Impact case study database
Search and filter
Filter by
- University of the West of England, Bristol
- 12 - Engineering
- Submitting institution
- University of the West of England, Bristol
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Connected Autonomous Vehicles (CAVs) can function without a driver and can communicate wirelessly with the surrounding infrastructure, and each other, to optimise passenger comfort, energy consumption and road network efficiency. Research at the University of the West of England has informed:
the approach to CAVs in the legal and insurance sectors;
development of safety standards;
industrial technology development;
local government public engagement around CAV technology.
UWE research into CAV technology, safety requirements and transport engineering issues, has increased the readiness of key stakeholders to deal with upcoming opportunities and challenges.
2. Underpinning research
CAV technology Starting in 2015 ( G1), an industry-academic consortium was amongst the first in the UK to implement fully integrated CAVs, operating in dense urban settings. UWE led technology integration into a converted Land-Rover car and a dual-seater Renault Twizy electric car. The UWE research demonstrated CAV technology operating safely in complex urban roadway scenarios, such as waiting for an oncoming cyclist to pass before overtaking a parked car. The research included:
A multimodal camera/radar/lidar sensing and perception subsystem;
A low-level vehicle controller using these perceptions to successfully modify local behaviour;
A high-level Decision-Making System (DMS) utilising a robust novel behaviour tree approach;
Wireless communications, e.g. the perception system above was mounted on a stationary bus, capable of passing traffic information to other vehicles in the area;
An immersive ‘hardware and human in the loop’ simulation suite.
Academic publication of the technology details of this work was initially restricted by commercially sensitive IP ownership of critical parts, but it is now summarised in ( R1).
Societal focus on use of CAV technologyFrom 2016 ( G2), UWE CAV research had an additional focus on low-speed, SAE level-5 fully autonomous ‘pods’. UWE led the FLOURISH project research on defining and implementing technology requirements for a pod’s Human-Machine Interface (HMI). This work predominantly focused on the requirements of older adults. The results suggested that participants prefer adaptive HMIs, with journey planner capabilities. There was also a strong preference for additional information and entertainment functions ( R2).
In projects G1, G2, G3 and G4, UWE led investigations into transport engineering, user-acceptance, and emerging legal and insurance issues. These projects looked at the interactions of a variety of CAVs with other types of road users, including buses, cyclists and pedestrians. Published outputs from these projects included a detailed study of vehicle control handover, which found that people often relax into ‘passenger’ mode quite quickly - posing significant issues for re-assuming control of the vehicle ( R3). A study of Local Shared Automated Vehicle Services (LSAVS) found that social considerations such as equity in access to mobility services, social inclusion, environmental protection, and concerns about control over interpersonal interactions were strong acceptance factors. Broad socio-political aspirations beyond transport policy were also found to be important. The study concluded that high levels of social acceptance would be required before introduction of LSAVS would be feasible ( R4).
Methodologies for ensuring CAV safety Starting in 2015 ( G1), UWE researchers led the creation of operational safety-cases for CAVs operating, initially, in UWE’s on-campus roadways. Later, UWE were a major contributor to safety-cases for CAVs operating in public areas ( G2, G3, G5, G6).
Since 2013 ( G7), researchers had been conducting research into ensuring the safety of close-proximity physical interactions between robots and humans in the context of the care of elderly or infirm adults. One of the main aims of this research was to adapt and transfer established Verification and Validation methods used in the silicon industries to the vastly more complex setting of human-robot interaction. The outcome was a hierarchical computer architecture that could verify and validate the correctness, safety and acceptability of the robot’s behaviour by efficient use of simulation. UWE researchers were able to achieve acceptable coverage of the state-space, whilst reducing the need for expensive and safety-critical real-world experiments, which could then be used in a more targeted way ( R5).
It was clear that this line of research could be useful in the CAV sector, because an unfeasibly large real-world distance would have to be covered before one could confidently state that close to all significant risks had been experienced by the vehicle. Because of this research, investigating the role of simulation in ensuring CAV safety, and developing tools to achieve it, became the main role of the UWE researchers in more recent CAV projects ( G3, G5, G6). UWE researchers have recently embedded a Multi-Agent System into the simulation architecture, allowing agents to act together antagonistically so as to create interaction scenarios that maximise stress on the vehicle controller under test. This has doubled the simulation system’s ability to generate effective tests compared to pseudo-random generation, while being time-efficient and robust ( R6).
3. References to the research
R1 Kent, T., Pipe, A., Richards, A., Hutchinson, J., and Schuster, W. (2020) A Connected Autonomous Vehicle Testbed: Capabilities, Experimental Processes and Lessons Learned. Automation, 1(1):17-33. https://doi.org/10.3390/automation1010002)
R2 Voinescu, A., Morgan, P., Alford, C., and Caleb-Solly, P. (2018). Investigating older adults’ preferences for functions within a human-machine interface designed for fully autonomous vehicles. Lecture Notes in Artificial Intelligence, 10927 LNCS, pp 445-462. https://doi.org/10.1007/978-3-319-92037-5_32
R3 Morgan, P., Alford, C., Williams, C., Parkhurst, G. and Pipe, T. (2017), Manual takeover and handover of a simulated fully autonomous vehicle within urban and extra-urban settings. In: Stanton, Neville A. ed. Advances in Human Aspects of Transportation: Advances in Intelligent Systems and Computing, vol 597 pp 760-771. http://dx.doi.org/10.1007/978-3-319-60441-1_73
R4 Paddeu, D., Shergold, I. and Parkhurst, G. (2020). The social perspective on policy towards local shared autonomous vehicle services (LSAVS). Journal of Transport Policy, vol 98, pp 116-126. https://doi.org/10.1016/j.tranpol.2020.05.013
R5 Webster, M., Western, D., Araiza-Illan, D., Dixon, C., Eder, K., Fisher, M. and Pipe, A. (2019) A Corroborative Approach to Verification and Validation of Human–Robot Teams, International Journal of Robotics Research, vol 39(1) pp 73-99. https://doi.org/10.1177/0278364919883338
R6 Chance, G., Ghobrial, A., Eder, K., Lemaignan, S. and Pipe, T. (2020) An Agency-Directed Approach to Test Generation for Simulation-based Autonomous Vehicle Verification. 2020 IEEE International Conference on Artificial Intelligence (AITest). 3-6 August. Oxford: UK. https://doi.org/10.1109/AITEST49225.2020.00012
Evidence of the quality of the underpinning researchG1 Pipe, T. Driverless Cars (VENTURER), Technology Strategy Board, 2015 – 2018, £594,636 (UWE Project Funding). G2 Pipe, T. FLOURISH, Innovate UK, 2016 – 2019, £433,991 (UWE Project Funding). G3 Parkhurst, G. CAPRI, Innovate UK, 2017 – 2020, £356,198 (UWE Project Funding). G4 Parkhurst, G. MultiCAV, Innovate UK, 2018 – 2022, £309,934 (UWE Project Funding). G5 Pipe, T. ROBOPILOT, Innovate UK, 2018 – 2020, £123,806 (UWE Project Funding). G6 Pipe, T. CAV-Forth, Innovate UK, 2019 – 2021, £625,139 (UWE Project Funding). G7 Pipe, T. Trustworthy Robotic Assistants (RoboSafe), Engineering & Physical Sciences Research Council, 2012 – 2016, £49,730 (UWE Project Funding).
