Abstract:
This thesis addresses the decoding of two dimensional (2D) barcodes in industrial environments. A new class of 2D barcodes is introduced that is based on low-density parity-check (LDPC) codes. The special requirements in direct part mark identification (DPMI) applications have been considered during the development in order to design a robust 2D barcode. The major accomplishments of this work include the design and the decoding of both codes, the LDPC code as well as the developed LDPC-based 2D barcode. A design method for short irregular LDPC codes is introduced that yields a better decoding performance for the additive white Gaussian noise (AWGN) channel and the Markov-modulated Gaussian channel (MMGC) compared to optimization methods known from literature. Estimation-decoding and a reestimation procedure of a 2-state channel-model’s transition probabilities have been extended for usage with a MMGC. For the application of LDPC codes on 2D barcodes an intelligent interleaver has been developed to place the code word’s symbols in the available grid of the 2D barcode in order to increase the error-correction capabilities of the resulting 2D barcode. In addition, a 2-state channel-model has been created that provides a good description of 2D barcodes in industrial environments. Considering the 2-dimensional arrangement of the 2D barcode’s modules, the estimation-decoding algorithm has been extended to operate based on a 2D hidden Markov model (HMM). This provided the estimation-decoding in 2 dimensions (ED2D) algorithm that is proposed for the decoding of LDPC-based 2D barcodes. Next to the utilization of regular LDPC codes for the design of the new 2D barcode, irregular LDPC codes have been applied. The latter have been designed by means of the introduced optimization method and the channel-model for 2D barcodes. For the evaluation of the new class of 2D barcodes, a test-environment and an appropriate test-procedure have been created. The test-environment enables one to compare different variants of 2D barcodes in a simulated industrial environment under fair conditions. This test-environment proves that the developed LDPC-based 2D barcode decoded with the ED2D algorithm has a substantially better error-correction performance compared to the standard Reed-Solomon (RS)-based Data Matrix code (DMC).
