Four-quadrant analog sun sensor

A  high-precision  error  compensation  method  was  proposed  for  the  four-quadrant  analog  sun  sensor  in order  to  improve  the  accuracy  of  the  micro-nano  satellite  attitude  determination  system.  A  completely  automated calibration process was designed. The process of measuring the photogenerated current of the four-quadrant silicon photocell was analyzed, and the projection relationship of the incident sunlight was modeled to extract the main error sources.  All  aspects  of  the  process  were  considered,  and  the  current  measurement  errors  in  each  channel  were separately corrected. The errors caused by machining and installation errors and the errors caused by neglecting the thickness of the optical mask were compensated, which formed a complete compensation method. The experimental results  showed  that  machining  and  installation  errors  were  the  main  error  sources,  and  the  influence  of  the  error caused  by  neglecting  the  thickness  of  the  optical  mask  was  slightly  greater  than  that  of  the  current  measurement error. The average accuracy before compensation was 3.072° (1σ) within ±40° of incident angle, and the average accuracy  was  0.177°  (1σ)  after  compensation.  The  accuracy  after  calibration  of  the  existing  method  was  0.5°(1σ).  The  calibration  accuracy  of  the  proposed  method  was  improved  by  about  three  times  compared  with  the existing  method.  A  set  of  automated  calibration  test  method  for  the  whole  process  was  designed  aiming  at  the production process of calibration test, which obviously improved the calibration efficiency and was suitable for mass application.

With the rapid development of microsatellite technology, solar sensors, as an important component of satellite attitude control systems, have been widely used for measuring the sun azimuth angle of microsatellites. The joint attitude determination system composed of solar sensors, magnetometers, and gyroscopes can meet the requirements of high-precision attitude determination in microsatellites According to the types of detectors, solar sensors are mainly divided into digital image sensors and analog photoelectric detectors In response to the requirements of small size, low cost, and low power consumption for micro/nano satellites, many scholars are committed to developing new solar sensors suitable for micro/nano satellites Antonello et al. developed a low-cost and small volume solar sensor based on precision machined pinholes and complementary metal oxide semiconductor (CMOS) image sensors, achieving 66.2 ° × 51.1 ° field of view and 0.03 ° (1 σ） The accuracy of, where σ  Is the standard deviation Wei et al. implemented profile detection and detector multiplexing technology based on image sensors, achieving a 100 ° field of view and 0.01 ° (1 σ） Accuracy Abhilash et al. achieved a ± 62 ° field of view and 0.5 ° (1 σ） Compared with digital sun sensors, analog sun sensors generally have lower accuracy, but their structure and principle are simpler, with advantages in volume, weight, power consumption, and cost Due to the balance of onboard resources, small satellites tend to use simulated solar sensors.

Four quadrant analog solar sensors have extremely low power consumption, minimal volume, low complexity, low cost, and high reliability. Multiple domestic and foreign units have developed various models of solar sensors based on this principle, which have been applied to multiple satellites Due to its excellent resistance to high and low temperatures, as well as space irradiation, the detector is particularly suitable for the use of a split design in the solar panel’s sun pointing and tracking process With the increasing complexity of space missions, it is required to improve the accuracy of four quadrant simulated solar sensors as much as possible. Without the use of high power consumption, large volume, and high-precision devices, it is necessary to achieve medium precision attitude determination tasks for micro and nano satellites, meeting the high-precision requirements of solar panels for solar orientation and tracking Practice has proven that accurate modeling and corresponding error compensation and precise calibration of solar sensors are an inexpensive and effective method to improve accuracy Fan et al. analyzed and simulated the error factors of encoded solar sensors, established an error compensation model that includes structural errors and fine code algorithm errors, and proposed corresponding model parameter calibration methods for precise modeling of internal and external parameters The experimental results show that the accuracy has been improved by an average of three times Yousefian et al. analyzed a solar sensor array composed of six photodiodes and divided the system error into manufacturing error, ambient light scattering, and photodiode model error. They provided corresponding calibration programs and ambient light models, and reduced the sensor accuracy from 2.63 ° (1 σ） Reduce to 0.83 ° (1 σ）. Many scholars have conducted corresponding analysis and modeling work on four quadrant simulated solar sensors Feng Miao et al. directly fitted the input and measured values of the solar vector of the four quadrant analog solar sensor by establishing a 5-degree surface fitting equation, achieving a field of view of ± 60 ° and a 0.5 ° (1 σ） The accuracy is not ideal due to the introduction of too many parameters and the lack of practical meaning in the model Wang Chunyu et al. proposed a quantitative analysis method for the main error sources in the assembly process of the four quadrant simulated solar sensor measurement error caused by assembly process deviation. However, they did not analyze the mechanical processing error and did not provide a complete compensation method and process Xu Xiaodan et al. analyzed the accuracy of photogenerated current collection in each quadrant of the four quadrant simulated solar sensor silicon photocell, analyzed the possible errors introduced in each link of the measurement link, calibrated and corrected the comprehensive errors of each link, and effectively improved the measurement accuracy. However, experimental measurement methods were used in each link of the testing process, which required high equipment accuracy and complicated testing process