4. Details of the impact
Informing the legal and insurance sectors
UWE researchers collaborated with Burges-Salmon, a large independent UK Law firm, and with AXA UK, a major motor insurer, on three CAV projects ( G1, G2, G3). UWE research contributed knowledge on the developing strengths and weaknesses of the technology itself, as well as many human-factor aspects. The latter included risks pertaining to handover of control between autonomous system and human, and issues relating to the individual and societal acceptability of the technology in use. A partner at Burges-Salmon noted that UWE research provided the ‘ evidence base produced and cited by Burges-Salmon in our engagement with stakeholders shaping the future regulatory framework for CAVs in the UK and beyond’ ( S1). In particular, UWE research also informed Burges-Salmon’s formal responses to the joint Law Commission and Scottish Law Commission preliminary consultation paper on automated vehicles (08.11.2018) ( S2), and follow-up consultation on autonomous passenger services and public transport. These consultations, in turn, informed the work of the Law Commission in developing recommendations on reforming UK law for the introduction of CAVs ( S1).
AXA UK’s Managing Director of Underwriting and Technical Services commented that UWE research on CAVs informed five official ‘ AXA statements on related vehicle insurance issues’ in 2017 and 2018 ( S3, S4) - AXA collaborated with UWE on four separate projects ( G1, G2, G3, G5). The company also acknowledged the contribution of UWE research to the sharing of knowledge through an AXA-funded film on Autonomous Vehicles (Sky, 2018) and an article on CAV technology in a leading newspaper ( Times, 2018), both aimed at increasing the acceptability of this technology for the general public. AXA noted that through these activities ‘ UWE research has assisted in our company’s marketing efforts on the future of driving’ ( S3).
Informing safety standardsThe British Standards Institute (BSI) has developed a number of standards associated with the introduction of CAVs. Standards are a crucial step towards the governance of this emerging technology and ensuring the safe trial, testing and deployment of CAVs on UK roads. The Head of Innovation Policy at BSI has acknowledged the use of UWE research ‘ to inform development of several CAV standards’, including Publicly Available Specification (PAS) 1880 (CAV control systems), PAS1881 (Assuring safety for automated vehicle trials and testing) and PAS1883 (Operational Design Domain taxonomy for an automated driving system) ( S5). BSI also referred to the contribution of UWE research in prioritising new areas for standardisation ( S5). In addition, Burges-Salmon commented on the continued application of UWE research to their involvement in the CAV standards programme at BSI ‘ where standards ranging from safety to trialling to data are being developed that will be used not just in the UK but globally’ ( S1).
Enabling South Gloucestershire Council to promote autonomous vehicles South Gloucestershire Council collaborated with UWE on G1, G2, G3 and G5. Their involvement with many of the experiments conducted by UWE researchers as part of these projects, has enabled South Gloucestershire Council to engage local people with CAV technology through a series of demonstrations at sites in the local area including:
UWE Frenchay Campus in 2017 as part of the VENTURER project ( G1, S6), and on a roadway near to the campus in 2018;
The ex-Filton Airport development site in 2019 and Cribbs Causeway shopping centre in 2020 as part of the CAPRI project ( G3, S6);
Autonomous parcel delivery on a 10-mile route on mixed-mode roadway in South Gloucestershire as part of the ROBOPILOT project ( G5, S6).
In 2020, two CAV simulator ‘kiosks’ were created for public dissemination of the technology in the region as part of the ROBOPILOT project ( G5); one for use by South Gloucestershire Council and one by UWE. The Council’s Strategic Economic Development Manager commented: ‘ UWE research has helped South Gloucestershire Council get communities in this region thinking about the benefits of autonomous vehicles’ ( S6).
Informing the development of CAV technology by industryUWE researchers collaborated with two businesses in particular on the development of CAV applications. Atkins Global is a large design and engineering consultancy with a worldwide presence across a broad range of sectors, including transport, infrastructure and energy. Atkins were overall project lead for the two first CAV projects ( G1 and G2). UWE researchers collaborated closely with Atkins during these projects, and UWE research played a critical role in developing Atkins’ policy on CAV technology ( S7). Atkins’ Technical Director acknowledged the role of UWE research in ‘ shaping advanced applications of CAV technology’, which has become ‘ a major area of growth’ for the company ( S7).
Fusion Processing is a British company that develops advanced sensors and control systems for the transport sector and smart cities. The company collaborated with UWE researchers on the VENTURER ( G1) and CAV Forth ( G6) projects. UWE’s technology integration research as part of the VENTURER project, enabled the development and testing of Fusion Processing’s sensor-perception pipeline and vehicle control technology, which led to their development of their CAVStar vehicle control product ( S8). UWE’s subsequent simulation-based Verification and Validation research is now a critical, integrated part of the process of ensuring the safety of CAVStar in the CAV Forth project. In this project, an autonomous bus will be operated as a commercial service across a 24-mile route in the Edinburgh region in 2021 ( S9).
