EPSRC TinyML UK Network

Coming Soon.

• Client: Innovation Driven Procurement
• Category: Education

Close Project

ESPRC TinyML UK Network

Comping Soon

• In Partnership with: Edge AI Foundations
• Category: Techncology & AI

Edge AI Foundation

Tag With Me

TagWithME.com

The ubiquity of mobile devices make them ideal platforms for exercise games (exergames) to promote physical activities. Advances in Internet of Things (IoT) technologies including Bluetooth Low Energy (BLE) beacons can be utilised for proximity detection to promote physical activities and the use of Artificial Intelligence (AI) in the form of object recognition can accelerate engagement with location-based pervasive games.

We are currently implementing a casual exergame, Tag With Me, in the form of a treasure hunt that provides the approximate location of nearby points of interest in real-time within the vicinity of Bluetooth beacons. The system exploits the signal strength of the BLE beacons to measure proximity which makes it suitable for outdoor and indoor functioning where GPS signals are not accessible. Once the player walks towards a point of interest one of the interactive challenges is activated such as the AI and camera challenges where the player must scan the nearby object with their smartphone camera, the tag challenge where the player must search for a tag and tap it or an education quiz.

• Client: Rufford Abbey Country Park, Highbury Hospital
• Category: Health and Wellbeing

Rufford Abbey Counrty Park and Nottinghamshire County Council

DigitalExposome

Multimodel Sensor Fusion Approach to study the impact of Environment Pollutants on Mental Wellbeing

The short and long term exposure to environmental urban factors (such as air pollution, gases, particulates and noise) can significantly impact an individual’s wellbeing and mental health. The World Health Organisation (WHO) found that 91% of people are living in places where the air quality guidelines are not met and the use of non-clean fuels and household emissions in the atmosphere are causing over 4.2 million deaths each year. In addition, those living in some locations in the UK have a higher risk of developing serious health conditions such as higher heart rate, asthma and cardio-cerebrovascular disease where a lifetime of exposure to high-levels of pollution can result in reduced life expectancy.

Repeated and continuous human exposure to the environment and highconcentrated air pollutants have been found to increase the risk of developing serious conditions such as respiratory and cardiovascular diseases or even death. Research recently has began focusing towards how the environment can impact physical health but it also is necessary to explore how the environment can impact mental wellbeing. Pollution within the urban environment is a continual problem contributing to rising health and mental wellbeing challenges. The ability to monitor air pollutants, physiology and mental wellbeing will enable the relationship between repeated environment exposures and mental wellbeing to be established.

The the term ’DigitalExposome’ as a framework to quantify an individual’s exposure to the environment by utilising a range of technological, mobile-sensing and digital devices. The concept aims to measure multiple environmental factors using mobile technologies and then quantify them in real-life settings. Combining multiple data collection methods helps to support DigitalExposome and gain a better understanding into how exposures to the environment can impact mental wellbeing.

Voronoi visualisations have given an indication of how changes within the environment can have an impact on mental wellbeing. Typically, it was found that where air pollution such as PM1, 2.5, 10 and Noise was increasing, participants labelled their wellbeing as very negative. This demonstrates consistent results with previous studies in this area. This form of spatial analysis, greatly helps in understanding the degree to which a place is similar to other nearby places.

Statistical analysis including PCA, Multi variant Linear Regression, Voronoi and data spatial visualisations were implemented to explore the variation in data and the factor importance. We found that physiological (on-body) sensor data is directly correlated to pollution (PM in particular) within the environment. In addition, DBNs have helped successfully classify five states of wellbeing with up to 80.8% accuracy using the fused physiological and pollution data.

• Category: Digital Monitoring

Close Project

Tangible Fidgeting Interfaces

Tangible Fidgeting Interfaces for Mental Wellbeing Recognition using Deep Learning applied to Physiological Sensor Data

The momentary assessment of an individual's affective state is critical to the monitoring of mental wellbeing and the ability to instantly apply interventions. This research introduced the concept of tangible fidgeting interfaces for affective recognition from design and development through to evaluation. Tangible interfaces expand upon the affordance of familiar physical objects as the ability to touch and fidget may help to tap into individuals' psychological need to feel occupied and engaged. Embedding digital technologies within interfaces capitalises on motor and perceptual capabilities and allows for the direct manipulation of data, offering people the potential for new modes of interaction when experiencing mental wellbeing challenges.

