ABSTRACT
An important motivation for developing a skill-capturing and guidance system based on optical means is to program and operate service robots, which serve users in a variety of ways such as feeding paralyzed or handicapped people at home and in hospitals or cleaning toxic air filters in laboratories. Because service robot users are usually ordinary people, an easy-to-use robot programming approach is vital if these kinds of robots are to become popularized. The traditional textual and key-in programming approach requires considerable learning and preparation time that is not suitable for service robots. Over the last decade an easy-to-use alternative approach, which has attracted much research attention, is
robot programming by human demonstration
[Delson and West, 1994; Ikeuchi and Suehiro, 1994; Yang et al., 1997]. This approach involves two phases of operation-programming and execution-that may usefully apply skill-capturing and visual-guidance technologies, respectively.
