ASTUTE PhD Studentships

PhD studentships commencing in October 2019 are available to support research within ASTUTE on the projects below. These studentships provide a maintenance allowance (currently £14,777 per year) and cover tuition fees at the home/EU level.

Applicants from outside the EU are welcome but they will need to pay the difference between the home/EU and international fee levels; that difference is currently £12,290/year. 

AST1: Connected Autonomous Vehicles with Cooperative Sensing and Mobile Edge Computing

Supervisor: Dr. Jianhua He -

Associate Supervisors: Dr. Richard Nock, Dr Xiahong Peng


Connected autonomous vehicles (CAV) is an emerging and promising technology to tackle road safety and pollution problems but also faces many technical challenges. In this project we propose cooperative sensing for CAV's, which exploits the increasingly powerful but underused computing and sensing resources of CAV's to improve sensing accuracy and range. Innovative cooperative sensing based road safety and efficiency applications will be developed and tested, such as cooperative advanced driving assistance system (ADAS) and cooperative platooning. New vehicle to everything (V2X) communication algorithms will be developed to provide a solid basis for information exchange among vehicles and network infrastructure. Innovative research will be conducted on cutting-edge 5G V2X networks, intelligent transport systems (ITS), autonomous driving, deep learning and mobile edge computing technologies. The project has external collaborators, including Huawei, Birmingham City Council (BCC) and Ranplan. It also benefits from international collaboration with the international partners of Horizon2020 project COSAFE (

Collaborators: Ranplan, Huawei, Birmingham City Council

Relevant Degrees: Including (but not limited to) Electronic Engineering, Communication Technology, and Computer Science.

Further information:

AST2: Training Self-Driving Cars using Deep Reinforcement Learning

Supervisor: Dr Diego Faria -

Associate Supervisors: Dr. Maria Chli, Assoc. Prof. Sergey Sergeyev


The pathway to make autonomous driving a truly ubiquitous technology is to ensure functional safety with the ability to effectively respond to unexpected events. This project puts forward the development of an autonomous system with capabilities to navigate in the absence of maps and explicit rules, relying - similarly to humans - on a comprehensive understanding of the environment, while following simple higher-level directions. It is expected that the Ph.D. candidate will design and develop an approach (e.g. Fig. 1) using the local scene captured from images and deep reinforcement learning for autonomous driving with a human in the loop. In addition, knowledge transfer will be exploited from the simulated environment to the real-world application to boost learning performance. A staged approach will be adopted: (i) utilising a simulated environment for training/testing; (ii) transferring the knowledge from the simulated task, and continuation of training in a real environment; (iii) instructing an autonomous vehicle through reinforcement.

Relevant Degrees: Computer Science or Electrical/Electronic/Computer Engineering or similar

Further Information:

AST3: 21st Century Smart and Sustainable Urban Logistics

Supervisor: Dr Marin Marinov -

Associate Supervisors: Mrs Karen Jones, Prof. Ed Sweeney


This PhD project focusses on smart and sustainable urban freight logistics with a particular emphasis on how changing consumer behaviour impacts the supply chain landscape. The overall aim is: to formulate an innovative urban supply chain strategy that will lead to the creation of a more integrated and sustainable freight logistics system.  The identification, mapping  and effective adoption of existing and emerging technologies in a smart city context is a key enabler for this research. A “systems approach” will be adopted with the research student supervised by a multidisciplinary team of academics. The project has three phases: (i) demand side (how changing consumer behaviour is changing the dynamics of urban markets); (ii) supply side (transport and logistics); integration of (i) and (ii) using appropriate modelling tools. The research will develop innovative approaches and concepts and propose new solutions for sustainable urban freight city logistics, using Birmingham as a living laboratory.

Relevant Degrees: Logistics and Supply Chain; Industrial Engineering; Engineering Systems Design; Engineering Management; Operations and Information Management; Mathematics and Computer Science.

Further Information:

AST4: Augmented Reality Agent for Smart Citizens

Supervisor: Dr. Ulysses Bernardet - 

Associate Supervisor: Dr. Peter Lewis


Our goal is to improve the well-being of city dwellers by developing an intelligent augmented reality agent that facilitates  the communication and interaction between citizens and Smart Cities. Heterogenous user interfaces to Smart City systems can make it hard to access relevant information and set priorities, and hence can be a barrier for citizens to interact with services. We will develop an augmented reality Smart Citizen Agent that supports active citizenship, participation, and co-design by facilitating the citizens' access to heterogeneous Smart City services and, allows the Smart City to collect data about citizen well-being, engagement, and the quality of services. The front-end of the Smart Citizen Agent is a humanoid Augmented Reality Agent, i.e. a computer-animated character that is integrated into the user's physical environment. Augmented reality agents are one of the most promising fields of application for virtual reality content and are particularity promising in a Smart City environment.

Relevant Degrees: Computer Science or a related Engineering discipline.

