Configuration of skilled tasks for execution in multipurpose and collaborative service robots

Abstract

Several highly versatile mobile robots have been introduced during the last ten years. Some of these robots are working among people in exhibitions and other public places, such as museums and shopping centers. Unlike industrial robots, which are typically found only in manufacturing environments, service robots can be found in a variety of places, ranging from homes to offices, and from hospitals to restaurants. Developing mobile robots working co-operatively with humans raises not only interaction problems but problems in getting tasks accomplished. In an unstructured and dynamic environment this is not readily achievable because of the high degree of complexity of perception and motion of the robots. Such tasks require high-level perception and locomotion systems, not to mention control systems for all levels of task control. The lowest levels are controlling the motors and sensors of the robots and the highest are sophisticated task planners for complex and useful tasks. Human-friendly communication can be seen as an important factor in getting robots into our homes. In this work a new task configuration concept is proposed for multipurpose service robots. The concept gives guidelines for a software architecture and task managing system. Task configuration process presents a new method which makes it easier to configure a new task for a robot. The idea is the same as when a person tells another how a task should be performed. Novel method for executing tasks with service robots is also presented. Interpretive execution, keeping the focus on only one micro task at a time, makes it possible to modify plans during their execution. Multimodal interaction is important feature to provide collaboration between humans and robots. Multimodal interaction reduces the workload of the user by administering task configuration and execution. A novel solution for using multimodal human-robot interaction (HRI) as a part of the task description is presented. This thesis is a case study reporting the results when developing a task managing (from configuring to execution) platform for multipurpose service robots and studying its performance and use with several test cases. The platform that was developed has been implemented with the WorkPartner multipurpose service robot. The structure and operation of the platform have proved to be useful and several tasks have been carried out successfully.