Search

As an electroluminescentdevice, the coupling relationship between light-emitting diode (LED) input currents, optical power, and LED junction temperature is a complicated multiphysics process. In this paper, a simplified LED photoelectron-thermal (PET) model by artificial neural network (ANN), which can translatemultiphysics field issue into a singlephysics field problem, ismentioned to study the coupling relationship. In the first, an LED lumens, optical power, and electric power at different temperatures are monitored in a temperature controlling integrating sphere. Then, a backpropagation (BP) ANN is trained by these data to construct an LED Photo-Electron-Thermal (PET) relationship. In addition, LED luminaire thermal analyzing is performed using a finite-element method on the outputs of the BP ANN. Finally, the advantage of this method in terms of saving computing resources and computing time is analyzed by comparing the degrees of freedom in different models. The result shows that at least 6.7 times computing resource are saved by this method, which will reduce the LED thermal management system analyzing time greatly. Finally, the application and extension of this model are discussed.

When detecting targets using active sonar in littoral environment, the echoes are always blended with high-level discrete clutters which lead to false alarms. Sonar and environment parameters will both make an impact on the quantity of clutters. Traditionally, it is difficult to quantify clutters what is essential for sonar design, actually. In this study, the reverberation statistical models are proposed to quantify clutters, such as K distribution and mixture distribution. We are trying to use particular parameters of these distributions to quantify clutters. The theory is declared herein and experimentally proved by lake trial. The method is useful for engineers to estimate the severity of clutters and adjust sonar parameters.

Traditional temperature error model of ring laser gyroscope (RLG) based on neural network faces great problems in the repeatability performance under different temperature conditions. To reduce the temperature error and improve the generalization ability of traditional neural network, a novel error model based on radial basis function neural network optimized by particle swarm optimization (PSO) and regularization approach was proposed. The temperature error is analyzed and preprocessed. The PSO method is used to search the optimal configuration of the network, and regularization method is used as the evaluation criterion to further optimize the coefficients of the network. The experimental results show that the proposed method can improve the precision and environmental adaptability of RLG, which had been practically applied in RLG position and orientation system.

A high rate sampling and digital filter can demodulate and remove the dither noise of RLG caused by mechanical dither. However, the time delay caused by antivibration system and digital filter severely reduce the measurement precision, and become a key problem. To solve the problem, a time delay modeling and compensation method of dithered RLG POS was proposed. An optimized multi-degree uncoupled antivibration system was designed. The dynamic response time delay model based on antivibrator characteristics was established. A linear FIR digital filter was designed to remove dither signal of RLG, the filtering time delay model was accurately established. The RLG POS can accurately measure space–time information by time delay compensation. The experimental results show that the attitude errors are reduced to 0.0028°, 0.0044° and 0.0021° from 0.0033°, 0.0058° and 0.0104°, using proposed time delay compensation method. It contributes to realize high quality image of optical camera without ground control points.

The Position and Orientation System (POS) is an integrated motion measurement system of Inertial Navigation System (INS)/Global Positioning System (GPS) for airborne remote sensing motion compensation. Multi-level lever-arm exists among each subsystem and each has its individual characteristics. The error model of individual lever arm was built up. In the INS subsystem, the accelerometer's output error caused by the size-effect error namely the displacement of the sensing center of accelerometer (CA) from the center of inertial sensing assembly (CISA) is of 100 µg level, which relates to the fixing orientation error of inertial measurement unit (IMU). The accelerometer's output error caused by the outer lever arm namely the displacement of CISA from mass center of vehicle (MCV) is of mg level, which is not determined for that the MCV is changing in the mission of flying. Between IMU and GPS, The position and velocity error which was caused by first level lever arm namely the displacement of ISA from the center of GPS was taken as the measurement for extended Kalman filter (EKF). Besides, between sensing load and IMU, the second level lever arm namely the displacement of ISA from the center of remote sensing load is determined, the motion error can be eliminated provide the measurement of POS is accurate. The simulation and flying experiment shows that the inner and outer level arm will introduce the ill acceleration which will produce the position error of 10−3 and 10−4 level in meters provide the GPS signal is normal, the first level and second level lever arm can produce position error in meter's level, the position error and velocity error can be taken as the measurement in EKF to realize integration of SINS/GPS.

Motion compensation technology based on Ring Laser Gyroscope (RLG) Position and Orientation System (POS) enormously improves the imaging quality and operation efficiency of airborne remote sensing systems. However, bias error of RLG, aroused by temperature variation, severely deteriorates the measurement precision of POS. To solve this problem, several error modeling and compensation techniques have been devised, including Linear Least Squares Fitting (LLSF), RBF Neural Network (RBF NN) and Least Square Support Vector Machine (LS SVM). Theoretical basis of these methods are introduced. Comparison among them with subjects on model complexity, computing speed, precision and generalization performance is drawn, and conclusions are verified via temperature circling experiment of real RLG. Approach based on LLSF acquires the advantages of high computing speed and low hardware resource occupancy, while superiority on precision and generalization performance of LS SVM is obvious. According to the hostile working environment and high precision requirement of POS, methods based on LLSF and LS SVM are adopted to work under online and offline modes of POS, which meet the demands of computing speed and compensation precision respectively. Airborne flight experiment results demonstrate that, six groups' average online inertial navigation error of RLG POS after 4 hours' flight was 9.5775 nmiles, while the average offline inertial navigation error was 4.0661 nmiles. Such result satisfied the application requirement of high resolution InSAR.