Multi-robot systems (MRS) have gained interest as a versatile paradigm for complex task execution across various domains such as healthcare, logistics, and maintenance. Often, they are called to operate in variable and dynamic environments, which makes uncertainties arise and affect those systems. Uncertainties require the system to be able to adapt its behavior at runtime, in response to the changing and unpredictable conditions in its operating environment. Moreover, often the behavior of the robots cannot be completely anticipated at design time. Consequently, static mission planning is not always suitable: mission specifications need to take into account the uncertainties and, hence, be dynamic and re-configurable at runtime, when the required knowledge is available.

This work focuses on the realization of adaptable multi-robot systems, which are capable of dealing with uncertainties by adapting their mission at runtime. We introduce the concept of “adaptable task” that is used in the global mission specification of the MRS to identify the mission tasks affected by uncertainties. Adaptation alternatives are modeled as sub-missions and associated with the adaptable task. At runtime, ad hoc written “trigger functions” executed by robots sense and evaluate the environment and select the most suitable adaptation alternative to be executed.

We have experimented with the approach by simulating a use case to assess its validity. The system was able to adapt its behavior in response to the environmental conditions, thus allowing the fulfillment of the mission goals. We also discuss the applicability of the use case on a set of known single- and multi-robot systems.