The aim of this fellowship programme is to design a socially responsible collective governance for Smart City commons: shared pool of urban resources (transport, parking space, energy) managed and regulated digitally. Smart City commons exhibit unprecedented complexity and uncertainties: transport systems integrate electric, shared and autonomous vehicles, while distributed energy resources highly penetrate energy systems. How can we manage Smart City commons in a sustainable and socially responsible way to tackle long-standing problems such as traffic jams, overcrowded parking spaces or blackouts? Failing to digitally coordinate collective decisions promptly and at large-scale has tremendous economic, social and environmental impact. Coordinated decisions require a digital (r)evolution, a new paradigm on where we decide, how we decide and what we decide. But which are limiting factors? 1.Online decision-making often disconnects citizens from the physical urban space for which decisions are made: choices are less informed and vulnerable to social media misinformation, while decision outcomes may show lower legitimation. What if collective choices could be made more locally as digital geolocated testimonies, creating opportunities for community interactions and deliberation? 2.Voting system design is another origin of poor collective decisions, with majority voting often failing to achieve consensus or fair and legitimate outcomes. What if we expanded the design space of voting systems with alternative voting methods, e.g. preferential, to encompass social values? While such methods have so far been costly and limited to low-cognitive exercises, negating their social value over majority voting, decision-support systems based on artificial intelligence (AI) emerge as game-changer. 3.With an immense computational and communication complexity, large-scale coordination of inter-dependent collective decisions remains a timely grand challenge. What if coordination could be digitally assisted and emerge as a result of smart aggregate information exchange, achieving privacy and efficiency? To address these challenges, I will combine Internet of Things, human-centred AI and blockchain technology with social choice theory and mechanism design. Using IoT devices, urban points of interest can be turned into digital voting centres within which conditions for a more informed decision-making will be verified in the blockchain, e.g. proving citizens' location. A novel ontology of voting features will provide the basis to predict voting methods that generate fair and legitimate outcomes. Using collective and active reinforcement learning techniques on the blockchain, human and machine collective intelligence will be combined to achieve a trustworthy coordination of collective decisions at large scale. In collaboration with high-profile partners from government/industry, I will demonstrate the applicability of these approaches via 4 innovative impact cases. 1.Using the developed solutions, citizens will geolocate problems and vote for transport planning solutions. 2.They will also vote on spot to implement participatory budgeting projects. 3.A smart parking system will be enhanced with load-balancing capabilities to alleviate crowded and polluted city centres. 4.Via citizens' coordination of transport modality, an urban traffic control system will be optimized for an equitable shift to public/sharing transport, while preserving low-carbon transport zones. These Smart City blueprints will open up new avenues for deeper understanding of digitally assisted collective governance. To master this inter-disciplinary research area and develop myself into a future leader, I will visit world-class leaders and, together with my team, enrol in novel training activities. Two esteemed mentors and an advisory board will further support me. I will engage with the broader community of citizens and policy-makers by organizing workshops and hackathons.

Across the world, growing population sizes and increasing urbanisation cause transportation networks to reach their capacity limits. In addition, the environmental impact of the transport sector, contributing an estimated 33% of carbon dioxide emissions in the UK for 2018, needs to decrease. Thus, environmental considerations and transportation needs necessitate an increase in trips completed by active, low-emission transport, such as walking. Walking is healthy, sustainable and plays a crucial role in how urban places of work, leisure and living are accessed and used. According to the National Travel Survey 2017 for England over 80% of trips under one mile are completed on foot and considering that over 70% of trips between one and five miles long are completed by car, the potential for an increase in walking is substantial. Getting more people to walk requires better infrastructure or policy interventions, such as clean air schemes, parking fees, or incentives for walking. Currently, planners and policy makers have to make do with data from surveys or localised pedestrian counts to inform their work. However, to decide which policies or infrastructure investments will work best in promoting walking, it is necessary to consider how pedestrian traffic varies over time across the entire street network of cities. For example, making walking more attractive in one part of a city centre may influence the footfall in other, potentially unexpected locations and possibly only at certain times, such as outside of rush-hour. Despite the evident use for such information, pedestrian traffic is currently not mapped over time for cities. This project aims to change this and develop a theoretical framework for robustly constructing time-dependent pedestrian traffic maps at the scale of cities. To future-proof the methodology, it will use pedestrian counts observed at distinct locations. These can be recorded via different, privacy-preserving technologies and do not rely on the voluntary participation of individuals or private sector service providers, as is the case for data obtained from personal devices, such as mobile phones. Crucially, to ensure the traffic maps are robust to sensor failures and the occurrence of events or unscheduled disruptions, the theoretical framework will incorporate several predictive methods, each of which contributes different desirable properties, such as accurately capturing regular patterns based on historic data, efficiently interpolating between count locations and the capability to predict traffic dynamics from initial values without further data input. To directly inform the deployment of measurement devices, suitable data collection protocols will be established. Outputs of this project will be useful beyond traffic monitoring. The ability of the methodology to forecast changes in pedestrian traffic caused by construction projects will be demonstrated and the relevance of pedestrian maps for assessing pedestrian exposure to poor air quality and for evaluating the success of businesses relying on passing trade will be shown. This project will develop our understanding of city-wide pedestrian traffic and will therefore be directly useful for monitoring, across large spatial scales, long-term transport developments, short-term effects of disruptions or planned alterations and it will help the economy by informing the positioning and running of businesses that rely on passing trade, for example.