【摘要】 提出了一套简便易行的空间光通信精跟踪演示系统方案,详细介绍了精跟踪演示系统的组成。信标光斑定位和跟踪算法决定了跟踪带宽,信标光斑定位的传统方法存在系统误差和随机误差,采用像元细分对误差进行校正;跟踪算法采用改进神经网络算法,利用神经网络的自学习和自适应能力,在线调整网络加权值,增强了系统的实时跟踪性能。最后分析了像元细分对定位精度的改善,比较了不同定位算法的跟踪性能,改进的神经网络算法提高了精跟踪系统的鲁棒性。采用400Hz的CCD,针对不同频率的信标抖动进行跟踪补偿实验,实验结果表明,50Hz的信标光抖动范围压缩了25.7%。更多还原

【Abstract】 A simple demonstration system for fine tracking is put forward,and the component is presented in detail.The centroid algorithm is the traditional method for beacon,and it has a systematic error and a random one.So it is obligatory to adopt the pixel subtraction.With the PID-like neural network algorithm,the weights of neural network and the parameters could be adjusted to reduce beacon vibration by the function of self-learning and adaptability in real time.The experimental results show the pixel subtraction improves the precision of centroid algorithm,and the PID-like neural network algorithm is robust.The fine tracking system can track 50 Hz beacon vibration with CCD of 400 Hz,and the range of beacon vibration reduces 25.7%,which is beneficial for the system to apply.更多还原