Tangible interfaces present an ideal opportunity to digitally enable physical fidgeting interactions along with physiological sensor monitoring to unobtrusively and comfortable measure non-visable changes in affective state. This opportunity initiated the investigation of factors that would bring about the designing of more effective intelligent solutions using participatory design techniques to engage people in designing solutions relevant to themselves.

Adopting an artificial intelligence approach using physiological signals created the possibility to quantify affect with high levels of accuracy. However, labelling is an indispensable stage of data pre-processing that is required before classification and can be extremely challenging with multi-model sensor data. LabelSens introduced new techniques for labelling at the point of collection using five custom built tangible labelling interfaces.

When classifying labelled physiological sensor data, individual differences between people limit the generalisability of models. To address this challenge, a transfer learning approach has been developed that personalises affective models using few labelled samples. This approach to personalise models and improve cross-domain performance is completed on-device, automating the traditionally manual process, saving time and labour. Furthermore, monitoring trajectories over long periods of time inherits some critical limitations in relation to the size of the training dataset. This shortcoming may hinder the development of reliable and accurate machine learning models. A second framework has been developed to overcome the limitation of small training datasets using an image-encoding transfer learning approach.

This research offered the first attempt at the development of tangible interfaces using artificial intelligence towards building a real-world continuous affect recognition system in addition to offering real-time feedback to perform as interventions. This exploration of affective interfaces has many potential applications to help improve quality of life for the wider population.

• Category: Wellbeing Monitoring

Close Project

DigitalPPE

The COVID-19 Pandemic brought about closure to work spaces, public venues and academic environments. These works investigated a solution that enables a safe exit strategy out of restrictions. Through the use of an IoT based Bluetooth Low Energy (BLE) device, individuals were reminded of the social distancing measures, facilitated through visual and haptic feedback. The device also enabled individuals to monitor close contacts, recorded on a database. The technology developed provides the additional benefit of no further equipment for contract tracing being provided. It also provides real-time alerts for larger gatherings, something that alternative systems of the time lacked in.

• Category: Health and Safety

Close Project

Prof. Eiman Kanjo

eiman.kanjo@ntu.ac.uk

Professor of Pervasive Sensing

Prof. Kanjo is currently a Professor of TinyML &Pervasive Sensing She is currently the director of the EPSRC Green+ Network and the director of the EPSRC TinyML UK network, and Co-lead of EPSRC ProSensing project. She is the Provost Visiting Professor at Imperial College London, Computing Department. Eiman was recognised as one of the top women in Engineering by Women in Engineering Society.

Prof. Kanjo was first to coin the phrase 'Mobile Sensing' and wrote some of the earliest papers on the subject ( GeoMobSens and Mobsens). She also built the first noise monitoring system using the phone-based microphone (NoiseSpy). Her current work (EnvBodySens) on studying and quantifying the impact of Environment on wellbeing using Mobile Sensing, Deep Learning, Data Science and AI, complements her work on Urban Computing in order to make sense of a place ( NeuroPlace, ShopMobia).

Eiman is also an expert in developing digital technologies for Mental Health and she has been involved in a wide range of projects in this area, including (NotiMind). She works closely with Mental Health networks and charities and currently developing novel Fidgeting Interfaces to reduce Anxiety and Stress among adult and school children. She often employs data science and on-device processing (on Edge Computing) to create privacy preserving pervasive tools that can transform wellbeing.

Eiman's team is active in developing tools and solutions to minimise the risk of COVID19, and recently developed a low-cost wearable for social distancing and contact tracing.

Awards:
• Top 50 Women in Engineering
• Edge AI Foundation Year Award 2025
• Vice Chancellor's Outstanding Research Team Award

Close Team Member