Improving ATSC 3.0 Channel Estimation using AI Techniques

Over the past several years, ATSC has continued to develop new technologies that can be integrated into the standard, making it a leading innovator in the broadcasting industry. The spectrum efficiency of the waveforms and coding, combined with new infrastructure based on the integration of a core network, has proposed ATSC 3.0 as a strong candidate for implementing new multimedia-related use cases. However, the requirements for certain use cases (e.g., AR/VR, 360-degree video) are very harsh, for instance, regarding the signal quality needed at the receiver end. A set of ATSC 3.0 physical layer configurations cannot be implemented with the current channel estimation and equalization solutions. Consequently, Artificial Intelligence (AI) is presented as the necessary enabler for these cases. Specifically, AI has demonstrated better use of the knowledge of the propagation channel than traditional processing methods. Therefore, this paper aims to propose AI-based channel estimation techniques for ATSC. Specifically, two solutions are proposed. The first is a super-resolution estimation enhancer, whereas the second performs the interpolation process within the AI. The results show that both techniques improve the performance of current processing techniques by around one order of magnitude.