Programmable Autonomous Systems
The increasing proliferation of Internet enabled things and services poses huge potential benefits for society.
With an expected 75 Billion connected devices by 2020, the flexibility of digital platforms must accelerate to cope with complex system requirements. Evolving cloud platforms, communications protocols, end user devices are all bringing new challenges to build effective end-to-end systems that can scale and intelligently self-manage. One of the most difficult challenges facing these digital platforms in the near future is how to cope with such rapidly changing requirements on a wide set of criteria.
The Programmable Autonomous Systems Unit seeks to research, develop and commercialise techniques, tools and algorithms that will ensure that citizens and businesses can benefit from this digital evolution. We advocate that future digital platforms must exhibit a set of fundamental characteristics if they are to cope with rapidly changing requirements in such a complex environment.
- Programmable – a system, service or platform whose logic is separated from its operation. By building programmable systems, we ensure that they are future proofed and flexible so that their features and functionalities can evolve as requirements evolve.
- Autonomous – a system, service or platform is autonomous if it can perform its activities in a fashion that requires minimal human supervision in the face of complex and changing environmental conditions.
By directing out research into these two highly complementary topics, we can ensure that society can take full advantage of the new digitally connected world.
An energy conscious society is pushing for more efficient and cost effective strategies to control energy usage and to increase the usage of renewables. This opens up researching into the following areas: Distributed energy systems – moving away from large decentralised power generation with long transmission systems to smaller, intelligent and local generation; Virtual network metering/Peer to peer sharing – allowing prosumers (those producing and consuming energy) to share that energy with neighbours with balancing of transaction costs; Supply/demand balancing, peak shaving – intelligent demand usage away from peak expensive times; Time of use tariff incentives – incentivisation of consumers to alter their energy profiles; Renewables intermittency – balancing of the system with high penetration of variable renewable energy resources.
The electrical grid is undergoing a paradigm shift from a large centralised, unidirectional system to an intelligent network where supply and demand are balanced with distributed and renewable sources of energy using aggregation, flexibility and empowering prosumers to take control of their energy. We are addressing how ICT systems will drive this new paradigm and facilitate the intelligence and control to allow end users become prosumers.
Incorporating variable and intermittent renewable energy technologies such as wind and solar is critical to decarbonising our environment and reducing our reliance on finite fossil fuels. This theme focusses on forecasting and predicting the generation capabilities of these distributed resources based on the myriad of environmental and infrastructural conditions that affect them.
- Steven Davy
- Michael Crotty
- SFI Connect Research Centre
The PAS Framework will aim to orchestrate other systems to take advantage of new data, new logic, new algorithms that would typically not be available, or only available to a human operator.
This approach aims to bring new levels of operational efficiency that
would typically not be possible.
The Orchestration will be carried out through APIs exposed by existing systems and will be from a centralise decision point or distributed through embedded agents instructing the systems.
Initially, we are seeking to have an impact in Smart Energy and Smart Manufacturing Automation.