Intelligence group report

Introduction.

The field of Biorobotics, and service robotics in particular is making rapid advances [Schraft94]. The Intelligence group of the Biorobotics Workshop held in Tsukuba, Japan in May, 1995 took on the role of surveying the state of the art of intelligent systems as they apply to service robotics, discussed the latest advances and needs for the next generation of these systems, and then charted out a roadmap for further development of intelligence in service robotics into the next decade. The group of researchers and practitioners who participated made brief presentations, and then discussed issues, challenges, and tasks to define the next generation of service robot systems. This report summarizes the different ideas within a cohesive framework, and then presents a set of grand challenges that are likely to govern research and development in the future.

Service robots are defined as adaptable, sensor-based mechantronic devices that perform useful services for humans. Kawamura et al. [Kawamura95] classify service robots between industrial robots, such as welding and assembly machines, which operate in well-defined, pre-programmed, structured environments, and field robots, such as fully autonomous vehicles like the Mars Rover, which often have to operate in dynamic, unknown, and uncertain environments.

Unlike field robots, service robots are designed to perform tasks in specific environments. Examples of service robotics applications include assistance systems for the disabled and handicapped, hospital patient aides, office automation systems, and worker robots that assist humans in difficult and high precision tasks. The very nature of the applications and the important requirement of symbiosis with humans requires these robots to perform in a semi-autonomous and interactive framework. Furthermore, safety, efficiency, reliability, and the ability to communicate, understand, learn, and adapt are key issues that govern the successful implementation of such systems.

Along with sensory ability and control, intelligence, i.e., the ability to reason about, understand, and solve problems, plays a key role in the design and implementation of service robotics systems. A first step in the group discussion was to identify primary features for characterizing the concept of intelligence. These are discussed briefly below.