When extreme weather strikes, the biggest obstacles are often the ones left behind. Fallen branches, debris and damaged infrastructure can block access for emergency services, making the final metres of an intervention the most difficult.
This challenge inspired Konstantinos Tsintotas, winner of the Third Open Call of dAIEDGE, to develop SPOT-LIFT, a project that combines robotics and Edge AI to assist civil protection teams in clearing debris safely and autonomously.
Built around the Boston Dynamics Spot robot, SPOT-LIFT equips the platform with an uncertainty-aware AI agent capable of perceiving, analysing and manipulating objects directly on-board. Rather than executing actions blindly, the system evaluates the stability of each grasp, estimates the centre of mass of an object and continuously assesses the confidence of its own decisions before lifting or repositioning debris.
Keeping every perception and decision-making process on the robot itself is fundamental to the project. By eliminating cloud dependency, SPOT-LIFT delivers predictable latency, greater robustness and reliable operation even when connectivity is limited—conditions that are common in emergency response scenarios.
Throughout the programme, Konstantinos collaborated with VERSES, whose technical guidance helped shape the project from both a deployment and performance perspective. We would like to thank VERSES for their commitment as hosting institution and for supporting the team throughout this journey.
The project also benefited from the dAIEDGE-VLab, where different optimisation strategies and deployment configurations could be explored under realistic Edge AI conditions. This environment allowed the team to validate performance against measurable KPIs and move from a research prototype towards a reusable SDK designed for the wider European Edge AI community.
SPOT-LIFT demonstrates that trustworthy Edge AI is not only about making robots autonomous—it is about enabling them to make safer decisions when every second matters.