80 and gyroscopes in an IMU, or between an IMU and its external aiding systems, such as GNSS receivers, Lidars or cameras. Scale-factor errors result from poor time synchronisation in inertial systems, which can then accumulate over time into severe navigation drift, owing to the discrepancy between an IMU’s clock and the uncrewed vehicle’s clock. The severity of that will increase with the dynamics of the vehicle, making higherspeed UAVs and UGVs particularly vulnerable to timing-related errors. Time synchronisation is a complex robotics problem, which companies often resolve through a robust eventsmanagement system to ensure timestamp accuracy within and without an IMU. Key elements of such a system include a number of different inputs for synchronisation purposes; for instance, the software can be defined such that every pulse of inertial data output triggers an event with a specific log timestamped using the IMU’s internal clock. Alternatively, an event-marker log can be sent whenever an event external to the IMU (such as a Lidar pulse, a camera shutter or a GNSS update) is received, be it by CAN, GPIO or other network bus, thereby timestamping each event. And, often, when integrating an IMU with a GNSS, a key part of the overall INS design is utilising the GNSS’s pulse-per-second (PPS) signal to realign and synchronise the IMU’s clock to the GNSS clock (the latter being typically defined by an atomic clock, placing them among the most trustworthy sources of timing data). Low latency Another critical part of ensuring accurate real-time control is minimising the output latency of inertial systems; that is, the time delay between when a motion is imparted on the IMU and when a packet of data capturing that motion has been generated and output. Keeping latency as low as possible is vital to ensuring precise control loops, both in navigation and gimbal pointing. This can largely be accomplished through software, specifically via an extended Kalman filter, optimised to calculate and output data packets using only minimised and condensed computations of the raw motion data. Also worth noting within the umbrella of timing- and latency-related performance parameters is timestamp repeatability. This refers to variances in timestamp accuracy, which are typically introduced via internal hardware-processing delays, and are hence resolved through the selection of higher-end components, as well as the testing and iteration of board designs to minimise concentrations of crosstalk or hotspots. Furthermore, mitigating problems stemming from hotspots, and generally ensuring the consistent performance of an IMU across temperatures, remains a vital guarantor for success in real-world applications, and failing to do so results in comparable degrees of output error to timing failure. Uncrewed vehicles increasingly undergo wide ranges of environmental temperatures during their missions. Robots running deliveries or industrial inspections, for instance, will drive (or walk) inside buildings with conditioned environments (airconditioning or central heating) before moving outside into the summer heat or winter frost. Inertial systems will be among the first to suffer severe performance deterioration without sophisticated compensation for temperature-induced errors in the Kalman filter. MEMS instruments Of the primary sensing elements within MEMS gyroscopes and accelerometers, numerous types are available and accounted for in technical literature. Types of inertial sensing among these vary, from capacitive MEMS that operate based on capacitance changes in the sensing mass while under acceleration or angular rate changes to piezoresistive MEMS that produce resistance changes in strain gauges, which form part of the core device, to piezoelectric systems that produce February/March 2024 | Uncrewed Systems Technology FOG-on-chip technology is enabling FOG IMUs to be produced slightly smaller at much lower cost in higher volumes (Image courtesy of Advanced Navigation)
RkJQdWJsaXNoZXIy MjI2Mzk4