Task monitoring and rescheduling for opportunity and failure management

Abstract

The CoBot robots, as other service robots, autonomously navigate in building environments performing different types of tasks that include item transportation and person guiding between locations. The CoBots can execute their planned routes, localize in the environment, avoid obstacles, and ask for help to humans to overcome their actuation limitations. However, they were not able to handle high-level unexpected events during execution, such as interruptions with new task requests that may need a careful analysis of rescheduling trade-offs. Unexpected events can be failures if their influence is on the pending tasks or opportunities if it is on the robot expectations. This work presents a new task-execution, monitoring, and rescheduling architecture, which includes a representation of new task features to be monitored to detect failures and opportunities, as well as a task scheduler to evaluate time and task features constraints. We demonstrate the new features in a task that needs to deliver hot coffee at some time, noting that the coffee gets cold with interruption delays.