In this paper, a physically interactive control scheme is developed for a manipulator robot arm. The human touches the robot and applies force in order to make it behave as he/she likes. The communication between the robot and the human is maintained by a physical contact with no sensors. The intent of the human is estimated by observing the change in control effort. The robot receives the estimated human intent and updates its position reference accordingly. The developed method uses the principle of conservation of zero momentum for position-controlled systems. A switching scheme is developed that goes between the modes of pure impedance control with a fixed-position reference and interactive control under human intent. The switching mechanism uses neither a physical switch nor a sensor; it observes the human intent and puts the robot into interactive mode, if there is any. When the human intent disappears, the robot goes into the pure-impedance-control mode, thus stabilizing in the left position.

Design often includes ill-structured problems that cannot be defined in a clearly identifiable and codifiable problem space. This paper tries to explain, from the viewpoint of design knowledge, how such ill-structured problems can be formed and how to tackle them. Particularly, the paper focuses on ill-structured problems that can be found in multi-disciplinary design problems. In this context, methods to deal with ill-structured problems in the context of computer supported cooperative work could be either to establish a unifying theory for relevant areas or to build ontological descriptions about relationships among those relevant areas

This paper proposes a new cellular-type machine to solve problems the present mechatronics machines have through their life cycle. It consists of homogeneous cells with modularity and autonomy, and employs a local communication mechanism. Each cell autonomously behaves based on local information accumulated through local communication. It can logically reconfigure itself, even when some cells are faulty. It can also be physically reconfigured by adding or removing some cells. Each cell periodically evaluates its local environment and its own action history, and selects its appropriate function. This leads to self-organization. It also has an advantage that no central software is needed to control the entire system. We designed and built a "cellular automatic warehouse" as an example, which demonstrated the advantages of the cellular machine.