A framework for formal specification of robotic constraint-based tasks and their concurrent execution with online qos monitoring

Combining tasks, melding their activities in sequence and in parallel, in order to achieve the desired goal is a challenging research topic. In many practical applications, tasks could be fulfilled even though the robot does not achieve a perfect matching with a given quantitative objective. Thus, it could be possible to carry out other activities, mildly conflicting with the current goal, yet providing benefits for the overall execution of a complex sequence of tasks. This paper proposes i) a criteria to evaluate the execution of a monitored task, ii) an enriched task specification to express the tolerance with which a goal is fulfilled, so that iii) a sequence of tasks can be executed concurrently, on the basis of monitored quantities. Furthermore, the paper reports a classification of conflicting scenarios for a finer selection of the scheduler policy. Finally, results from experimental scenarios show the potential benefits of the proposed methodology.