Locally Invertible m-Dimensional Convolutional Codes

A new approach for decoding multidimensional (mD) nonsystematic convolutional codes is described. The possibility of coding at different rates along the dimensions of the sequence space is explored and the notion of ordering of information and encoded sequences based on coding rate is explained. Using this technique, a class of fast-decodeable, locally invertible m-D convolutional codes based on the concept of one-to-one mapping between information and encoded subsequences of equal size is defined. Error syndrome generation using orthogonal paritycheck sums is described and methods for error correction using error syndromes are suggested.