Ex Machina – Integrating maritime robots into human endeavours
Development in marine robotics over recent decades has generally focussed on the technical challenges. Many of these core challenges are solved, at a price. We can certainly continue to work on better/cheaper solutions, but this may be missing the point. This presentation argues that the bottleneck to bringing robotic power to bear on maritime operational problems is now largely clustered around necessary changes in people and process, e.g. culture, trust, ethics. This is compounded by the shift from one-on-one human systems-to-robotic systems operations to distributed robotic architectures, where the power of emergent behaviour and large spatial-temporal apertures become possible. The fundamental constraints on communication bandwidth and range inevitably drive solutions to address the need for AI and autonomy in the UW world, contrasting with UAS developments. How might processing, command and control best be distributed in a multi-asset system? There’s also something exciting going on with cloud architecture, edge computing, deep learning and AI that translates into the UW Maritime domain. The future lies in locally-smart devices at the edge, with distributed sensing, processing and decision-making performed locally at each node to enable agile action, with slower global deep learning loops running back to the cloud through non-real-time data harvesting and concentration through a gateway. This is a dawning game-changer and will require mastering distributed control, emergent behaviour, managing trusted networks and much more, all in a bandwidth-constrained domain.