1. Structure and working principle

The bottom of the four quadrant analog solar sensor (referred to as the analog solar sensor) is a square silicon photocell with a side length of 2l, which is etched into four silicon photocells (Q1, Q2, Q3, and Q4), with a center point of O, as shown in Figure 1 (a). The four quadrant silicon photocell forms an Oxyz coordinate system A light shield with a square light introducer is installed parallel above the four quadrant silicon photocell. The center O1 of the light introducer is located on the axis Oz, and the four edges are parallel to the axes Ox and Oy. A, B, C, and D are the four corners of the light introducer, located in the first, second, third, and fourth quadrants of the Oxy plane

2. Error analysis

Exploring the current measurement process of simulated solar sensors and the incident projection process of sunlight, analyzing the impact of current measurement errors, mechanical processing and installation errors, and ignoring the thickness of the light shield on errors

3. Error compensation method

Error compensation is based on an accurate model simulating a solar sensor, considering the errors present in each link and providing a vector output formula with undetermined parameters According to the calibration experiment, adjust the parameters to effectively improve the accuracy of the sensor

4. Calibration process design

The above error compensation method can effectively improve the accuracy of the sensor, but it relies on a large amount of test data and precise experimental environment construction methods This article uses a model parameter calibration method that accurately models internal and external parameters for calibration. Ki, bi, dx0, dy0, L1, L2, h, m are selected as internal parameters (i=1, 2, 3, 4), and the angle between the initial position of the solar simulator and the turntable coordinate system, as well as the installation matrix of the solar sensor in the turntable coordinate system, are selected as external parameters. The four quadrant photo generated current sampling values output by the solar sensor and the turntable rotation angle are collected as experimental data All parameters are solved using the nonlinear least squares method, and the initial values and empirical ranges of each parameter are given based on the actual parameters, avoiding the convergence problem caused by optimizing too many parameters simultaneously

Among the three error sources used to simulate solar sensors, the order of impact on accuracy is in descending order: machining and installation errors, errors caused by ignoring the thickness of the light shield, and current measurement errors, with machining and installation errors being the main sources of error By using this error compensation method, the accuracy of simulated solar sensors can be stably and reliably improved

Before applying the above error compensation methods, the error source should be controlled to minimize the impact of the error The error compensation method improves the accuracy of the sensor as much as possible through experimental calibration methods, but cannot solve the problem of field of view angle degradation caused by the aforementioned errors The source of the upper limit of accuracy for the four quadrant analog solar sensor is the noise in the photogenerated current measurement circuit of the four quadrant silicon photocell. This noise can be suppressed using more advanced circuit design methods or processed using subsequent filtering algorithms In the future, we should optimize the structure and circuit design scheme based on high-precision models, reduce machining and installation errors, and design solar sensors with higher accuracy