Work Packages

WP1 Enhanced Machine-Machine Teaming

Workpackage one focuses on two main research questions:

•RQ1 Multi-party Teaming: How can multiple autonomy architectures negotiate responsibilities?

•RQ2 Robotic communications: How can autonomous systems efficiently share data and information to have a coherent view of the world and of the mission goals?

WP2 Robust Robotic Perception

Workpackage two focuses on three main research questions:

•RQ3 Continual learning: How can a mix of real and simulation environmental data be used by learning-based algorithms to improve performance, adapt to new sensors and environments, and build reliability? 

•RQ4 Common Scene Understanding: Unmanned systems now collaborate across multiple domains, equipped with a wide range of sensor modalities. These systems produce data generated by different types of sensors, with each sensor type revealing a particular feature of the observed phenomenon. What can be done to break the barrier between different sensing modalities to enable true multi-robot collaboration? 

•RQ5 Quality Assurance: How can Quality Assurance (QA) for data and analysis be estimated automatically? Trust and successful manned-unmanned teaming is dependent on reputable, reliable and consistent output from the onboard perception algorithms, including a quantifiable Quality Assurance process for both the input data and the resultant analysis.

WP3 Human-machine interaction and explainability

The final technical workpackage has the following research questions:

•RQ6 Joint Optimisation of User and System Objectives: How can we take into account both the system and user’s objectives, in terms of actions to complete and information to be given to maintain situation awareness? Does this personalised, human-in-the-loop approach help to manage user trust and facilitate human-robot teaming, thus increasing adoption?

•RQ7 Explainable Deep Machine Learning for Robot Perception: How can we generate explanations of decisions made by complex machine learning algorithms performing robot perception?