5. Sources to corroborate the impact
S1 Testimonial from a Partner at Burges-Salmon Ltd
S2 Burges Salmon response to joint Law Commission and Scottish Law Commission
S3 Testimonial from Managing Director, Underwriting and Technical Services, AXA Insurance Ltd
S4 AXA and Burges Salmon Insurance and Legal Report 2018
S5 Testimonial from Head of Innovation Policy, British Standards Institute
S6 Testimonial from Strategic Economic Development Manager, South Gloucestershire Council
S7 Testimonial from Technical Director, Atkins Global Ltd
S8 CAVStar product page on Fusion Processing Ltd website https://www.fusionproc.com/automated-vehicle-systems/cavstar-automotive-sensing-and-control-system/
S9 Testimonial from CEO, Fusion Processing Ltd
- Submitting institution
- University of the West of England, Bristol
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Research into 3D imaging and deep learning by the Centre for Machine Vision (CMV) at the University of the West of England, has led to innovations in the agricultural industry. An automated cattle monitoring system based on CMV research is improving welfare and productivity at dairy farms across the UK and has played an important role in the sustainable farming programme of the UK’s largest milk supplier. A system for the real-time monitoring of root crop production information during harvesting informs marketing, planning and agricultural practices, and has been successfully brought to market by a newly formed company. An automated system for weed detection has been used to guide application of herbicides, saving on costs and environmental impacts. The application of research into machine vision and deep learning has made farming more efficient and sustainable, with benefits to business, consumers, animal welfare and the environment.
2. Underpinning research
From 2007 to 2019 the CMV team worked on the application of 3D imaging for detecting and describing objects of interest in challenging applications. This included novel methods for characterising the morphology of aggregate particles, facial recognition, and detecting hidden objects for security purposes. The technology was later adapted for animal condition monitoring. CMV pioneered new technologies for animal condition monitoring in dairy farming ( R1), for in-field crop data capture, and for automated weeding in pasture ( R2).
Animal condition monitoring Using state-of-the-art data capture, analysis, and machine learning techniques, CMV developed an automated data-driven approach, bettering unreliable, subjective, and labour-intensive methods of manual scoring of animal condition. A novel ‘rolling ball algorithm’ (RBA), which the CMV team previously developed for characterising complex 3D morphologies (e.g. in petrographic analysis of aggregate particles) ( R3), was applied for the first time for quantifying animal body condition (essentially how lean/fat the animal is). RBA’s significant advantage is global operation; obviating the requirement to detect and track local animal features, such as the ‘hook’ and ‘pin’ pelvic bones which are usually considered to be critical features for body condition monitoring. CMV also successfully introduced 3D imaging for tracking and analysing gross body movements, as a proxy for conventional leg gait motion in lameness detection ( R4). Height movements, readily captured from the animal’s back while walking beneath a sensor, form locomotion signals analysed using a Hilbert transform and machine learning, to give better than 96% accuracy in lameness detection. Hence, via a sequence of related projects ( G1, G2, G3), CMV pioneered work on improved farm productivity, farm animal husbandry and welfare ( R1).
In-field root crop data capture Research on isolating 3D structures in plant phenotyping tasks ( R5), formed the basis for developing technology for live crop metrology data-capture on the harvesters ( G4). In this case, the size and shape distribution of fast-moving potato tubers is captured using an overhead RGB-D camera mounted within the vehicle, where segmentation, sizing, and shape characterisation are accomplished in real-time. A Yolov3-tiny object detection network identifies individual potatoes that are then converted to real world measurements using 3D camera parameters. Double counting is avoided by triggering this process only when the entire frame changes through use of an optical flow tracker. Combining with real-time GPS ( R2) enables the generation of timely production statistics, such as crop counting and sizing distributions, together with produce-yield field heat maps, allowing producers to monitor yield, market crops more profitably and, via targeted agronomic performance data, to assess field performance and plan future soil treatments.
Automated weed detection Research on the synergistic use of computer vision and state-of-the-art deep learning techniques, in the form of convolutional neural networks (CNN), allowed pioneering work in the detection, recognition and localisation of plants species in the natural environment, with significantly improved results, setting a new standard in accuracy and applicability ( R6). Using an SSD-mobilenet-v2 CNN, accuracy is better than 96% for a three-class dataset detecting dock and clover leaves in grass - a task that has proved impossible using conventional computer vision and hand-crafted features. Significantly, the use of transfer learning and COCO (Common Objects in Context) weights, means that fewer than 50 training samples are required per class. This is important for agricultural businesses because, due to the high cost of collecting and processing large amounts of training data, they have not yet been able to employ Neural Network models. The employment of a CNN and transfer learning offers a very powerful method for classification of weeds in grassland, making the practical precision-targeted application of herbicide possible from a moving vehicle, equipped with cameras and a controlled spray boom.
CMV research has demonstrated:
The benefits of rich 3D data for revealing, and helping to better describe and characterize, attributes of interest to farmers, such as body condition, lameness and weight in animals ( R1, R4).
The benefits of richer 3D data for isolating and tracking features of interest in plants and crops ( R2, R5).
The power of deep learning in complex and challenging detection and recognition tasks, within highly unstructured agricultural applications ( R6).
3. References to the research
R1 Hansen, M., Smith, M., Smith, L. (2018) Automated monitoring of dairy cow body condition, mobility and weight using a single 3D video capture device. Computers in Industry. Vol.98 pp. 14-22 https://doi.org/10.1016/j.compind.2018.02.011
R2 Smith, L., Zhang, W., Hansen, M., Smith, M. (2018) Innovative 3D and 2D machine vision methods for analysis of plants and crops in the field. Computers in Industry. Vol 97 pp. 122-131 https://doi.org/10.1016/j.compind.2018.02.002
R3 Lee, J., Smith, M., Smith, L. (2007) A new approach to the three-dimensional quantification of angularity using image analysis of the size and form of coarse aggregates. Engineering Geology, 91, pp. 254-264. https://doi.org/10.1016/j.enggeo.2007.02.003
R4 Abdul Jabbar, K., Hansen, M., Smith, M., Smith, L. (2017) Early and non-intrusive lameness detection in dairy cows using 3-dimensional video. Biosystems Engineering. Vol 153 pp. 63-69 https://doi.org/10.1016/j.biosystemseng.2016.09.017
R5 Bernotas, G., Scorza, L., Hansen, M., Hales, I., Halliday, K., Smith, L., Smith, M., McCormick, A. (2019) A photometric stereo-based 3D imaging system using computer vision and deep learning for tracking plant growth. GigaScience, 8/5. pp.1-15 https://doi.org/10.1093/gigascience/giz056
R6 Smith, L., Byrne, A., Hansen, M., Zhang, W., Smith, M. (2019) Weed classification in grasslands using convolutional neural networks. Applications of Machine Learning, SPIE Optics Photonics. Vol.11139. https://doi.org/10.1117/12.2530092
Evidence of the quality of the underpinning researchG1 Smith, L. Precision cow health management, Technology Strategy Board, 2013 – 2016, £210,046. G2 Smith, M. Automated welfare monitoring of dairy cows using 3-dimensional imaging and deep learning, BBSRC, 2019 – 2020, £35,192. G3 Smith, M. Agsenze KTP, Technology Strategy Board, 2019 – 2021, £182,834. G4 Smith, M. Smart Storage' solution for the potato sector to reduce waste by 110kt and improve packer profitability by £108 million per annum, Innovate UK, 2018 – 2019, £99,966. G5 Smith, M. GrassVision: Automated application of herbicides to broad-leaf weeds in grass crops, Innovate UK, 2016-2018, £107,053. G6, Smith, M. GrassVision 2 - A retainable, smart-camera, vision system for agriculture - SKAi, the Soil Essentials KORE Artificial Intelligence platform, UKRI, 2019 – 2022, £137,961.
