QoS Metrics-In-the-loop for better Robot Navigation
MIRoN ITP
Navigation is an essential capability in most robotic solutions, being basic in a wide range of scenarios, such as maintenance, inspection or factory intralogistics. However, traditional navigation approaches are no longer adequate due to their little flexibility. Dealing with variability in open-ended environments requires robots to adapt themselves according to the current situation in order to achieve the required quality of service. In this sense, runtime adaptation allows moving autonomous navigation one step forward. The ambition of MIRoN is to provide a complete framework able to endow robots with the ability of self-adapting its course of action according to the external and internal context at runtime.
The MIRoN framework, delivered as an Eclipse plug-in, will provide both modelling and code generation tools enabling the creation of RobMoSys-compliant systems with adaptive navigation capabilities. At design-time, the framework is intended to support the modeling of
- variation points, which determine the decision space of the adaptation process, i.e., the answer to what can be adjusted. Variation points will be linked to elements in the specification of the robotic behavior (e.g., tasks and skills);
- contexts, mainly expressed in terms of RoQME[1] QoS metrics; and
- adaptation policies, explicating how to configure the variation points depending on the current context in order to optimize relevant non-functional properties (NFP), such as safety or performance.
Expected Impact:
- MIRoN is contributing to RobMoSys by developing a model-based framework for dealing with adaptive robot navigation. In particular, the proposal is built on the Flexible Navigation Stack[1] to provide a novel approach to adaptive navigation based on the systematic use of models for dynamically reconfiguring the robot behavior, defined in terms of Behavior Trees (BT), according to the runtime prediction and estimation of QoS metrics defined on NFPs.
Follow @MiRON_ITP on Twitter
