Precision and fine motor skills are necessary for the execution of various tasks and can be crucial whether robots can perform a given task. Sensor fusion is a fundamental tool for the accurate execution of robotic tasks. For this purpose, sensor fusion aims at capturing a complete picture of the environment. This goal is achieved by combining multiple sensors with different sensor modalities. For example, an object can be „seen“ and „felt“ by a robot.
In the Festo System Laboratory, surfaces are processed in various ways using a robotic system. Therefore, sensors to measure in particular the interaction between the robot, the surface to be processed and the tool are used. For the interaction it is interesting, for example, how the workpiece and the tool are aligned to each other and how much distance there is between the tool and the surface. Sensors that work with laser beams or inertial measurement units (IMU) can record this information. It is also important to obtain the interaction forces with which a tool acts on a surface during the process. Combined force/torque sensors are available for this purpose. Other sensors that provide information about the interaction of robot, tool and workpiece are, for example, capacitive or inductive sensors.
In this project, sensor fusion is used to reliably and accurately estimate the position and orientation of a tool mounted on the robot end-effector. When the tool is processing a surface, sensor fusion is the basis for estimating the contact forces that emerge between the tool and the processed surface. The selection of the sensors and the sensor fusion methods thus form the basis for precise processing of object surfaces performed by a robot.