Bidirectional search application for polar codes with large kernels

The paper introduces an algorithm for processing large polarization kernels. The proposed algorithm admits lower decoding complexity compared to other kernel processing algorithms, such as recursive trellis algorithm based processing, on certain kernels. The discussed kernel processor is based on bidirectional trellis decoding of linear codes, where the algorithm grows trees of trellis nodes originating from the first and the last layers of the trellis until these trees intersect.