Unmanned Systems Technology 023 I Milrem Multiscope I Wireless charging I Logistics insight I InterGeo, CUAV London & USA show reports I VideoRay Defender I OS Engines GR400U-FI I Ultrabeam Hydrographic Ultra-2 I IMUs

88 Focus | IMUs, gyros and accelerometers examining the accelerometer’s data output to check for inconsistencies. It is also critical to take measures that minimise all external noise during accelerometer manufacturing and calibration. Sinking granite blocks down to bed rock may be necessary for calibration facilities near busy roads. Random walk must be tested closely, so that bias estimation algorithms can be developed for real-time mitigation during vehicle operations. Bias estimates are difficult to obtain from the manufacturer of the sensing element, and impossible to calibrate-out before integration. Gyroscopic scale-factor error, by comparison, can be easily corrected by taking readings from test motions with known angular velocities, comparing the differences between the axes’ reading values and the control values, and using a matrix of vector values to calibrate these differences. Another important target for calibrations is gyroscope misalignment. This refers to mistakes made during the manufacturing and soldering processes, resulting in errors from components on the PCB not being mounted perfectly horizontally, or the three single-axis gyros not being oriented orthogonally to each other. A horizontal motion table test (in which the IMU or gyro is moved horizontally across the table to check for anomalous roll readings) can be important for establishing the degree of this misalignment error and the level of mathematical correction necessary for mitigating it. And after being used on unmanned aircraft, many high-accuracy gyros exhibit poor recovery from the high- g manoeuvres and fast turn rates UAVs may have to conduct. Tests to detect such issues can then be followed with calibrations by the end-user or manufacturer to try to return system performance to the required levels. Headings When looking to generate heading estimates along with acceleration and angular velocity, engineers have two main choices. The first is to integrate a magnetometer which, when combined with accelerometer, gyroscope and a processor for calculating pitch and roll, comprises an attitude and heading reference system (AHRS), also known as a magnetic, angular rate and gravity sensor. There are several types of magnetometer. A fluxgate magnetometer, for example, integrates two coils around a ferromagnetic core, which is an excitation coil that produces an AC field to cyclically saturate the core, and a sensing coil which is used to measure external magnetic fields. This measurement is used to detect the Earth’s magnetic field, the direction of Magnetic North and thus the vehicle’s heading. Fluxgate sensors are perhaps the most widely used type of magnetometer in navigation systems, with several dozen times less noise and temperature sensitivity than other types. Hall effect magnetometers, the most common type, work by first applying a current through a proof mass such as a conductive plate. Lorentz forces cause a displacement in the flow of electrons in response to magnetic fields affecting the plate, generating a small corresponding voltage. By equipping an amplifier to the plate, this voltage can be detected, measured and used as a compass. Anisotropic magneto-resistive magnetometers are made of nickel-iron (or ‘permalloy’) material in a thin film. The material has a magnetisation vector in a single axis, which rotates when an external field is applied to it. The heading is derived from deviations in this vector caused by the December/January 2019 | Unmanned Systems Technology Unmanned systems can use magnetometers or dual GNSS antennas to track headings independent of vehicle motion (Courtesy of XSens) Many tests conducted during IMU calibration rely on readings taken from repeated motions, and measured against control values for those motions (Courtesy of UAV Navigation)