Implementation of Prospective Memory in a Mobile Robot

This paper presents implementation of prospective memory in a mobile robot. Predefined images, sounds are stored in the memory of the robot. When the robot is in use, it can carry out predefined tasks like identifying similar events in the working environment, when such tasks are stored in the memory of the robot. The robot memory can be loaded with different tasks based on requirements on daily basis. Based on the amount of expert knowledge the robot is built in, all defined tasks will be carried out by the robot. The work implements a robot for monitoring of production machines status in an industrial environment. The robot acts as the supervisor and switches a machine on when a task has to be done by the machine. In any case the machine is in idle status, it will be switched off by the robot. Video information of the robot moving in a manufacturing with the presence of sounds from machines are recorded and used for further analysis. The sounds acquired have been analyzed using cepstral analysis to identify the machines working status and the video information is processed with contextual clustering for proper image segmentation to enable robot move along the path of working cell for further monitoring.