Advanced Signal Processing Techniques for Machine Type Communications

The rapid evolution of wireless technology has led to the emergence of new use cases and the deployment of various types of communication devices across broad geographic areas. Notably, the fifth generation of wireless communications is the first technology to natively support enhanced mobile broadband communication (eMBB), ultra-reliable low latency communication (URLLC), and machine-type communication (MTC) services, with future standards expected to follow this trend. To provide cost-effective connectivity in such scenarios, existing cellular network infrastructure can be used to accommodate both the existing and the large number of new devices. However, co-hosting diverse devices with potentially conflicting requirements poses significant challenges for existing signal processing techniques, such as, detection, channel estimation, and decoding algorithms.

While all the services bring forth novel challenges, the integration of MTC into cellular networks poses major drawbacks that need to be addressed before full integration. For instance, MTC is anticipated to generate large volumes of uplink traffic from a massive number of machine-type devices (MTDs) employing non-orthogonal pilot sequences, thereby increasing the complexity of signal recovery. Moreover, due to their energy constraints, MTDs typically require low-complexity transmission schemes to conserve energy, thereby, limiting their ability to perform complex signaling operations.

Motivated by the aforementioned challenges, this thesis focuses on devising novel signal-processing techniques for MTC in cellular networks. Specifically, this work proposes comprehensive solutions that start with the introduction of novel pilot generation strategies to promote the coexistence of MTC with other services. The proposed pilot sequences guarantee low complexity transmissions for the MTDs, thus, contributing to the energy efficiency of the MTDs. Additionally, the present work introduces novel signal recovery techniques specifically designed for massive MTC scenarios, where the MTDs appear in clusters. For this, the approximation error method (AEM)-alternating direction method of multipliers and AEM-sparse Bayesian learning are proposed. In summary, the results presented in terms of channel estimation and device activity detection accuracy demonstrate the significant potential of our proposed solutions in alleviating practical challenges associated with signal processing in cellular networks hosting MTC.