A dynamic safety system based on sensor fusion

Machines in industry, including industrial robots, have in many cases drama tically reduced the man-made work and improved the work environment. New ma chines introduce, however, new risk factors. Traditionally machines are saf eguarded by means that more or less rigidly separates the machines from the personnel. This works well in many traditional areas, i.e., where industri al robots are involved. There is however a risk that the safety system limi ts the valuable flexibility of the robot, which can be considered as a qual ity that tends to become even more valuable in the progress of programming possibilities and sensor technology. This article shows an example how a sa fety system can be designed to achieve increased flexibility in co-operatio n between human and production safety strategy. The proposed safety system is totally based on sensor information that monitors the working area, calc ulate the safety level and improve the system dynamically, e.g., reduce the robot capability in conjunction to the system safety level. The safety sys tem gain information from the sensors and calculates a risk level which con trols the robot speed, i.e., the speed is reduced to achieve a sufficiently low risk level. The sensor data is combined with fuzzy-based sensor fusion and fuzzy rules. The safety system is based on sensor information, hence i t automatically adjusts to changes in the guarded area as long as the funct ionality of the sensors is maintained. Finally, we present a system impleme ntation in an industrial robot application.