Two Methods for the Determination of Inertial Sensor Parameters

Inertial sensors, gyroscopes and accelerometers are widely applied. Inertial navigation systems (INS) usually consist of the inertial measuring unit (IMU) with three mutually perpendicular gyroscopes and accelerometers that represent three input axes to the system and navigation algorithm. Input values for the navigation algorithm are signals of gyroscopes and accelerometers. It is very important to know sources of navigation algorithm errors. One of the sources is a bias instability. A bias instability is a parameter that defines a class of sensors. The class of gyroscopes is defined in deg/h or deg/s and the class of accelerometers is defined in mg or µg. Other parameters defining a random walk of sensors are the angle random walk of gyroscopes and the velocity random walk of accelerometers. Although manufacturers declare these parameters in their data sheets, it is very important to confirm all declared sensor parameters before application. Two methods of angle random walk, velocity random walk and bias instability are presented and discussed with reference to their precision. These two methods are based on the Allan variance and power spectral density (PSD). The analysis was performed on the Analog Devices ADIS 16365 six degree of freedom sensor