Governments around the world are planning the use of cutting-edge AI and robotics technologies to support better care delivered closer to the patient in the community; they can assist patients in playing a greater role in managing their health. Assistive and rehabilitation robots, with the underpinning technologies (machine learning, sensors, actuators, etc.) are being developed to sense and adapt to changing user needs, provide physical support for mobility, rehabilitation, and social engagement.
Recent research has shown potential benefits of using intelligent robots in health and wellbeing, e.g. supporting independent living (Di Nuovo et al., 2018), supporting individuals with autism and intellectual disability (Di Nuovo et al., 2018), the assessment of cognitive skills (Di Nuovo et al., 2019). However, the healthcare sector poses unique challenges for translational research compared to other domains. Devices manage sensitive and private data that must be preserved from being inadvertently disclosed by the AI. This is particularly critical as in a multimodal robotic system, interaction is performed via multiple ways, e.g. video, speech, tablet, gesture, movements, which increase the risk of exposure of private information (Calo, 2020). Indeed, a challenge for healthcare robots is related to users’ low level of trust in the AI that controls the robot, which has a significant impact on how much patients comply with decisions made by the AI.
Currently, there are no established approaches to guarantee data security and privacy in multimodal interaction, in fact, this topic is mostly unexplored and requires further research to identify suitable AI algorithms and to define the protocols for human-machine interaction. Robots are equipped with the ability to sense, process, and record the world around them, and particularly for domestic robots that have access to individual's private homes, the privacy risks have raised concern (Urquhart et al., 2019).
This project will develop intelligent access control and dynamic authentication mechanisms for securing multimodal robotic interaction in healthcare and wellbeing applications. The research will co-design and experimentally validate a Self-Supervised Robotics System (SSRS) with a prototype of an intelligent data manager (IDM), an artificial intelligence component that will monitor the interaction and guarantee conformity to the security and privacy requirements. The IDM will be created with interpretable algorithms that will conjugate the advantages of data-driven models with the trustworthiness of blockchain technologies. Thus, the IDM component is secure by design and user's data protected and the knowledge learned by the robotic system should be preserved and not exposed or disclosed to any unauthorized users intentionally or unintentionally. The expected outcome of this project is to improving data security and privacy of human-robotic interaction to improve trust and support the wider adoption of robotic solutions for health and wellbeing.
The methodology includes the study of computational intelligence and cybersecurity methods and algorithms for the IDM to safeguard real-time interaction with multimodal robots; then, to validate the increased trustworthiness via Proof-of-Feasibility (PoF) study with a robotic health assistant prototype. The IDM will provide verifiable and secure remote communication with the external actors in the healthcare domain (nurses, doctors, caregivers). To preserve the recorded data from external intrusions, the system will also incorporate an efficient blockchain framework to ensure trust and safe adoption by making it tamper-proof, auditable, and transparent during external engagement with an access control mechanism to protect the user’s privacy.
This research project will produce humanly interpretable and verifiable models that will facilitate scrutiny by the user, medical practitioners, allow manual changes to remove algorithmic biases, and support the user’s “right to explanation” that will increase the trust in the robotic system. The research approach will also involve an independent group of stakeholders, with expertise in health and care co-design the requirements for the PoF study and certify the conformity of the prototype to the relevant medical procedures and good practice.
In this project, additional opportunities for training and development will be provided by the Advanced Wellbeing Research Centre (AWRC). The AWRC is a catalyser for cross-sectorial research and innovation in the health and technology sectors with close partnerships with industries, local communities, local authorities, and charities. The AWRC will provide state-of-the-art robotics equipment like Care-o-bot (£190k) and other social robots such as NAO, Pepper, MiRo.
Information on entry requirements can be found on our GTA program page
How to apply
We strongly recommend you contact the lead academic, Dr Jims Marchang ([Email Address Removed]), to discuss your application.
Please visit our GTA program page for more information on the Graduate teaching assistant program and how to apply. Any questions on the graduate teaching assistant programme requirements can be addressed to the postgraduate research tutor for this area which is Dr Marjory Da Costa Abreu ([Email Address Removed]).
Start date for studentship: October 2022
Interviews are scheduled for: Late June – Early July 2022
For information on how to apply please visit our GTA program page
Your application should be emailed to [Email Address Removed] by the closing date of 31st May 2022.