To control the stability of objects where continuous motion and deformation are expected, measurement techniques are used that provide continuous kinematic monitoring of the object's motion with respect to a stable environment. Modern robotic total stations (RTS) in kinematic measurement mode allow position determination with frequencies around 10Hz. In the case of oscillations of objects with higher frequencies, it is no longer possible to determine the representative trajectory of the object with these methods. In the dissertation we investigate the possibility of improving the results of kinematic measurements of a robotic total station by combining high-frequency measurements of a low-cost inertial measurement unit (IMU). We analyzed the quality of the results obtained by using both measurement systems separately. For processing the kinematic RTS measurements, we presented the Kalman filter method. For processing the measurements from low-cost IMU sensors, we developed a model and a method for calibrating the sensor. By using a standalone low-cost IMU sensor without further processing of the results, we determined the trajectory of the object in which a drift of several hundred meters occurred after one minute of operation. We analyzed methods that can largely eliminate the drift, and the best results were obtained with a high-pass Zero Phase filter. Kinematic RTS measurements and IMU measurements were combined using an extended Kalman filter measurement model in combination with a smoothing procedure and a method for combining trajectories obtained by linear interpolation of the differences between trajectories determined by RTS and IMU measurements processed with the Zero Phase filter. We have confirmed the suitability of the methods with experimental work, and in our case the results of the kinematic RTS measurements were improved by almost 50%.
|