4. Details of the impact
‘HerdVision’ animal condition monitoring system: increasing productivity and animal welfareThe ‘HerdVision’ system, initially developed under a 3-year BBSRC feasibility study grant ( G1) for which underlying techniques were reported in ( R1), was developed into a commercial product via follow-on BBSRC ( G2) and KTP ( G3) funding. The research has had a significant impact on agri-tech company, Agenze, which has been applying the research to challenges in dairy farming. Since 2014, Agsenze has secured GBP700,000 of investment to develop the technology based on UWE research, and has created eight new jobs to work on this project (6.5FTE) ( S1). The Managing Director at Agsenze commented that:
‘the scoring is produced at least two times a day as opposed to the… [previous] manual system of once every two months – leading to better welfare of the livestock and better productivity for the farmer’ ( S1).
As of 2020, the HerdVision system is being used to monitor 3,300 cows on 11 farms throughout the UK ( S1, S2), 10 of them owned by Arla, a farmer-owned dairy cooperative. The farmers and cows at these farms benefit from automated body scoring and lameness detection. Arla, the largest producer of milk in the UK, supplies major supermarkets such as Morrisons and Aldi. The Arla farms using HerdVision are part of the ‘Arla 360’ programme, which uses innovative technology to make dairy farming more sustainable and to improve animal welfare. HerdVision has been used as part of the Arla 360 promotional campaign, in which Arla UK positions itself as a sustainable and ethical brand ( S3). This has increased the confidence of retailers stocking Arla products and their customers. The Senior Agricultural Manager at Morrisons UK commented ‘ our customers care about animal welfare, so to know that these trials can improve the well-being of the animals supplying their milk is reassuring’ ( S3).
‘HarvestEye’ in-field crop data capture: providing real-time insight and improving profitability‘HarvestEye’ is a unique system that provides valuable insights on root crop performance during harvest across a whole field, rather than conventional, limited, sample data. The system is based on advanced vision techniques for surface analysis developed by CMV and reported in R2 ( S4). The Director of B-Hive Innovations, the company developing HarvestEye, noted ‘ UWE research contributed materially to the decision to set up B-Hive’ and that ‘ the Harvesteye system, based on UWE research, is an important part of our portfolio’ ( S5). In November 2019, B-hive Innovations launched HarvestEye Ltd and announced a partnership with GRIMME UK, to market the system in the UK ( S6). The successful development of this system has meant that Harvest Eye Ltd now has three new employees (3FTE), has sold 42 units to date, and had plans to employ another three employees in 2020/2021 to cater for further growth in sales ( S4).
‘GrassVision’: upskilling precision farming specialists to reduce environmental damage through targeted herbicide delivery Since 2016, CMV has worked in partnership with SoilEssentials Ltd to develop a system for detecting and spraying weeds in pasture. The system is based on findings from an InnovateUK feasibility project ( G5, R6). A follow-on 3-year InnovateUK project, ‘GrassVision2’ ( G6), supported the development of a marketable device, to allow herbicide to be precisely targeted only where needed, increasing productivity, reducing costs and reducing the environmental impact of over-application of herbicides. The resulting Soil Essentials KORE Artificial intelligence platform uses ‘SKAi’, a re-trainable smart camera vision system developed at UWE for agricultural applications. The system is able to reduce the total amount of crop protection products required by 90%, by identifying weeds and delivering herbicide in a targeted way. SoilEssentials is an established provider of precision farming products and services, approaching its 20th anniversary ( S7). The expertise of the SoilEssentials team has been in database applications, data modelling and agricultural hardware integration. UWE research enabled the SoilEssentials team to develop a high level of expertise in two new areas; machine vision and deep learning. The Managing Director of Soil Essentials commented:
‘Engagement with UWE research, as reflected in publications such as the 2019 paper on Weed classification in grasslands using convolutional neural networks [ R6], has been invaluable in upskilling our team and creating the potential for new commercial opportunities. The team at SoilEssentials have now successfully trialled the weed detection and spraying technology… and have developed many of the skills required to be at the forefront of this pioneering technology. This would not have been possible without our engagement with UWE research’ ( S8).
5. Sources to corroborate the impact
S1 Testimonial from the Managing Director of Agsenze
S2 Agsenze slide showing UK farms using Herdvision as of September 2020
S3 Arla UK press release July 2019
S4 Testimonial from the Director of B-Hive Innovations Ltd August 2020
S5 Testimonial from the Director of B-Hive Innovations Ltd March 2019
S6 HarvestEye press release 21.11.2020
S7 SoilEssentials Ltd website
S8 Testimonial from the Managing Director of SoilEssentials Ltd
- Submitting institution
- University of the West of England, Bristol
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Societal
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Over 2 billion people globally lack basic sanitation facilities, and over a billion are also without electricity. Researchers at the University of the West of England (UWE) have developed an inexpensive Microbial Fuel Cell (MFC) technology, Pee Power®, which addresses both these challenges. Pee Power urinals use unprocessed urine to generate sufficient electricity to light cubicles. With funding from the Bill and Melinda Gates Foundation, this technology has been developed to the point of practical application. Units have been installed in schools in Uganda and Kenya, and in an informal settlement in South Africa. They have helped several thousand students and local residents feel safer while using toilet facilities at night. In addition, Pee Power toilets have changed awareness, knowledge and public attitudes through three years of installations at Glastonbury Festivals. The technology has also been evaluated by humanitarian organisation, Oxfam, who see it having major potential for future provision of sanitation in disaster zones.
