The communication networks play a crucial role in the healthcare sector in delivering safe and mission-critical health services for the patients. However, the currently deployed communications infrastructure is not designed to tackle the huge surge of patients due to pandemics, such as COVID-19. That necessitates introducing new communication technologies that can foster remote patient treatment, such as remote robotic surgery, to limit the close interaction between the surgeon and the patient. The tactile internet (TI), an evolutionary leap of the Internet-of-Things, is a telecommunications network that aims to enable the remote control of objects employing haptic communications.
The remote robotic surgery via TI can be viewed as a practically viable solution to separate the surgeon and the patient physically. The TI enables the surgeon to perform remote surgery by sending several command and feedback signals through a communication network. It is important to note that the TI-enabled remote healthcare system requires an ultra-low latency of ∼ 1 ms and ultra-high reliability of 99.999%. The fifth-generation (5G) communication systems combined with visible light communication (VLC) technology can partly meet the stringent requirements of TI-enabled remote healthcare.
The VLC systems offer various inherent advantages for the high-speed and short-reach wireless communication systems, such as the availability of a broad license-free spectrum, high data security, low power consumption, low latency, and immunity to electromagnetic interferences. That makes the VLC system a suitable and potential candidate for mission-critical applications in healthcare systems. However, communication-induced artefacts such as packet delay/packet loss may have a detrimental impact on the safety and stability of the TI-enabled remote healthcare system. Recently, there has been an increased thrust towards applying machine learning (ML) techniques to efficiently deal with the communication-induced artefacts in the TI-enabled remote healthcare system.
On this ground, the proposed research work aims to apply various ML approaches, such as Gaussian process regression (GPR) and long short-term memory (LSTM)/gated recurrent unit (GRU)-based recurrent neural networks (RNNs), for the prediction of the packet delay/loss. Besides, the proposed research work also investigates various online training methods to adapt to the time-dependent dynamics of the learning environment. We envisage that this proposed research work on the use of ML in the 5G-VLC-enabled TI network systems can encourage further research innovations for future intelligent healthcare systems.
The various design stages of this research work include the development of the simulation platform for the system modeling of the 5G-VLC-enabled TI network, the implementation and the validation of the considered ML approaches, such as GPR and LSTM/GRU-based RNN, the implementation of the efficient and effective online training methods, and the performance evaluation of the proposed approaches with that of the existing techniques in the literature. Software such as Matlab/Python will be employed for system modeling, whereas the TensorFlow/Keras will be adopted to implement the ML models.
The candidates are expected from the electronics/computing discipline, and a background in communication networks and machine learning is desirable. Also, the candidates should have the ability to pursue independent research and work as part of a team.
Link to full details and how to apply: Design of Intelligent Tactile Internet for Future Healthcare Systems - Doctoral College (ulster.ac.uk)
With 5G wireless communication in roll out phase, there is a significant research momentum on 6G where Artificial Intelligence (AI) will play a pivotal role in the design of Beyond 5G and 6G communication systems. Ultra reliable low latency communication (URLLC) is anticipated to be one of the three fundamental blocks of future wireless research, imposing very confined requirements to achieve very low latency and extreme reliability. In order to achieve low latency and ultra high reliability, the use of AI technology have been proposed that primarily includes radio resource management and selection of radio access technologies.
In this regard, one of the primary objective of this research is to investigate the utilization of AI for solving the significant capabilities and service abilities related with the physical layer of communication systems, which is still an amenable challenge. To name a few, this includes impairment mitigation, modulation, detection, channel estimation and End to End learning while reducing the latency to extremely low value. The primary objective of this research is the realization of AI based channel estimation for (multiple-input multiple-output) MIMO systems in general and massive/ multi user MIMO detectors in particular for end to end learning as novel technologies for future 6G cell free wireless communications systems. Moreover, this includes the optimization of the machine learning (ML) and deep learning (DL) models to reduce the computational complexity and overheads. This optimization problem will be addressed with optimization theory and data driven approaches. The utilization of knowledge driven ML and meta learning approaches will also be investigated.
In the later part of the study, a digital predistortion (DPD) with reduced-complexity will be proposed for the MIMO transmitters. In particular, the understanding of the wireless channel models and state of-the-art receivers is of fundamental importance for benchmarking the novel study item. Furthermore, MATLAB, Python and domestic simulators will be used for the link level and system level implementations along with in house experimental and measurement setups. In addition, dissemination of scientific results through high impact factor journals, transactions and international conferences is expected.
Finally, this PhD study will provide the opportunity to develop technical skills in machine learning, analytical and experimental skills, the attributes required to work with commercial and academic leaders in this field .
The ideal candidate will have a degree or masters in telecommunications, electronic, electrical engineering, computer science or related discipline. Excellent mathematical and analytical skills, experience in programming, at least in Matlab and Python is essential.
Link to full details and how to apply: Edge Intelligence for 5G Networks and Beyond - Doctoral College (ulster.ac.uk)