Unmanned Systems Technology 015 | Martin UAV V-Bat | William Sachiti | Sonar Systems | USVs | Desert Aircraft DA150 EFI | SeaCat AUV/ROV | Gimbals

Sofradir’s Sirius is an example of a cooled long-wave quantum well IR photodetector, which requires cooling down to 73 K and so needs a relatively large Stirling cycle cooler (Courtesy of Sofradir) 74 Core sensor evolution Developments in core sensor technologies have enabled today’s small payloads to offer performance comparable with those of larger ones of a few years ago. Detectors are improving constantly in respect of their reliability, sensitivity, resolution and technical parameters such as the size of each pixel. All these factors affect the design of the overall system. One systems engineer points out that high-resolution focal plane arrays, and new video formats such as HD, enable small payloads to fly closer to the ground and use relatively short focal lengths to achieve a ‘pixel-to-the-ground’ ratio that is as good as or even better than that offered by larger payloads using long lenses from higher altitudes. That applies to the visible as well as the IR channels. However, flying close to the ground requires a sensor capable of generating high frame rates to avoid blurring of the image. For example, for a thermal camera with a 7 Hz frame rate installed in a turret on a small fixed-wing UAV flying typically at 20 m/s at less than 400 ft, the imagery would be very blurred. A frame rate of 25 Hz or more would be needed to solve this problem. Algorithmic influence Digital image processing is an area in which companies can distinguish their products from those of the competition, as it helps extract more information from the signal coming out of the detectors. Many of the leading payload houses do this themselves, and create algorithms that are their intellectual property. Several of these functions not only enhance and exploit the image, they also feed back into the control system, and ultimately help determine the instructions sent to the servo motors. While target tracking is the most common task for processing, stabilisation in translation and roll also add significant value, an engineer tells us. August/September 2017 | Unmanned Systems Technology on their platforms, as well as high- quality low-friction bearings and careful placement of cable runs, can reduce these problems, but residual effects remain to be dealt with by active control systems. While a gimbal permits movement about an axis, it can neither sense movement nor control it. That requires sensing systems such as IMUs and computer-controlled electric servo motors. These motors can either drive the payload directly or via a gear set if the required torque is greater than the motor can generate directly. However, the use of geared motors increases coupling forces significantly and also limits the stabilisation system’s bandwidth by an order of magnitude. Bandwidth in this sense means the vibration frequency range, expressed in Hertz, over which the system is effective. Direct-drive motors provide more bandwidth, but generate high-frequency torque forces on the gimbal and turret supporting structure, which therefore has to be as stiff and strong as possible within the size and weight budget to make the most of that bandwidth, as any unwanted flexibility will allow uncontrolled movement. All of the high-performance two-axis turrets use direct-driven motors these days, and indirectly driven designs are being eliminated very quickly, an engineer tells us. Motion sensing systems based on smaller and more affordable MEMS gyros with low noise have played an important role in the development of gimballed payloads, he adds. A further issue is aligning the sensors’ optical axes with one another precisely – a process known as boresighting – so that as the operator switches from one sensor to another the crosshairs remain over exactly the same spot. This is particularly critical in target engagement, for obvious reasons. Again, that is not as easy as it might seem, as the sensors are in different places in the turret and therefore the light follows slightly different optical paths into them. Boresight alignment must be maintained throughout the turret’s field of regard and throughout the zoom range for each sensor individually, and among all of them collectively. For this reason gimbals must have high-quality bearings with as near zero play as possible, and be mounted on very stiff structures that remain dimensionally stable under all anticipated conditions. One reason that is so difficult is that UAV payloads are subject to very large temperature changes across different times of the day and between ground level and high altitude, and a gimbal’s structural materials and its optical and mechanical components have different thermal coefficients of expansion. To prevent that from causing misalignments between the sensors, for example, careful selection of materials and the design of mechanical components and electronic control loops are essential.