Enhancing critical ecosystems

This work aims to explore how combinations of sensing instrumentation, actuation and a spectrum of data analytics to not only improve and protect the assets within an ecosystem, such as a City, but to permit more ‘joint-up’ thinking about how such assets interact and impact upon each other in that ecosystem. For example, water companies want to optimise customer quality of service and pipe lifetimes, and are not really concerned with transport infrastructures, and vice versa. Imagine a future where when one network detects a leak, the road traffic is automatically re-routed to avoid this. Better still the water network predicts a leak and schedules maintenance activities. In doing so this work will enable more efficient ecosystem operation which in turn can save resources making them more sustainable and this shall improve the long-term safety of such systems (ie water delivery and food production) and of the person and property.

This is a multi-disciplinary proposal bringing together sensor network, food production (farming) and water networks researchers and user partners that addresses key priorities, end user needs, and challenges; with a diversity meaning that solutions will be truly transformative and impactive. The work is scheduled to run over 5 years and aims to answer the following overarching research question:

Through data analysis, how can we enhance critical ecosystems and the cyber-physical systems that support them continuously?

There are a number of research objectives we require to answer this question:

  • Understanding sensor selection and placement to optimise observation
  • Understanding the edge device analytics algorithms to improve the efficiency of the IoT-based system maintaining usefulness of the data from sensors
  • Deriving alternative control methods (valve closure, fertilizer release etc.) that incorporate advanced event-based control, aperiodic control and protocols/algorithms to support this efficiently.
  • Experimenting with Data driven modelling of critical infrastructures & dynamic model update
  • Exploring how the physical world, self-monitoring, privacy and security impacts the trustworthiness of data for such systems

 

Taking water distribution networks and precision agriculture as two examples of critical infrastructure this work stream will initially instrument a sensing computing infrastructure to obtain data from each. Using data driven modelling, the operation of these will be mapped to a control model and the control devices will be deployed to close the loop. Water distribution networks are large scale and topologically complex with many constraints that impact their operation (customer demand, weather, pipe lifetime etc.). Agriculture also brings large scale and a diversity of data (soil, foliage, weather etc.). Further for the latter the major barrier to developing precision agriculture systems concerns the trust of such system; though farmers acknowledge the benefits that can be obtained from being part of a network of farms (e.g. early warning of infestations etc.). For the former, after cost, resilience of the system is the next barrier in the water industry. In these kinds of systems the data required to operate the control loop and guarantee computer system and water systems lifetimes and to keep the them both secure are not necessarily the same, therefore understanding these trade-offs is very important.

Part of The Alan Turing Institute-Lloyd’s Register Foundation Programme for Data-Centric Engineering

Funded by :

Lloyds R