Lorry transportation accounts for 19% of vehicle carbon emissions in the UK but 30% of those lorries are empty, with most of the remainder not running at capacity. Reducing haulage of empty space or "air" would have a direct and substantial impact on carbon emissions as well as bringing down costs. This proposal optimises the load factor (percentage of capacity used) for hub-and-spoke logistics and, specifically, for the partner organisation, Pallet-Track. Hub-and-spoke networks enable hauliers or "depots" to deliver goods from customers in their own areas to recipients who may be located at the other end of the country or even abroad. Depot A takes pallets to the nearest warehouse or "hub", where they are picked up by depots coming from the target delivery area. At the same time, Depot A will pick up goods destined for its own area that have been dropped off by depots from elsewhere. This transport model reduces distances depots travel and extends their delivery range. However, it creates a different problem: they do not know how much freight they will need to collect when they reach the hub. If the delivery and collection numbers do not match, they will transport empty space or they will not have enough lorries, which is compounded by incorrectly estimated pallet numbers at the hub. The proposal uses machine learning to predict the number of pallets depots need to process and the network hub's role in managing the flows between members. Artificial intelligence will interpret these numbers within a cognitive model of decision making that helps depots collaborate with each other to share resources. The numbers of pallets delivered and collected by depots will be analysed to produce accurate predictions for each depot, every day of the year. The cognitive model will interpret these predictions and translate them into transport decisions that optimise the load factors. It will work out which combination of lorries of different sizes are best able to deliver and collect their own pallets. At the same time, it will flag up to neighbouring depots any spare capacity that could be used by them or requests to use spare capacity the neighbours may have. This innovative technology will optimise lorry loading, reduce carbon emissions, and cut costs, making Pallet Track a more attractive hub-and-spoke network. It will also save space at the hubs, which will increase capacity and the throughput of pallets, reducing delivery times.

One-in-six adults will have suffered mental health problems in the past week. Mental health carries an economic and social cost of £105bn/annum, with suicide the leading cause of death for those between 10 and 34 years. There are many vulnerable people needing mental health assessments, but they require trained people to carry them out and the data collected is not easily analysed or shared. Even practitioners find it difficult to create meaningful management plans from their assessments and patients do not feel that they have much of a say. Aston University has been researching these issues for more than a decade. It has developed a web-based questionnaire, GRiST, which models the way practitioners think and reason about mental health and associated risks of suicide, self-harm, harm to others, vulnerability, and self-neglect. GRiST breaks down each risk concept into simple questions that help assessors work out a person's risk level from 0 (minimal risk) to 10 (maximum risk). GRiST has been adopted by several NHS Trusts and other mental-health services over the years and has built up a unique data set of 1.5 million risk assessments. Each one selects a person's relevant subset of answers from 300 precisely-quantified questions and links them to a specific risk level. EGRIST Ltd was spun out of Aston to capitalise on this technology. Our proposal is to develop an automated, digital, structured decision support system known as eGRiST, that can predict levels of mental-health risks and provide advice on how to manage them safely. Innovation arises from the combination of its intuitive psychological model for representing mental health expertise and the development of appropriate artificial intelligence and machine learning algorithms. These will improve the accuracy of risk evaluations of assessors by benchmarking them against their colleagues and help them generate more effective plans for reducing risks, both in the immediate and longer term. The expertise will also be shared with patients through self-assessment technology, so that patients can understand and manage their own mental health in collaboration with their clinical team. The resulting "canopy of care" will deliver reliable, safe, and effective mental-health support for anyone, wherever they are, at the click of a button. It will streamline mental-health services by more accurate referrals, improve communication between patients and practitioners, and give people more control over their own mental health.