Trustworthy Autonomous Systems

Our School undertakes research into a wide range of technologies relevant to the development of trustworthy autonomous systems. These include: formal techniques to reason about behaviour; methods to analyse human-robot interaction; and the development of robot vision systems and network protocol infrastructure to support robot-to-robot communication and access to edge-compute resources.

Theme Lead: Professor Alice Miller

Banner for Trustworthy Autonomous Systems

Track record - staff

The theme brings together researchers from across the School who are involved in a wide range of projects relevant to the development of trustworthy autonomous systems. These include: formal techniques to reason about behaviour; methods to analyse human-robot interaction; system intelligence and self-awareness; and the development of robot vision systems and network protocol infrastructure to support robot-to-robot communication and access to edge-compute resources. Examples of activities in this area are outlined below.

  • FATA (Miller, Norman, Calder, Andrei, Gay, Dardha, Enright and McCreesh)
    • Development of a range of models for the simulation and verification of Unmanned Aerial Vehicles. Funding includes a variety of EPSRC IAA awards.
    • Modelling and reasoning about learning robots, satellites and smart railway systems. 
    • Computational modelling and automated reasoning about the behaviour of complex, interactive, and sensor-driven adaptive and/or autonomous systems.  Funding includes BBSRC and EPSRC projects such as Science of Sensor Systems Software.
    • Programming language approaches, especially type-based, to reason about correct communication in concurrent and distributed software.
      Funding includes the EPSRC ABCD project  the EU BehAPI project.
    • Formal data-driven models and analysis methods for interactive systems and social interactions to interpret human behaviour in the context of pervasive, autonomous technology. 
    • Verifiability of constraint based tools for autonomous decision making and scheduling.
  • IDA (Aragon Camarasa, Siebert, Anagnostopoulos)
    • Robotic manipulation and grasping, chemical robotics and machine perception.    Funded by Innovate UK and EPSRC.
    • Visual sensing and computer vision architectures based on biological systems for control and learning in hand-eye robotics applied to advanced manufacturing and autonomous systems.
    • Distributed intelligence algorithms for unmanned vehicles in resource-constrained enviroments. GNFUV project funded by EU/H2020.
  • GIST (Brewster, Foster, Chalmers)
    • User interface design for interaction with autonomous systems, including the design of user interfaces for autonomous and semi-autonomous vehicles, the effective handover of control between driver and vehicle, and user experience design for passengers of autonomous vehicles.
    • Models of both human and robotic behaviour and the measurement of the performance of said models.
    • Ethical systems design, especially those using the concept of 'human-data interaction design’. This involves legibility, agency and negotiability with regard to how complex data-driven systems, such as autonomous systems operate. 
  • GLASS (Trinder, Cano Reyes, Pezaros, Singer, Michala, Perkins)
    • Reliable Scalable Software Systems, many of which use, or reason about, a reliable distributed actor model like Erlang or Scala with Akka. Important application areas are large-scale servers and robots. Supported by a variety of funding including the EU RELEASE project.
    • Providing Machine Learning capabilities to mobile/embedded edge devices (e.g. robots, drones and satellites).  Applications include autonomous navigation, image classification, video detection, speech recognition. Funding includes the EU BONSEYES project. 
    • Measurement-based, closed-loop network and service management mechanisms for adaptive and resilient operation of mission-critical networked systems. Focus is on self-management, self-configuration, self-optimisation, and self-protection properties.
    • Investigation into programming language runtimes with predictable and/or autonomic behaviour (e.g. bounded resource utilization).
    • Data analytics and machine learning on data streams to optimise network  usage, information extraction, privacy, stakeholder access requirements and clearance. Applications include autonomous and unmanned ships and autonomous satellite systems. 
    • Robust and secure network transport protocols for real-time traffic, including audio/visual data; network protocol standardisation; and specification. Scalable edge and in-network computation architectures, funding includes the EPSRC FRµIT project.