Performance improvement of a robot photographer using a multiple human detection system to activate the community

Service robots for daily work activities have been anticipated for several years even though it is difficult for companies to estimate their market demand and future usage. To overcome these problems, various types of robots are to be developed and verified by demonstration tasks using standard software. If many developers use the standard software to develop robots, the possibility to find a killer application of the robots will be increased and the service robot market will be widespread. Thus, we developed a prototype of a robot photographer using Robotics Technology Middleware to activate the community. The robot was developed quickly and easily. However, the success rate for capturing pictures was low in various environments. Therefore, we analyzed the cause of failures and proposed a correction algorithm. To improve the robot performance, we developed a multiple human detection system with recovery motion.