently decodes the portion of the trellis within the section using forward recursion starting from a known state at the beginning of the section and moving forward to the end of the section to define a set of forward recursion state metrics within the section;
• a decoder calculates a soft output at each stage of the section using the forward recursion state metrics, the backward recursion state metrics, and branch metrics at each state, and outputs the soft output for that stage;
• and
• performing the forward and backward recursions of the soft output calculations of within a sliding window decoding method within the section wherein the sliding window decoding method is characterized by a window length A, wherein A is less than L.