In the current Circular Economy (CE) landscape, many activities fail to fully realise the potential value of products, components, and materials. Rather than being repaired, reused, or remanufactured, a significant number of products end up in mixed waste streams and are recycled, which is the least valuable end-of-life CE option. This leads to a substantial loss of residual value. One of the critical challenges in the CE is the absence of effective and efficient methods for large-scale separation of products, which are in very different used or end-of-life conditions, into various CE options such as reuse, repair, remanufacture, repurpose, or recycle. The presence of mixed waste streams poses a significant barrier to creating tighter loops of circularity and preserving materials at their highest value for an extended period. The current methods of evaluating end-of-life options are inefficient and inaccurate as they heavily rely on human judgment, which is often based on experience and prone to bias and error. Similar to how triage can help create priorities and organisations in healthcare systems, RoboTriage proposes a new concept, circularity triage, referring to the process of rapidly examining products, components and parts to determine their best CE option, e.g. reuse, repair, remanufacture, repurpose, or recycle. We aim to create and develop robotic systems that can perform circularity triage by capturing the health condition data of used products, allowing for a swift evaluation of product conditions so as to group products of similar conditions, avoid mixed waste streams and recommend the highest-value CE options. RoboTriage has five objectives: O1: To create robotic systems that can perform triage operations (at the operation level). O2: To create system intelligence that enables smart planning of triage operations (task level). O3: To identify new patterns and connections between product history data and product conditions using autonomous large-scale robotic triage (task level). O4: To identify value opportunities and develop circular business models with the new patterns and connections obtained to facilitate high-value retention (system level). O5: To support the uptake of CE and sustainability considerations and practices by industrial partners through three flagship case studies. RoboTriage's academic impact transcends the manufacturing and CE domains, extending to ICT, AI, and data science. The influence of RoboTriage extends into economic, societal, and environmental domains. RoboTriage technologies have the potential to be deployed for CE purposes, thereby enhancing their scale and productivity. On average, a one per cent increase in robot density correlates with a 0.8 per cent increase in productivity. The impact of ideas in RoboTriage will be exemplified by our 11 industrial partners and over £700k contributions (including over £400k cash contributions) from the three host institutions and external partners. Facilitated by RoboTriage technologies, the promotion of high-value CE options such as remanufacturing could lead to a 90% reduction in primary material usage and a 55% reduction in energy and emission impact. The impact of this project will extend to international organisations through our United Nations partners, ITU and UNESCO, both of which are also our project partners.

The Circular Economy requirements and sustainability goals have been set out by the UK government and the United Nations to address the climate crisis and maintain our standard of living. The environmental impact from the global consumption of engineering materials is expected to double in the next forty years (OECD: Global Material Recourses to 2060, 2018), while annual waste generation is projected to increase by 70% by 2050 (World Bank What a Waste 2.0 report, 2018). A radical departure from traditional forward manufacturing is needed that no longer exclusively focuses on the original manufacturing process and the end of life dispose of manufactured products, parts, and materials. Processes are needed that will significantly prolong the useful life of engineering and especially critical materials (minerals with high economic vulnerability and high global supply risk e.g. rare earth elements for batteries, magnets and medical devices) by increasing the effectiveness of reuse, repurpose, repair, remanufacture, and recycle (Re-X) manufacturing processes. These Re-X processes are currently 3-6 times more labour intensive than traditional manufacturing processes. They are often not economic resulting in many engineering materials being disposed on landfill sites, degraded, or incinerated. UK businesses could benefit by up to £23 billion per year through low cost or no cost improvements in the efficient use of resources. The vision of this hub is to pursue an integrated, holistic approach toward creating a new manufacturing ecosystem for circular resource use of high value products through advances in AI and intelligent automation, empowering the UK to be a world leader in circular manufacturing. To deliver this ambition the hub will focused on two grand challenges: GC1: Radically transform the sustainable use of critical materials. (Goal: >75% Critical components reuse; >20% critical material use decrease; >50% component reclaim increase). GC2: Radically improve the productivity of Re-X manufacturing processes on par with or exceeding traditional forward manufacturing processes (Goal: >10 times improvement). To address these, the hub will establish a truly interdisciplinary team cutting across Manufacturing, Robotics, AI and Automation, Materials Science, Chemical Engineering, Chemistry, Economics, and Life Cycle Assessment.?The hub will focus on three major fronts: Research excellence, community building and user engagement. The new research required to address the grand challenges and overcome the barriers and limitations preventing the transition to a truly circular manufacturing ecosystem will investigate: - New smart processes for disassembly, remanufacturing, separation, and recovery of critical products, components, and ultimately materials. - New sensing and analysis processes to track and determine the state of critical materials throughout their life. - New design methodologies for circular manufacturing. - New testing and validation methods to certify the remaining useful life of crucial products, components, and materials. - New circular Re-X business models. Our research programme will enable rapid scale up of Robotics and AI solutions that are compatible with sector practice, extensible via modular design, and can be repurposed initially in four flagship sector scenarios: energy, medical devices, electric drives, and large structures. Consequently, this Hub will directly address the 80% of the environmental impact of high-value products (Circular Economy Action Plan, European Union, 2020), and save more than 8M tonnes of CO2 emissions annually (HM Government Building our Industrial Strategy report, 2017).