2. Underpinning research
Microbial fuel cells (MFCs) generate power using wastewater as fuel. Bacteria inside the MFCs break down organic matter, releasing electrons and cleaning the liquid. MFC research at UWE began in 2002, with Professor Ieropoulos developing the world’s first autonomous MFC-powered robots (EcoBots) ( R1). Realising the potential of MFC technology to change lives, the Bristol Bioenergy Centre (BBiC) at UWE focused on developing MFC technology for wider societal implementation. This meant working with higher energy density and lower-cost materials ( R2). Such a consideration was rare amongst other laboratory-based groups internationally, who use expensive and unsustainable materials (such as noble metals) for research. BBiC projects began to focus on optimising MFC components for wastewater treatment, and in 2011 the group reported for the first time that unprocessed urine works extremely well as fuel for MFCs ( R2). This finding was instrumental in Ieropoulos securing both an EPSRC Fellowship ( G1) and a follow-up EPSRC commercialisation grant ( G2), and also obtaining a linked series of grants from the Bill and Melinda Gates Foundation ( G3, G4, G5, G6).
In 2012-13, work by Winfield et al. led to the development of ceramic MFCs ( R3). This has been crucial for implementing the technology in remote and/or poor areas, where clay can be sourced locally at very little cost. Research by Gajda et al. focused on improving the design of ceramic MFCs by developing cylindrical reactors ( R3). These novel MFCs produce three times as much power. The cylindrical stacks also produce a by-product with antimicrobial properties that can be used as a disinfectant ( R3).
With a focus on size and power scale-up, BBiC developed expertise on how best to configure multiple MFCs into three-dimensional stacks. This opened up opportunities for new applications of MFC technology, including its potential in the development of smart ‘living architecture’ ( R4). MFCs were integrated into three different kinds of conventional house bricks (from Europe and Uganda) to test their ability to produce usable power (e.g. to operate motorised windows) whilst cleaning the property’s wastewater ( G7). The research demonstrated that the conversion of existing and future buildings into micro-power stations and micro-treatment plants was achievable using MFC technology.
BBiC have developed a range of designs, configurations and patents for MFCs ( R5) with the power output now orders of magnitude higher than a decade ago. In 2003, original BBiC MFCs produced nine nanowatts of power. BBiC MFCs now produce two million nanowatts, and are a fraction of the size. An important technological breakthrough in stack development came with the production of novel multi-electrode boxes, generating power measured in watts rather than microwatts. BBiC’s systems today produce between two and three watts of power continuously.
3. References to the research
R1 Melhuish, C., Ieropoulos, I., Greenman, J. and Horsfield, I. (2006) Energetically Autonomous Robots: Food for Thought, Autonomous Robots 21:(3), pp 187-198. https://doi.org/10.1007/s10514-006-6574-5
R2 Ieropoulos, I., Greenman, J. and Melhuish, C. (2011) Urine utilization by microbial fuel cells: energy fuel for the future, Physical Chemistry Chemical Physics 14: (1), pp 94-98. http://dx.doi.org/10.1039/C1CP23213D
R3 Gajda, I., Greenman, J., Melhuish, C. and Ieropoulos, I. (2015) Simultaneous electricity generation and microbially-assisted electrosynthesis in ceramic MFCs, Bioelectrochemistry 104, pp 58-64. https://doi.org/10.1016/j.bioelechem.2015.03.001
R4 You, J., Rimbu, G., Wallis, L., Greenman, J. and Ieropoulos, I. (2019) Living Architecture: Toward Energy Generating Buildings Powered by Microbial Fuel Cells, Frontiers in Energy Research 7. https://doi.org/10.3389/fenrg.2019.00094
R5 Ieropoulos, I., Greenman, J. (2015) Patent: Microbial fuel cell, method of controlling and measuring the redox potential difference of the fuel cell, WO2016120641A1, WIPO (PCT).
R6 Ieropoulos, I., Stinchcombe, A., Gajda, I., Forbes, S., Merino-Jimenez, I., Pasternak, G., Sanchez-Herranz, D., Greenman, J. (2016) Pee Power urinal – microbial fuel cell technology field trials in the context of sanitation, Environmental Science: Water research and technology, 2, pp 336-343. https://doi.org/10.1039/C5EW00270B
Evidence of the quality of the underpinning research
G1 Ieropoulos, I. Waste made useful by using Microbial Fuel Cells for energy generation, EPSRC, 2010 – 2014, £539,721.
G2 Ieropoulos, I. New Directions: MFC Commercialisation through continued Research, Networking and Collaboration, EPSRC, 2013 – 2015, £247,108.
G3 Ieropoulos, I. Grand Challenges Explorations, Urine-tricity: Electricity from urine, Bill and Melinda Gates Foundation, 2011 – 2013, £55,555.
G4 Ieropoulos, I. Grand Challenges Explorations, Phase-II Urine-tricity ++: Electricity from urine, Bill and Melinda Gates Foundation, 2014 – 2015, £499,701.
G5 Ieropoulos, I. Urine-tricity III: Electricity from urine, Bill and Melinda Gates Foundation, 2014 – 2015, £1,052,337.
G6 Ieropoulos, I. Phase 4 Urine-tricity: Development of Microbial Fuel Cell Platform Technology and In-Field Testing and Support, Bill and Melinda Gates Foundation, 2019 – 2021, £1,073,748.
G7 Adamatzky, A. Living architecture, European Commission, 2016 – 2019, £699,828.
4. Details of the impact
BBiC successfully translated their findings into a pee-powered urinal installation (named Pee Power), which used the urine to fuel the MFCs that generated power for the cubicle’s lights. The first Pee Power urinal was installed at UWE’s Frenchay campus in March 2015 and then scaled-up for testing in the field at the Glastonbury music festivals (2015-2019) ( R6). The field trial demonstrated, for the first time, that it was feasible to use MFCs to generate power as well as treat urine.
Pee Power urinals are already having practical impact. Installations in Uganda, Kenya and South Africa have improved safety and sanitation for over 3,600 students and community members. Field trials at Glastonbury Festival raised significant public awareness of MFC technology and its potential for providing sustainable and safe sanitation solutions. These urinals have also offered a huge step forward for humanitarian organization, Oxfam, in their future provision of sanitation in disaster zones.
