Mission Critical Communications (MCC) are all communication related to the safety and security of the civil society such as public safety services, including police forces, firemen, rescue and ambulance services, or employee critical infrastructures, like energy and transportation suppliers. MCC are conveyed by Professional Mobile Radio (PMR) networks. Such services being dispensible in critical circumstances; group communication, wide-coverage, security from eavesdroppers become indispensable requirements from MCC. In this work, I focus on the design of MCC systems to facilitate these inherent needs.
Wireless communication is prone to attacks from malicious eavesdroppers due to its broadcast nature. Physical layer security is a powerful tool that utilizes concepts of signal processing to provide security at the physical layer of the network. It mainly exploits the characteristics of the eavesdroppers' channel for the benefit of legitimate users. In MCC, the need for secrecy is crucial than ever to prevent from any misleading effects of compromised security in critical applications. Thus at Telecom-Paris, I work on developing physical layer-secured transceiver designs for MCC. The secrecy is added against eavesdroppers by various means, including artificial noise-aided security, security through MIMO beamforming, and by considering imperfections in available knowledge of eavesdroppers' channels. The concepts developed during this work can also be adapted in other wireless communication applications.
With its ability to learn complex non-linear functions to perform a task, machine learning has found its way into the domain of wireless communication. It has become an active participant in a vast range of research problems, including resource allocation, channel estimation, transceivers filter design, localization, etc. I am utilizing machine learning, specifically deep neural networks, for the millimeter-wave base-station and beam selection in a multi-base-station heterogeneous (sub-6GHz and millimeter-wave) network architecture. The learning is performed by using the channel characteristics from only the sub-6GHz legacy base-stations.
Millimeter wave (mmWave) communication is a promising technique for next generation wireless (5G and beyond-5G), as it can accomodate a large number of users and offers data rates in Giga-bits-per-second (Gbps) due to the availability of huge unlicenced bandwidth at mmWave frequencies. However, it is associated with many challenges such as free-space path loss, reduced coverage as it is susceptible to blockages, high cost and power consumption etc. Thus for efficient implementation of mmWave technology and to avail its benefits for next generation wireless networks, it is important to overcome these associated challenges. Driven by this, in my work, I focussed on design and evaluation of low complexity mmWave transceiver designs to achieve reduced hardware and computational complexity for various communication scenarios which includes multi-user MIMO downlink, MIMO interference channel, half-duplex (HD) and inband relay-assisted mmWave communication.