Further Information:

AST5: Road Health Warning and Management System for Smart Cities' Infrastructure

Supervisor: Dr. Yuqing Zhang -

Associate Supervisor: Assoc. Prof. Sergey Sergeyev


Traffic Speed Deflectometers (TSD) are employed by UK Highways England to monitor the national road health conditions and decide the maintenance schedule. A challenge exists around how to process and interpret the TSD laser-based deflection data to provide a more reliable prediction of the road failures (structure bearing capacity, rutting, cracking, potholes etc.). By re-interpreting the TSD data, this project aims to develop an autonomous road health warning system (see figure) to effectively quantify the road structural conditions, achieve advance warning of the developing deteriorations and predict road residual life, in which laser signal processing and mechanics principle-based machine learning methods will be adopted. The implementation of this system will permit preventative maintenance and rational rehabilitation of the national road networks in smart cities, leading to the longest road service life, the least maintenance costs and the minimum delay costs for the road users.

Collaborators: Transport Research Laboratory (TRL), Highways England, KIER, Jiangsu Transportation Institute

Relevant Degrees: Civil Engineering, Engineering Mechanics or Electronic Engineering.

Further Information:

AST6: Reducing air pollutions and greenhouse gases in urban and rural areas using atmospheric solar photocatalysis

Supervisor: Dr. Wei Li -

Associate Supervisors: Dr Kaiming Zhou (Photonics), Dr Sotos Generalis (Mathematics) and Prof Patricia Thornley (Energy)


Gas emissions from human activity are threating our health and environment. For instance, motor vehicles are a significant source of urban air pollutions (e.g. NOx, CO and SO2) and are increasingly important contributors of anthropogenic greenhouse gases (GHGs, e.g. CO2, and NO2). From livestock sector in rural area, the emission of CH4 in particular is also one of the largest sources of GHGs globally. A novel technology, which combines solar up-drafting devices and photocatalysis was proposed recently by Dr de Richter et al to transform those air pollutions and GHGs into benign atmospheric gases. The figure illustrates the principle and some examples. This studentship will involve building a rig, and aim to investigate and optimise the choice of catalysts and their operating conditions for the elimination of different target gases in urban and rural environments.

Collaborators: Pilkington Technology Management Ltd., Dr Renaud de Richter (University of Montpellier, France), Prof Philip Davies (University of Birmingham), and Prof Tingzhen Ming (Wuhan University of Technology, China).

Relevant Degrees: Chemical Engineering, Chemistry, Mechanical Engineering, Material Science.

Further Information:

AST7: Ellipsometric dual comb LIDAR (ELIDAR)

Supervisor: Assoc. Prof. Sergey Sergeyev -

Associate Supervisor: Dr Yuqing Zhang 


The Dual-Comb Spectroscopy (DCS) technique explores two laser pulse trains with slightly different repetition rates toenable detection of the beating (interferogram) between the signal reflected from the moving target (Doppler LIDAR) and reference comb by a single detector (Fig.1). DCS-based LIDARs are unable to classify scene objects especially to recognise humans against a background of objects of the same size or clutter. The successful applicant will work on the novel technique of distance ranging and obstacles signature recognition based on themode-locked fibre laser with the states of polarisation evolving at different time scales.  

Relevant Degrees: Applicants with a Master of Science degree in Electrical Engineering or Physics or equivalent are  especially encouraged to apply.

Further Information:

AST8: Generative models for conterfactual inference in monitoring applications

Supervisor: Dr. Yordan P. Raykov -

Associate Supervisor: Dr Maria Chli


In recent years, we have seen an explosion of machine learning innovations a lot of which have only focused on supervised learning applications with well defined classification objective. In a lot of these problems we have large amount of stuctured data and we assume that every variation of input that can be observed during the test phase of our system would be observed during the training phase as well. However, this is not a realistic assumption to make in large scale monitoring problems where mistakenly applied black box models can lead to wrong causal estimation, mistaken patient diagnosis, dangerous machine response and misleading inference. To address this issue we will seek to build up on the current  understanding of the  theoretical and empirical properties of state-of-the-art methods in real time monitoring applications as well as develop novel machine learning tools for addressing different challenges in monitoring applications. 

Relevant Degrees: Computer Science or related subjects. 

Further Information: [To follow]

AST9 Deep Learning in Graph Domains for Sensorised Environments

Supervisor: Dr Luis Manso -

Associate Supervisor: Dr George Vogiatzis


The In the future, robots and smart environments will cooperate to assist people, performing monitoring tasks to enhance people's comfort and safety. These environments are characterised by a large number of sensors generating vast amounts of interrelated data that are naturally expressed as graphs. The figure depicts a simplified version of the kind of data structure that these environments can generate. The predominant approach to deal with graph structures is two-stage. In the first stage, input data is transformed into vectors. In the second stage, vectors are processed to learn and extract conclusions. Unfortunately, valuable information is generally lost in the graph-to-vector conversion. Techniques natively working with graph-representations are a less explored approach. This PhD will focus on enabling robots and environment sensors to share a common graph-like world model representation and to perform reasoning tasks using machine learning techniques specially designed to work with graph structures.

Collaborator: Ortelio Ltd.

Relevant degrees: Computer Science, Electronic Engineering, Mathematics, Physics or similar.