Improving safety and sanitation: Pee Power in the Global SouthIn July 2017, the first overseas Pee Power installation was commissioned at the Seseme girls’ secondary boarding school in Kisoro, Uganda, funded by the Bill and Melinda Gates Foundation. The school serves 600 pupils, who have access to four toilets. Prior to the installation of the MFCs, these toilets had never been lit, with serious safety implications. In a video interview, the Headteacher commented that ‘ the girls used to fear to go to the latrines at night’ because of intruders getting into the school ( S1).
A brick structure was constructed adjacent to the toilets to house the MFC modules. Urine from the first two cubicles in the toilet block flows down into the MFCs, powering lights in the toilets and providing a safer environment for the girls at night. Following the Pee Power installation, the Headteacher noted that ‘ since we got this power…[the girls] can freely go to the latrines at night, and they are not scared at all, and those strangers no longer come into school’ ( S1). The International Water Security Network’s survey of the girls at the school who used the Pee Power system, highlighted that 86% of respondents found the toilets safer to use at night ( S2).
In June 2018, a second Pee Power system was installed at the Brainhouse Academy, a mixed-sex primary and secondary academy in Mathare, Nairobi, Kenya. The Academy has around 500 pupils. New toilets at the school would have been unlit without the integration of Pee Power. One Brainhouse teacher commented that Pee Power had ‘ made the school secure’ and that ‘… pupils feel much safer when they visit the toilet during night hours’ ( S3). Another teacher at the school echoed this point, and also commented on the reduction of rodents in the toilets due to the lights ( S4). The technology has also enhanced his teaching, as he refers to it when teaching related topics ( S4).
A third Pee Power system was installed at an informal settlement community of about 2,500 people in Thandanani, Durban, South Africa. This was part of the Bill and Melinda Gates Foundation engineering field-testing platform, used to evaluate technologies sufficiently mature for potential commercialisation. As well as demonstrating feasibility in practice and at scale, at Thandanani, Pee Power has again enabled people to use the toilets at night without the need for torches. A community member at Thandanani commented ‘ we feel comfortable having lights – you don’t have to carry your phone any more’ ( S5). A Prototype Engineer in the Pollution Research Group at the local University of KwaZulu-Natal praised the Pee Power installation at Thandanani for its ‘ eco-friendly electricity provision and waste treatment’, noting that the system ‘ has positively surprised the community and visitors that we can use waste as an energy source’ ( S6).
Field-testing and support with successive phases for funding from the Bill and Melinda Gates Foundation has supported development to the point where large-scale deployment is now practical in a humanitarian, not-for-profit-context. Early-stage industrial collaboration has also been undertaken (subject to confidentiality) to support commercial development of the technology in selected markets, in line with the terms of agreements with the Gates Foundation.
Pee Power shaping the future of sanitation in refugee campsHumanitarian organization, Oxfam, became interested in the Pee Power urinals in March 2015. The prototype urinals offered an entirely new solution for Oxfam’s delivery of sanitation in disaster zones. The Head of Water and Sanitation at Oxfam, identified Pee Power urinals as a potential ‘ game changer’ for refugee and displacement camps without lighting for toilet areas, where often ‘ women going to the toilet at night are facing abuse and being molested’ ( S7). The Pee Power urinals have several advantages over, for example, solar panels, the performance of which deteriorates with time and which are often stolen. Oxfam are now aware of a viable solution which does not have these limitations ( S7).
Pee Power enhances environmental profile of Glastonbury Festival and improves public understanding and confidence in sustainable toilets Field trials at Glastonbury Festival have done much to spread knowledge and understanding of the technology and its potential with the broad-spectrum, developed-world festival audience. This resulted in increased media coverage and resulted in successively larger Pee Power installations, eventually accommodating up to c.5000 users per day. Pee Power urinals operated at all the Glastonbury Festivals from 2015 to 2019, with a five-year memorandum of understanding being signed by Glastonbury and UWE. Glastonbury’s sanitation manager commented:
‘Since 2015, the Pee Power installations at the festival site have helped Glastonbury Festival to position itself as an environmentally and socially responsible event. The installations help raise awareness of environmental issues among festival-goers, in addition to awareness of the situation of the 2 billion people without access to basic sanitation and 1 billion without access to electricity’ ( S8).
The 2019 Glastonbury installation improved confidence in sustainable toilets for 92% of the 4,745 users who responded using the ‘FeedbackNow’ pads located inside the urinal - quite an achievement for a music festival ( S9)! A more detailed survey conducted by the Pee Power team, showed improved awareness among urinal users of the power of innovative technology to make all toilets sustainable, and to address the lack of electricity and sanitation in the developing world (97% of respondents, n=77). 84% of respondents also reported increased awareness of Glastonbury Festival’s commitment to sustainability after using Pee Power ( S10, pp 3-5).
5. Sources to corroborate the impact
S1 Video interview (and transcript) with Headmistress at Sesame Girls School, Kisoro, Uganda
S2 International Water Security Network Kisoro survey report
S3 Testimonial from Teacher A at Brainhouse Academy
S4 Testimonial from Teacher B at Brainhouse Academy
S5 Testimonial from member of the community at Thandanani, Durban
S6 Testimonial from member of the Pollution Research Group, University of KwaZulu-Natal
S7 Video interview (and transcript) with the Head of Water and Sanitation at Oxfam
S8 Testimonial from the Sanitation Manager, Glastonbury Festivals
S9 FeedbackNow Glastonbury survey report
S10 Pee Power at Glastonbury survey report
- Submitting institution
- University of the West of England, Bristol
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Societal
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Professor Alan Winfield’s research and engagement activities have contributed materially to robot and Artificial Intelligence (AI) ethics, building on academic discourse to inform and impact on intense public and policy debate, both nationally and internationally. The research, conducted at the University of the West of England, has:
informed the development of new national and international standards for the ethical design and application of robots and robotic systems;
influenced the development of new organisational standards within the robotics industry, and defined best practices;
enhanced wider public understanding and informed public debate on robot ethics;
guided the work of the UK government, and UK and EU parliaments, and informed policy debate in the House of Commons;
helped shape the NHS strategy for preparing the healthcare workforce to deliver digital healthcare technologies in the future.
