Gyro Stabilization of a Positioning Unit

Mounting cameras on motorized objects has become possible in greater extent due to the emerging camera technology during the past few decades. This application could prove useful in several areas such as search and rescue operations, surveillance or even news monitoring. One hardship that this brings is the difficulty of keeping the camera unit stable while its setting is not. In the situations mentioned above, it would be desired to keep the camera unit in the same attitude during the operation, regardless of the impacts from the surrounding environment. The goal of this thesis has been to solve this problem using gyro stabilization by implementing a stabilization algorithm to be added in the camera unit’s software. The idea behind gyro stabilization is to counteract for external disturbances causing the camera unit to dislocate from its desired attitude. This is realized by controlling two BLDC, Brushless DC, motors in a positioning unit (PU) provided by Axis Communications AB. The PU could be rotated about two axes (using the two motors) perpendicular to each other, pan and tilt, and could hold either a network video camera, or a set of illuminators. The control of the motors is achieved by utilizing data from an IMU, Inertial Measurement Unit, including a gyroscope and an accelerometer, whose measurements are fused in a Kalman filter. This data, together with an estimation of the orientation of the PU, is used to calculate the error between the actual position and the desired position of the PU. The orientation estimation is found by relating the coordinate system of the PU to a fixed, global coordinate system. The calculated error is then minimized by using a PID controller to control the velocity of the pan- and tilt motors and ultimately the position of the PU. Most of the calculations and simulations regarding the stabilization algorithm were carried out in MATLAB Simulink, and later implemented in the PU’s software, written in C code. The final algorithm was tested both on an ABB IRB2400 robot, and mounted on a small boat driven by an outboard motor, made possible by an collaboration with the research program WASP. The results were satisfying and the PU was successfully stabilized, which was the main goal of the thesis. However, there is room for some minor improvements, for example improving the modeling of the motors