针对滚转高动态变化的旋转弹丸,考虑成本和体积限制,提出了一种仅依靠磁阻传感器的智能测姿方法。该方法根据旋转弹的飞行特点,通过飞行转台模拟弹丸飞行姿态,采集数据作为人工神经元网络的训练样本,训练后通过前馈网络在线辨识和估计俯仰角后计算滚转角。采用自主研制的低成本微惯性组件与商用惯导产品进行了对比实验,实验结果表明,所提方法可行,解算的姿态精度控制在1.5°以内。

According to spinning projectile of rolling in dynamic environment, an intelligent attitude determination method based on magnetometers data only was proposed with consideration of cost and size. Experimental signal data were used for training samples of an artificial neural network (ANN) depending on simulating flight attitude of turntable. Pitch angle estimation was completed by ANN. Based on it, roll angle was calculated eventually. Prototype of low cost micro inertial sensors and commercial inertial navigation system were used for implementing experiments. Experimental results demonstrate that the proposed method is feasible. The proposed intelligent attitude determination method by magnetometer data only can provide accuracy of less than 1.5 degree.