2. Underpinning research
Considering the ethical impacts of robots and work on standardsUWE research into robot ethics and the related field of ethical robots, emerged from Professor Winfield’s 2006-2009 EPSRC grant Walking with Robots ( G1) and 2009-2013 EPSRC Senior Media Fellowship, Intelligent Robots in Science and Society ( G2). Walking with Robots, which won the Royal Academy of Engineering Rooke Medal in 2010, considered the ethical implications of robotics research, and Intelligent Robots in Science and Society looked at the ethical impact of robotics on society. Winfield was asked to present findings from these projects to the EPSRC Societal Impact Panel, which led to him being asked to co-organise a joint EPSRC/AHRC workshop, which in turn resulted in the publication of the EPSRC/AHRC Principles of Robotics ( R1).
These principles ( R1) directly influenced subsequent ethical principles in robotics and AI. They also prompted the formation of a working group on robot ethics, which led directly to the development of British Standard BS 8611 – the world’s first published ethical standard in robotics (see Section 4). Winfield’s engagement with the EPSRC principles and the British Standards Institute (BSI) in turn led to his invitation to join the IEEE Standards Association’s global ethics initiative. Winfield’s contributions to the work of this initiative have been the development of new general ethical principles for autonomous and intelligent systems, and new IEEE ethical standards.
Introducing the ‘ethical black box’ approach and pillars of ethical governance
An ongoing collaboration with Professor Marina Jirotka (University of Oxford) on ethical governance, resulted in a proposal in 2017 that all robots and AI should be equipped with the equivalent of an aircraft flight-data recorder to support robot accident investigation ( R2). Winfield and Jirotka are currently being supported by a five-year EPSRC grant ( G3) to develop this ethical black box. The work on R2 led to a paper ( R3) that developed the framework linking ethical principles to standards and regulations. The paper argued that ethical governance was essential to building public trust in robotics and AI, and proposed four ‘pillars’ of good ethical governance for companies and organisations:
Publish an ethical code of conduct;
Provide ethics and responsible research and innovation training for all members/staff;
Practice responsible innovation, including the engagement of wider stakeholders within a framework of anticipatory governance;
Be transparent about ethical governance.
Using simulation-based architectures for ethical robots and introducing verifiability
Winfield’s work with Professor Michael Fisher (University of Liverpool), brought work on formal methods from computer science together for the first time with robotics, to develop completely new approaches to the validation of robot systems and robot safety. Also, in parallel, Winfield’s EPSRC project The Emergence of Artificial Culture in Robot Societies ( G4), led to the idea of robots with simulation-based internal models – that is, a simulation of the robot itself, other robots, and its environment inside itself. Building on this work, the EPSRC project Verifiable Autonomy ( G5) focused on explicitly ethical robots, i.e. robots that can take ethical considerations into account when deciding how to behave. The inclusion of models of other agents was a novel feature of this research, enabling the robot to predict the consequences of both its and another agent’s actions. UWE researchers conducted a series of successful experimental trials and demonstrated the world’s first transparent and verifiable ethical robot. Outputs from this work include R4, R5 and R6. R6 appeared in a special issue on machine ethics of the Proceedings of the IEEE, co-edited by Winfield. This special issue represents the most comprehensive survey of the emerging field of practical machine ethics to date, with papers on both the engineering and governance of ethical machines.
3. References to the research
R1 Boden, M. Bryson, J., Caldwell, D., Dautenhahn, K., Edwards, L., Kember, S., Newman, P., Parry, V., Pegman, G., Rodden, T., Sorrell, T., Wallis, M., Whitby, B. and Winfield, A. (2017) Principles of robotics: Regulating robots in the real world. Connection Science, 29 (20), pp. 124–129. https://doi.org/10.1080/09540091.2016.1271400
R2 Winfield, A. and Jirotka, M. (2017) The case for an ethical black box. In: Gao, Y., ed. (2017) Towards Autonomous Robot Systems. Springer, pp. 1-12. https://uwe-repository.worktribe.com/output/904084
R3 Winfield, A. and Jirotka, M. (2018) Ethical governance is essential to building trust in robotics and AI systems. Philosophical Transactions A: Mathematical, Physical and Engineering Sciences, 376 (2133). ISSN 1364-503X. https://doi.org/10.1098/rsta.2018.0085
R4 Winfield, A., Blum, C. and Liu, W. (2014) Towards an ethical robot: Internal models, consequences and ethical action selection. In: Mistry, M., Leonardis, Aleš, Witkowski, M. and Melhuish, C., eds. Advances in Autonomous Robotics Systems: Proceedings of the 15th Annual Conference, TAROS 2014, Birmingham, UK, 1-3 September 2014, pp. 85-96. http://dx.doi.org/10.1007/978-3-319-10401-0_8
R5 Vanderelst, D. and Winfield, A. (2018) An architecture for ethical robots inspired by the simulation theory of cognition . Cognitive Systems Research, 48. pp. 56-66. ISSN 1389-0417. https://doi.org/10.1016/j.cogsys.2017.04.002 R6 Bremner, P., Dennis, L., Fisher, M. and Winfield, A. (2019) On proactive, transparent and verifiable ethical reasoning for robots. Proceedings of the IEEE, 107 (3). pp. 541-561. ISSN 0018-9219. https://doi.org/10.1109/JPROC.2019.2898267
Evidence of the quality of the underpinning researchG1 Winfield, A. Walking with Robots, EPSRC, 2006 – 2009, £249,557. G2 Winfield, A. Intelligent Robots in Science and Society, EPSRC, 2009 – 2012, £112,078. G3 Winfield, A. RoboTIPS: Developing Responsible Robotics for the Digital Economy, EPSRC, 2019 – 2024, £428,068. G4 Winfield, A. The Emergence of Artificial Culture in Robot Societies, EPSRC, 2007 – 2011, £735,507. G5 Winfield, A. Verifiable Autonomy, EPSRC, 2014 – 2018, £340,338.
4. Details of the impact
The development of new national and international ethical standards
UWE research has informed the development of national and international standards for the ethical design and application of robots and robotic systems:
The principles articulated in R1 prompted the formation of the BSI working group that drafted BS 8611, the world’s first published standard (2016) in robot ethics ( S1).
As founding co-chair of the IEEE Standards Association ethics initiative General Principles Committee, Winfield brought the insights of his research to bear on the development of new general ethical principles for Intelligent and Autonomous Systems, a foundational part of IEEE framework ‘ Ethically Aligned Design’ 2017 ( S2). For example, S2 (p30) makes use of R3.
Drawing on R2 and R3, Winfield led a proposal that one of the IEEE framework’s general ethical principles related to transparency should form the basis of a new IEEE standard (IEEE P7001). In 2017, that proposal was accepted, and he now chairs Working Group P7001 drafting a new standard on Transparency in Autonomous Systems (for an outline of P7001 see S3).
