Simplified trellis decoding of block codes by selective pruning

As channel conditions improve we can augment data throughput by using error-correcting codes with higher coding rates. However at these rates convolutional codes suffer from performance loss due to puncturing and can be difficult to implement. On the other hand the trellis based decoding of block codes is computationally intensive. For this reason we introduce an innovative method of simplifying the decoding of block codes by way of selective trellis pruning. This algorithm can produce quasioptimal results with only a fraction of the operations required by conventional methods. We also introduce a way of obtaining a pruned trellis directly from a simplified generator matrix. By using this method one can avoid the construction and pruning of full trellises, which for some codes can be extremely large, thus simplifying decoding even further.