We will train a cohort of students at the interface between the physical and computer sciences to drive the critically needed implementation of digital and automated methods in chemistry and materials. Through such training, each student will develop a common language across the areas of automation, AI, synthesis, characterization and modelling, preparing them to become both leader and team player in this evolving and multifaceted research landscape. The lack of skilled individuals is one of the main obstacles to unlocking the potential of digital materials research. This is demonstrated by the enthusiastic response toward this proposal from our industrial partners, who span sectors and sizes: already 35 are involved and we have already received cash support corresponding to over 27 full studentships. This proposal will deliver the EPRSC strategic priority "Physical and Mathematical Sciences Powerhouse" by training in "discovery research in areas of potential high reward, connecting with industry and other partners to accelerate translation in areas such as catalysis, digital chemistry and materials discovery." The CDT training programme is based on a unique physical and intellectual infrastructure at the University of Liverpool. The Materials Innovation Factory (MIF) was established to deliver the vision of digital materials research in partnership with industry: it now co-locates over 100 industrial scientists from more than 15 companies with over 200 academic researchers. Since 2017, academics and industrial researchers from physical sciences, engineering and computer sciences have co-developed the intellectual environment, infrastructure and expertise to train scientists across these areas. To date, more than 40 PhD projects have been co-designed with and sponsored by our core industrial partners in the areas of organic, inorganic, hybrid, composite and formulated materials. Through this process, we have developed bespoke training in data science, AI, robotics, leadership, and computational methods. Now, this activity must be grown scalably and sustainably to match the rapidly increasing demand from our core partners and beyond. This CDT proposal, developed from our previous experience, allows us to significantly extend into new sectors and to a much larger number of partners, including late adopters of digital technologies. In particular, we can now reach SMEs, which currently have limited options to explore digitalization pathways without substantial initial investment. A distinctive and exciting training environment will be built exploiting the diverse background of the students. Peer learning and group activities within a cross-disciplinary team will accelerate the development of a common language. The ability to use a combination of skills from different individuals with distinct domain expertise to solve complex problems will build the teams capable of driving the necessary change in industry and academia. The professional training will reflect the diversity of career opportunities available to this cohort in industry, academia and non-commercial research organizations. Each component will be bespoke for scientists in the domain of materials research (Entrepreneurship, Chemical Supply Chain, Science Policy, Regulatory Framework). External partners of training will bring different and novel perspectives (corporate, SMEs, start-ups, international academics but also charities, local authorities, consultancy firms). Cohort activities span the entire duration of the training, without formal division between "training" and "research" periods, exploiting the physical infrastructure of MIF and its open access area to foster a strong and vital sense of community. We will embed EDI principles in all aspects of the CDT (e.g. recruitment, student well-being, composition of management, supervisory and advisory teams) to make it a pervasive component of the student experience and professional training.

Despite being far from having reached 'artificial general intelligence' - the broad and deep capability for a machine to comprehend our surroundings - progress has been made in the last few years towards a more specialised AI: the ability to effectively address well-defined, specific goals in a given environment, which is the kind of task-oriented intelligence that is part of many human jobs. Much of this progress has been enabled by deep reinforcement learning (DRL), one of the most promising and fast-growing areas within machine learning. In DRL, an autonomous decision maker - the "agent" - learns how to make optimal decisions that will eventually lead to reaching a final goal. DRL holds the promise of enabling autonomous systems to learn large repertoires of collaborative and adaptive behavioural skills without human intervention, with application in a range of settings from simple games to industrial process automation to modelling human learning and cognition. Many real-world applications are characterised by the interplay of multiple decision-makers that operate in the same shared-resources environment and need to accomplish goals cooperatively. For instance, some of the most advanced industrial multi-agent systems in the world today are assembly lines and warehouse management systems. Whether the agents are robots, autonomous vehicles or clinical decision-makers, there is a strong desire for and increasing commercial interest in these systems: they are attractive because they can operate on their own in the world, alongside humans, under realistic constraints (e.g. guided by only partial information and with limited communication bandwidth). This research programme will extend the DRL methodology to systems comprising of many interacting agents that must cooperatively achieve a common goal: multi-agent DRL, or MADRL.

The UK has fallen significantly behind other countries when it comes to adopting robotics/automation within factories. Collaborative automation, that works directly with people, offers fantastic opportunities for strengthening UK manufacturing and rebuilding the UK economy. It will enable companies to increase productivity, to be more responsive and resilient when facing external pressures (like the Covid-19 pandemic) to protect jobs and to grow. To enable confident investment in automation, we need to overcome current fundamental barriers. Automation needs to be easier to set up and use, more capable to deal with complex tasks, more flexible in what it can do, and developed to safely and intuitively collaborate in a way that is welcomed by existing workers and wider society. To overcome these barriers, the ISCF Research Centre in Smart, Collaborative Robotics (CESCIR) has worked with industry to identify four priority areas for research: Collaboration, Autonomy, Simplicity, Acceptance. The initial programme will tackle current fundamental challenges in each of these areas and develop testbeds for demonstration of results. Over the course of the programme, CESCIR will also conduct responsive research, rapidly testing new ideas to solve real world manufacturing automation challenges. CESCIR will create a network of academia and industry, connecting stakeholders, identifying challenges/opportunities, reviewing progress and sharing results. Open access models and data will enable wider academia to further explore the latest scientific advances. Within the manufacturing industry, large enterprises will benefit as automation can be brought into traditionally manual production processes. Similarly, better accessibility and agility will allow more Small and Medium sized Enterprises (SMEs) to benefit from automation, improving their competitiveness within the global market.