Further Information:

AST10: Simulating the City

Supervisor: Dr Peter Lewis -

Associate Supervisors: Dr. Antonio Garcia Dominguez and Mr. Ben Senior (Arcadis)

Digital twins are fast becoming an essential feature in understanding how to build and maintain assets in complex, connected, intelligent urban environments. A digital twin is a model: according to the UK National Infrastructure Commission, a “dynamic representation of a system which mimics its real-world behaviour. This will typically be a real-time, updated collection of data, models, algorithms or analysis. A digital twin is a virtual representation of a physical object or system across its lifecycle, using real-time data to enable understanding, learning and reasoning”. In the academic literature, a digital twin is considered to be "a digital replica of a living or non-living physical entity”, which exists simultaneously with the physical entity. This project will tackle the challenge of how to build and unlock the benefits of large-scale digital twins. It will be driven by the aspiration to build a digital twin of Birmingham, which will serve to expose the challenges and viable solutions needed to do so.  

Collaborator: Arcadis

Relevant Degrees: Computer Science, or a related discipline

Further Information:

AST11: Epidemics, marketing and opinion setting - competitive and collaborative spreading processes

Supervisor: Prof. David Saad -

Associate Supervisors: Roberto Alamino (Mathematics, specialising in networks), Maria Chli (Computer Science, specialty in multi-agent methods), Bo Li (Mathematics MSc Fellow, working on spreading processes), Simon Thompson (BT Research, Head of Practice, Big Data & Customer Experience)


Epidemic spreading has existed as long as life itself. Alongside their impact on community health, spreading processes became a modelling tool for opinion-setting, information dissemination and marketing on social and information networks. Information, supporting ideas or products, is passed between network constituents in a probabilistic process akin to epidemic spreading, where competing news/advertisements compete on the largest audience. On the other hand, some processes exploit previous exposure to concepts/diseases to facilitate the effective spread of new information/diseases, e.g., the risk of developing tuberculosis is 16-27 times greater in HIV carriers. Optimising the use of limited resources for maximising spread, blocking hate/fake information or the containment of infectious diseases is of great societal relevance. We will analyse interacting spreading processes and develop principled algorithms for resource optimisation. We will apply the methodology to applications that are of interest to our industrial collaborator-BT, including containment, mitigation and source identification of malware infections.

Collaborator: BT

Relevant Degrees: Physics, Mathematics, Computer Science or similar

Further Information:

AST12: Analysing drone footage for detecting change in man-made environments

Supervisor: Dr. George Vogiatzis -


One of the major applications of aerial surveillance is the monitoring and detection of change in man-made environments. The causes of change can be attributed to climate variations, plant and wildlife, human activity and natural disasters, each of which must be monitored for a variety of scientific, economic or regulatory reasons. Focusing on urban and semi-urban landscapes, there is growing interest in monitoring how these change as a result of construction, for planning purposes, but also as a result of natural disasters (fire, floods, earthquakes) for the purposes of coordinating disaster response or settling insurance claims. In this project we will work with Geospatial Insight Ltd to investigate the use of drones for the automatic surveillance and change detection in a man-made environment. We will be building an automated surveying system that would revolutionize the change monitoring of urban scenes by lowering the costs and increasing the accuracy.

Relevant Degrees: Computer Science, Engineering, Maths or a related discipline.

Further Information:

AST13: Coached Knowledge Transfer for Deep Reinforcement Learning

Supervisor: Dr. Maria Chli -

Associate Supervisor: Dr. George Vogiatzis


In this project we will work with to radically change the way we currently train and deploy Artificial Intelligence (AI) agents in real-life problems arising from Logistics and Robotics. Inspired by the way humans learn in complex, challenging domains such as sport, we are putting forward a very different approach to Deep Reinforcement Learning (DRL). Instead of developing new agents and training them from the ground up for each new problem encountered, this PhD project will develop a reliable framework for knowledge transfer between widely varying tasks. This will then allow us to create an agent “coaching” architecture similar to the way athletes train on a variety of tasks that lead them to mastery of their sport. Our DRL agents will be accumulating their knowledge as they are trained through a specially designed curriculum of tasks, eventually gaining expertise in their target task faster and more reliably than “from scratch” training.


Relevant Degrees: Computer Science, Engineering, Maths or a related discipline.

Further Information:

AST14: Personalised digital health

Supervisor: Dr. Maria Chli -

Associate Supervisor: Dr. George Vogiatzis

In collaboration with RowAnalytics Ltd, this project will build an, IoT enabled, intelligent patient monitoring and response system, where complex edge analytics and disease knowledge models are combined with a patient’s health data to provide fully personalized real-time advice. The student will leverage our previous successes in deep reinforcement learning for real-time control problems to develop new learning architectures that will allow fast convergence of networks despite the presence of non-stationarity. The complexity introduced by the highly non-stationary sensor inputs will be reduced with state of the art probabilistic models. At the same time the student will extend our deep reinforcement learning transfer work in a multi-agent setting. We envision that this will empower seamless cooperation and coordination in distributed sensor and actuator networks, ultimately leading to efficient decision-making in these challenging decentralised settings. 

Collaborators: RowAnalytics Ltd

Relevant Degrees: Computer Science, Engineering, Maths or a related discipline.

Further Information: