CAJ | 학술논문

介绍一种只需少量计算就可实现四轴无人机姿态的估计和控制的新颖方法。整个算法采用四元数表示，四元数运算基于惯性测量单元（IMU）的加速度和陀螺仪传感器。姿态估计使用四元数代替三角函数，允许用加速度数据分析推导和优化梯度下降算法来减少由陀螺仪传感器产生的误差。姿态控制将当前姿态与目标姿态的姿态误差反馈到PD控制器来实现想要的目标姿态。姿态估计中用到的梯度下降算法通过四元数对应的欧拉角的均方误差验证，静态均方误差小于0.945°。实际的飞行测试可以实现一个稳定的飞行，同时验证了姿态控制算法。
개소일충지수소량계산취가실현사축무인궤자태적고계화공제적신영방법。정개산법채용사원수표시，사원수운산기우관성측량단원（IMU）적가속도화타라의전감기。자태고계사용사원수대체삼각함수，윤허용가속도수거분석추도화우화제도하강산법래감소유타라의전감기산생적오차。자태공제장당전자태여목표자태적자태오차반궤도PD공제기래실현상요적목표자태。자태고계중용도적제도하강산법통과사원수대응적구랍각적균방오차험증，정태균방오차소우0.945°。실제적비행측시가이실현일개은정적비행，동시험증료자태공제산법。
A novel method with few calculations is presented to realize estimation and control of quad?axis UAV. The whole algorithm is expressed with quaternion,and the quaternion calculation is based on the acceleration of inertial measurement units (IMU)and gyroscope sensor. The quaternion is applied in attitude estimation instead of trigonometric function. Gradient descent algorithm is analyzed,deduced and optimized with acceleration data to reduce the error generated by gyroscope sensor. The error of current attitude and target attitude is fed back to PD controller with attitude control,to realize the expectant target attitude. The gradient descent algorithm applied in attitude estimation is verified with RMS error which is obtained from Euler angle corre?sponded with quaternion. The static RMS error is less than 0.945°. The stable flying is realized by actual flight test,meanwhile the attitude control algorithm is verified.