Zhi-yuan Wang’s research while affiliated with Zhejiang University and other places

An artificial intelligence enhanced star identification algorithm is proposed for star trackers in lost-in-space mode. A convolutional neural network model based on Vgg16 is used in the artificial intelligence algorithm to classify star images. The training dataset is constructed to achieve the networks’ optimal performance. Simulation results show that the proposed algorithm is highly robust to many kinds of noise, including position noise, magnitude noise, false stars, and the tracker’s angular velocity. With a deep convolutional neural network, the identification accuracy is maintained at 96% despite noise and interruptions, which is a significant improvement to traditional pyramid and grid algorithms.

Infrared Earth sensors are widely used in attitude-determination and control systems of satellites. The main deficiency of static infrared Earth sensors is the requirement of a small field of view (FOV). A typical FOV for a static infrared Earth sensor is about 20° to 30°, which may not be sufficient for low-Earth-orbiting micro-satellites. A novel compact infrared Earth sensor with an FOV of nearly 180° is developed here. The Earth sensor comprises a panoramic annular lens (PAL) and an off-the-shelf camera with an uncooled complementary-metal-oxide-semiconductor (CMOS) infrared sensor. PAL is used to augment FOV so as to obtain a complete infrared image of the Earth from low-Earth-orbit. An algorithm is developed to compensate for the distortion caused by PAL and to calculate the vector of the Earth. The new infrared Earth sensor is compact with low power consumption and high precision. Simulated images and on-orbit infrared images obtained via the micro-satellite ZDPS-2 are used to assess the performance of the new infrared Earth sensor. Experiments show that the accuracy of the Earth sensor is about 0.032°.