Software-defined radio implementation of a video transmission system: a study of non-uniform digital constellations

Emerging new wireless technologies and devices to capture and broadcast high-resolution videos in real-time, demand higher data transfer rates. To increase the bit rate, the communication systems always are under development. In the meantime, the other objective is low power consumption, especially for mobile devices like phones and cameras. This achievement requires an optimum design of the elements in a transmission chain. One of the possibilities is optimizing the modulator block. Most of the standard communication systems like DVB-T, use the traditional uniform constellations to map multiple information bits to a constellation symbol. Optimizing the constellation leads to obtain "non-uniform" constellations. In this work, we studied and designed the new "non-uniform" constellations which are used to close the capacity gap to the Shannon limit as much as possible. In this regard, we used the maximization of the BICM capacity approach. BICM, which is a scheme to design a transmission chain and its blocks, tries to increase the code diversity, and because of its flexibility, the channel encoder and modulator can be chosen independently. This makes it a powerful tool to focus on the constellation in the modulator, without worrying about the impact of constellation optimization on the other blocks. By doing this, the improved constellations have a higher capacity which can be translated to gain in other terms like reduction in the power, or, lowering the Bit Error Rate (BER). However, this improvement is not without the cost. Non-uniform constellations will impose another level of complexity to the demodulation procedure in the receiver, because, in contrast to the uniform constellations, the modulation algorithm can not use the symmetry features that are present in the constellation shape. Fortunately, this problem is resolved by designing new methods to simplify the demapping procedure. Another issue is choosing the best performing constellation. Since the SNR factor is present in the channel capacity equation, we have a specific optimum constellation for each SNR value. Therefore, the target receivers and operation parameters of the system must be selected carefully. Nevertheless, the performance evaluations showed the optimized constellations outperform the uniform one regardless of SNR. To evaluate and verify the performance of resulting constellations, first, we used them in a simulation of a full transmission chain, and then, applied them to an SDR implementation of a video transmission system to observe the performance in a real system. The outcome of the evaluation proved the maximum power gain of 0.8 [dB]. One of the most interesting parts of this work is that, in general, thanks to the BICM approach and its flexibility, the new optimized constellations can be used in a wide range of applications and new technologies that are following that design scheme.