Research Opportunities

5G will become the dominant connectivity for many wireless devices, including smart phones, communications module embedded on robots, forklifts, etc. These wireless devices are usually enabled with other wireless connectivity solutions, such as Wi-Fi, Bluetooth, etc. This project will study and propose 5G-enhanced indoor positioning solutions. Wireless signals from the multiple communication protocols will be used to derive high-accuracy indoor positioning estimation. Performance study will be conducted, and enhancement algorithms based on, e.g., machine learning or deep learning, will be proposed and evaluated in realistic real-world scenarios.

Students will have the opportunity to conduct inter-disciplinary research that combines signal processing and machine learning. Moreover, they will also get to work with industry partners to apply their research output on products that are ready to be commercialised.

University Supervisor
Professor Sun Sumei
Sumei.Sun@singaporetech.edu.sg
View Profile

In the era of 5G, machine learning (ML) is becoming more and more prevalent for supporting many different applications for industry 4.0, autonomous driving and intelligent transportation. Current ML has mostly taken a centralized approach which requires all the data to be aggregated on a central sever for training. This introduces significant cost for data transmission. In addition, the privacy of the data may be sacrificed.

This project intends to explore the possibility of using federated learning to address the above issues, in different settings such as in industry 4.0, health care, automobile traffic network or in a setting where there are several robots collectively working on an assignment. For instance, if a car is experiencing a traffic situation that can be of relevance to other vehicles in the vicinity, an updated traffic model based on the current situation will be uploaded to the cloud and this becomes a shared model for other vehicles that may be experiencing a similar situation.

University Supervisor
Associate Professor Benjamin Premkumar
Benjamin.Premkumar@singaporetech.edu.sg
View Profile

5G has become increasingly important in supporting many critical industry systems by enabling functions such as train signalling, drone navigation, and robotic control. However, the undesired radio frequency (RF) interference due to jamming, spoofing, or malfunctioning of wireless devices could lead to performance degradation, service disruption or even failure of such systems. Therefore, real-time monitoring and detection of RF interference is very important for the safe and reliable operation of such systems that rely on 5G.

University Supervisor
Associate Professor Pei Yiyang
Yiyang.Pei@singaporetech.edu.sg
View Profile

Moving towards 5G and beyond, the next-generation network (NGN) use-cases such as tactile Internet, aerial vehicles/drones ad-hoc networks, etc., are the key enablers of a smart nation. Establishing the 5G ecosystem is a key enabling strategy and an essential building block of Singapore’s Smart Nation initiative. Singapore is progressively deploying the 5G core infrastructure across the island to move from 5G non-standalone (NSA) to standalone (SA) 5G speeds. This essentially means that there are going to be fundamental changes to the core network architecture such as network virtualisation, software-defined networking and cloud-based edge technologies, to cater to the wide variety of 5G use-cases.

In 5G networks, diversified services on different market verticals need to be provisioned as end-to-end services distributed across mobile and stationary devices, as well as edge and core data centers with differentiated QoS requirements. For example, the enhanced mobile broadband (eMBB) services need to support transmissions of large payloads at peak data rates of 20 Gbps while the ultra-Reliable and Low Latency Communications (uRLLC) services require transmissions of small payloads with very high reliability. This fragments the market verticals and brings up the need for an end-to-end dynamic network slicing through real-time traffic characterisation through the 5G Core, Cloud, and Edge networks.

The outcome of this project aims to contribute to the realisability of diversified 5G applications across different market verticals, focusing on anomaly detection and network security. The solution will be tested at the 5G testbed setup at SIT in collaboration with IMDA and major telco companies in Singapore. Our telco industry partners can leverage our testbed to analyse various resource constraints and security gaps when supporting diversified services – moving towards an end-to-end 5G infrastructure.

University Supervisor
Assistant Professor Purnima Murali Mohan
Purnima.Mohan@SingaporeTech.edu.sg
View Profile

The roll-out of 5G across the globe has created a lot of excitement for both consumers as well as businesses with the promise to enable a range of new use‐cases. With the capabilities of enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC) and massive machine‐type communication (mMTC), 5G could bring about new applications not possible with earlier mobile communication technologies.

However, it is unlikely that 5G will completely replace other wireless communication technologies such as WI‐SUN, Wi‐Fi, Zigbee and many others. These technologies offer unique advantages for different applications. In such a heterogenous operating environment, 5G could potentially provide last-mile connectivity to bridge these wireless communications to the Internet. We envisage such architecture to be prevalent for a variety of use cases. The interconnectivity between 5G and legacy wireless networks (5G‐X) is essential to address the needs of different applications and communication requirements. The co‐existence and inter‐networking of 5G‐X in an IoT environment such as a smart building can enhance the service offerings, thus optimising bandwidth, system operations and enhancing system and data security.

This research envisions a smart building that is equipped with hybrid connectivity consisting of 5G networks interconnected with a Wi‐SUN network. Two modes of 5G‐X connectivity, namely (1) 5G‐Wi‐SUN, (2) 5G‐60GHz high-capacity link will be investigated, developed, and demonstrated through innovative applications related to persuasive energy management, video analytics, direct device control, indoor localisation, and artificial intelligence applications.

The overall performance will be studied for these applications that require near real‐time processing.

University Supervisor
Associate Professor Forest Tan
Forest.Tan@SingaporeTech.edu.sg
View Profile

The roll-out of 5G across the globe and its potential to enable a range of new use cases has created a lot of excitement for both consumers as well as businesses. With the capabilities of enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC) and massive machine‐type communication (mMTC), 5G could bring about new applications that were previously not possible with earlier mobile communication technologies. Beyond these important enhancements, the latest Release 17 also sees several new additions such as the support for device‐to‐device (D2D) communication and the increase of the frequency spectrum in mmWave up to 71 GHz.

In this research, the candidate will delve into research and development to enable 5G devices to perform direct D2D communication without the need for a central arbitrator.

The candidate will also investigate the use of this feature to explore new and innovative use cases that were previously not possible. D2D communication involving robots, robot-to-infrastructure, and human‐to‐machine communication via 5G is essential to enable proximity‐based communication and direct control of IoT devices without relaying the information to a base station and/or server.

With the ultra‐fast speed of 5G-to-5G connectivity, users can share files directly without going through a network server. With the D2D communication capability using 5G, the location or position of IoT devices and robots in smart buildings can be determined, thus facilitating real‐time navigation and collision avoidance in an indoor GPS-denied environment. 5G Release 17 has included what is known as sidelink expansion to support D2D communication. It will bring a slew of new and enhanced capabilities to expand sidelink to new use cases such as public safety, IoT and many others.

University Supervisor
Associate Professor Forest Tan
Forest.Tan@SingaporeTech.edu.sg
View Profile