88 Focus | IMUs, gyros and accelerometers accelerations and rotations to examine its output readings, and hence program compensatory subroutines into the Kalman filter. Automated calibration quality control checks and product functionality tests can then follow to ensure devices conform to standards expected by customers in every respect, from solder correctness to bias over temperature. While it is rare, some MEMS IMUs are now supplied with their individual calibration reports, enabling customers to see input-to-output discrepancies, scale-factor errors and other biases specific to an IMU to understand what has been calibrated for. FOG expands New generations of FOG take at least five years to come out. A principal reason is because engineering and bulk manufacturing a new system can require complete changes to the associated machinery and tooling to ensure fibre can be manipulated, welded and coated differently, and that new integrated optical chips (IOCs) and other associated components can be handled, particularly as these are reaching sizes of 10-40 µm and require more precise equipment. Manufacturing a FOG begins with extrusion of an optical fibre to the desired thickness and length (which descends from the extrusion point under gravity in most cases), which is followed by the winding of that fibre into a coil. That coil is then coupled with the light source (typically a form of diode), a splitter to ensure light can be sent in opposite directions through the fibre, and the signal detector. The complete coils can then be installed inside the IMU enclosure, along with an accelerometer. Among the FOG IMU makers seeking to minimise packaging volume, an unusual approach of manufacturing components such as polarisers directly on to the fibre has been developed, instead of performing any splicing (the forced joining of two ends of fibre together, typically by electric arc welding). This involves using heat to shrink the end of a fibre to around one-sixth to one-seventh of its usual thickness, and then growing the polarisation crystal on that shrunken end. Doing it this way saves on component costs and avoids the performance losses that can result from splices – resulting in a more power-efficient IMU. Other changes and indeed expansions to production are being undertaken by many leading FOG manufacturers to enable outputs in higher volumes and at lower unit costs, with at least one looking to grow its manufacturing capacity by 10 times. Doing so requires exhaustive planning, foresight and investment, going down to the very foundations of a production facility. The calibration of FOGs, while technically a similar process to calibrating MEMS, is even more sensitive to vibration; hence, a very large quantity of concrete in the foundation slab (and pinning all of that to bedrock) is imperative for reliable calibration and hence high throughput of FOGs. Coil winding Ensuring such foundations minimises the risk that heavy production equipment – such as coil winding machines, composite potting systems for ensuring the coils’ solidness and thermal stability via polyurethane, or three-axis temperature chambers for calibration – will cause vibrations that resonate through to the rest of the facility and the products inside. The predominant method for coil winding today is quadrupole winding, in which two spools take turns supplying fibre on to a product spool from the centre outwards. As FOGs are highly sensitive to winding nonlinearities, minimising vibration is again critical to successful quadrupole winding, as is minimising error in the winding process. New generations of production lines are therefore being designed with greater automation for traceability and repeatability, with specialised machinery and machine learning applied to identify any errors that have occurred during winding, potting or other processes. Polishing of fibre ends, for instance, has traditionally relied on machines activated and operated by human workers, but today this can be fully automated, with consistent results to ensure a flat February/March 2024 | Uncrewed Systems Technology On top of integrating with better GNSS sources, the use of visual, radar and laser odometry (with exhaustive programming and error modelling) are hallmarks of the latest INSs (Image courtesy of OxTS)
RkJQdWJsaXNoZXIy MjI2Mzk4