Ideas and practice within robotics industry The robotics industry and affiliated unions are using Winfield’s research to inform their thinking and practice on robot ethics. His research has underpinned the development of organisational standards and helped define best practice:
The UNI Global Union – an association of over 650 unions in 140 countries with over 20 million members – has developed a set of principles for ethical AI. Principle 2 ( S4, p7) is Equip AI Systems With an “Ethical Black Box” – an idea first proposed in R2. UNI Global Union cites R1 and an article in Futurism, which draws on R2, as providing inspiration and insight for the development of the principles ( S4, p10).
Ethical standards are becoming incorporated into industry practice; Sastra Robotics, for example, reference the IEEE initiative (which draws on R3) in their statement on Rules and Ethical Considerations of Robotics Technology ( S5).
Informing public debate on robot ethics
UWE research has extensively informed public debate on robot ethics through Winfield’s wide-ranging engagement with the public – drawing on the training and development received as a science communicator through his EPSRC Senior Media Fellowship award. From August 2013, Winfield contributed to over 50 public lectures and panel debates on ethical challenges in robotics and AI. For example, he gave the Campaign for Science and Engineering (CaSE) 2018 annual lecture alongside Dame Wendy Hall and Jim Al-Khalili at the Institute of Physics, and he debated AI with Professor Brian Cox and Robin Ince on BBC’s popular radio programme The Infinite Monkey Cage in January 2016. Winfield is frequently called upon by the press and media to comment on topical issues in robot ethics; notably he was a guest on BBC R4 programme The Life Scientific in February 2017, and interviewed for BBC News HARDtalk in October 2017 ( S6). Winfield’s blog, which deals mainly with issues relating to robot ethics and ethical robots, has been visited over 500,000 times since August 2013.
Informing the work of the UK government, and UK and EU parliamentsWinfield’s research and profile within robot ethics has also led to prolific engagement with UK government departments and parliament, and has informed policy debate in multiple Parliamentary Committees and similar bodies:
An invitation by the Foreign Office to brief the G8 non-proliferation committee on the risks of robotics and AI in October 2013 (co-presenting with Lord Martin Rees) ( S7, p1).
Invitations by the UK Chief Scientific Advisor, Sir Mark Walport, to attend expert round table briefings in January 2015 and January 2016; the latter on the ‘ risks and opportunities in the use of artificial intelligence in government decision-making’ ( S7, p2, p3).
Winfield was invited to submit written evidence to the House of Commons Select Committee on Science and Technology inquiry on Robotics and AI, and is cited in connection with the need to be able to investigate the logic by which AI decisions are made and the implications of this for public confidence ( S8, Robots and artificial intelligence, p18, p22, cf. R2).
Winfield was invited to give oral evidence to the 2017 House of Lords Select Committee on Artificial Intelligence’s session on AI ethics. This oral evidence is cited in the Committee’s 2018 report in connection with technical transparency ( S8, AI in the UK, p38).
Both the work of the House of Commons Select Committee on Science and Technology on Robotics and AI, and the concept of a ‘logging mechanism’ to give a step-by-step account of processes involved in decision making, (first proposed in R2) informed a debate held in the House of Commons on 17 January 2018 where the House considered ethics and AI ( S9, Column 353WH).
Winfield has attended several meetings of the All-Party Parliamentary Groups (APPGs) on AI and on Data Analytics. Winfield and Jirotka’s work on transparency ( R2) and ethical governance ( R3) was cited in the 2019 APPG Data Analytics report on Trust, Transparency and Technology ( S10, p23, p46).
In 2020, the European Parliament Panel for the Future of Science and Technology published a report on The ethics of artificial intelligence, which cites R3 ( S11, p32, p34).
Informing NHS strategy
Finally, UWE research has helped shape the NHS strategy for preparing the healthcare workforce to deliver digital healthcare technologies in the future. Winfield was invited by Health Education England (HEE) to join the Topol Review as the Robotics and AI ethics advisor, contributing to the final report published in February 2019 ( S12); Winfield co-drafted two sections of that report: Ethical Considerations and AI & Robotics. The NHS’ Interim people plan (June 2019) identified implementing the recommendations of the Topol report as a major objective for the NHS ( S13, p54, p71).
5. Sources to corroborate the impact
S1 Robots and Robotic Devices - Guide to the Ethical Design and Application of Robots and Robotic Systems BS 8611:2016 (British Standards Institution, 2016)
S2 IEEE Standards Association (2017) IEEE Global Initiative on the Ethics of Autonomous and Intelligent Systems - Ethically Aligned Design
S3 Winfield, A. F. (2019) Ethical standards in Robotics and AI. Nature Electronics, 2. pp. 46-48. ISSN 2520-1131
S4 UNI Global Union Top 10 Principles for Ethical Artificial Intelligence
S5 Sastra Robotics (2017) Rules and ethical considerations of robotics technology
S6 BBC R4, The Infinite Monkey Cage: Artificial Intelligence, first broadcast 19.01.2016; https://www.bbc.co.uk/programmes/b06wcsng; BBC R4, The Life Scientific: Alan Winfield on Robot Ethics, first broadcast 27.02.2017 https://www.bbc.co.uk/programmes/b08ffv2l; BBC World News, HARDtalk, interview by Stephen Sackur, first broadcast 31.10.2017 https://www.bbc.co.uk/programmes/n3ct2km5
S7 Invitations from the Foreign Office and the UK Chief Scientific Advisor
S8 House of Commons (2016) – Robotics and artificial intelligence, Select Committee on Science and Technology, Fifth Report of Session 2016-2017; House of Lords (2018) AI in the UK: ready, willing and able? Select Committee on Artificial Intelligence, Report of Session 2017–19
S9 Hansard Ethics and Artificial Intelligence Volume 634: debated on 17.01.2018
S10 All Party Parliamentary Group on Data Analytics (2019) Trust, Transparency and Technology: Building Data Policies for the Public Good
S11 European Parliament Panel for the Future of Science and Technology (STOA) The ethics of artificial intelligence: Issues and initiatives
S12 The Topol Review Preparing the healthcare workforce to deliver the digital future: an independent report on behalf of the Secretary of State for Health and Social Care, NHS Health Education England, Feb 2019
S13 NHS Interim People Plan June 2019