Algorithm Enhancements for the SS-411 Digital Sun Sensor

This paper discusses the evolution of the SS-411 series digital sun sensor. The earlier SS-256 and SS-330 models have proven themselves in orbit on low cost satellite missions. The SS-411 represents a further hardware revision with enhanced robustness and improved attitude estimation performance. To complement the latest hardware improvements, researchers in the Space Avionics and Instrumentation Laboratory (SAIL) at Ryerson University in Toronto have developed advanced signal processing routines compatible with the SS-411. These routines significantly improve the accuracy of the sensor's estimation without increasing manufacturing complexity. With this advanced processing, the sensor maintains a mean accuracy of 0.16° over the entire field of view. INTRODUCTION As the microsatellite industry matures, the competing demands for performance and simplicity create niches for new devices. The Sinclair Interplanetary SS-411 digital sun sensor (Figure 1) represents the latest revision in a series of small and capable microsatellite sensors. This study details the latest changes to the sensor design with particular focus on a set of new signal processing algorithms. The sensor employs narrow slits to create a series of bright peaks on a linear detector array. The accuracy in the output sun vector is directly related to the accuracy of the peak position estimate (PPE). Previous laboratory and simulation studies [1][2] have post-processed images from the SS-256 and SS-330 model sensors to localize the peaks to a small fraction of a pixel, and thus produce very precise sun-estimates. This study details the adaptation of these algorithms from off-line use to embedded implementation on the SS-411's internal 8-bit processor. The SS-411 design. Analog two-axis sun sensors are commonly made from four-quadrant photodiodes. The sun angles can be determined by a simple function of the four output currents. However, these sensors are vulnerable to incoming light from non-solar sources: the earth, the moon, or glint from spacecraft appendages. This unwanted light cannot be distinguished from sunlight and so it decreases the sensor's operational accuracy. Digital sun sensors use an array of many photosensitive pixels. By increasing the number of scalar measurements it becomes possible to resolve multiple light sources and to discard those that are not the sun. Many two-axis digital sun sensors are available. Some use a two-dimensional detector array such as a CCD or active pixel sensor [3][4]. Others use two orthogonal linear 70° ± Figure 1: SS-411 digital sun sensor with penny for scale. Enright 1 21st Annual AIAA/USU Conference on Small Satellites detector arrays[5]. The Sinclair Interplanetary sensors are unusual in that they use only a single linear detector but are able to determine the sun location in two axes. Incoming sunlight passes through a metal mask in which four thin slits have been cut (Figure 2). An image of this pattern is projected onto a linear detector array, creating a one-dimensional brightness profile containing four strong peaks. One pair of peaks is close together, while the other pair has a wider spacing (Figure 3). This allows a computer algorithm contained within the sensor to distinguish between them and to locate the twodimensional offset between the mask and the detector. A second algorithm models the sensor geometry including refractive effects and produces the final output: a vector from the sensor to the sun. Current Design The sensor is built from high-end COTS components. At its heart is a 16mm long linear detector array with 256 pixels. The detector has a fast electronic shutter, but even at the maximum speed the pixels can saturate in direct sunlight. An optical filter with a transmission of 10% or less attenuates the incoming light and allows reasonable shutter speeds. All of the optical parts use the industry standard 25mm diameter circular form factor. A concern early in concept development was how to vent the sensor without letting stray light strike the detector. Light-tight venting filters are available, but present mechanical integration concerns. Instead, the problem of venting was avoided by designing the device so that it contains no air. The optical elements are bonded directly to each other and to the detector using transparent adhesive. The rest of the unused internal volume is completely filled with opaque potting compound. The detector is mounted to a printed circuit board that also contains a microcontroller (Figure 4). An on-board ADC digitizes the detector output and the CPU executes the image processing algorithms. A serial bus connects the microcontroller to the satellite's ACS computer. Various different bus technologies are supported: asynchronous, SPI, I2C, and CANbus. The sensor produces digital sun position vectors, as well as exposure and temperature telemetry. It accepts synchronization and reconfiguration commands, or even software patches while on-orbit. Heritage The first Sinclair Interplanetary sun sensor design was designated the SS-256. It was followed in succeeding years by the SS-330, and recently the SS-411 (Figure 5). Figure 2: The SS-411 Mask, backlit to show slits. The mask is 25mm in diameter. Figure 3: Sample Images from the SS-411. The dotted peaks have been labelled with their generating slits. Figure 4: The SS-411 Electronics. The detector and microcontroller can be seen in the left figure. x y Lx2 Ly1 Ly2 Lx1 50 100 150 200 250 0 50