标题: STM32代码MPU6050的姿态结算,摘用cleanfight的姿态解算代码 [打印本页]

作者: chrain5    时间: 2018-1-20 10:53
标题: STM32代码MPU6050的姿态结算,摘用cleanfight的姿态解算代码
STM32代码MPU6050的姿态结算,摘用cleanfight的姿态解算代码,,,效果可以

单片机源程序如下:
  1. #include "imu.h"
  2. #include "imu.h"
  3. #include  "math.h"
  4. #include "flycontro.h"

  5. #define SPIN_RATE_LIMIT 20

  6. #ifndef sq
  7. #define sq(x) ((x)*(x))
  8. #endif
  9. #define sin_approx(x)   sinf(x)
  10. #define cos_approx(x)   cosf(x)
  11. #define atan2_approx(y,x)   atan2f(y,x)
  12. #define acos_approx(x)      acosf(x)
  13. #define tan_approx(x)       tanf(x)
  14. // Use floating point M_PI instead explicitly.
  15. #define M_PIf       3.14159265358979323846f




  16. Attitudat_Struct Accel,Gyro,Angle; //MPU姿态数据
  17. Attitudat_Struct Gyro_Offset;  //陀螺仪零漂数据
  18. short Accel_X,Accel_Y,Accel_Z,Gyro_X,Gyro_Y,Gyro_Z;
  19. void MPU_GetAngleDat(void)
  20. {
  21.   int  Accel_Xtemp,Accel_Ytemp,Accel_Ztemp,Accel_Xtempout,Accel_Ytempout,Accel_Ztempout;
  22. /// static float Yawerr=0,lastYawerr=0;
  23.   MPU_Get_Accelerometer(&Accel_X,&Accel_Y,&Accel_Z);        //得到加速度传感器数据
  24.   MPU_Get_Gyroscope(&Gyro_X,&Gyro_Y,&Gyro_Z);        //得到陀螺仪数据
  25.   
  26.         Accel_Xtemp  = Accel_X;
  27.         Accel_Ytemp  = Accel_Y;
  28.         Accel_Ztemp  = Accel_Z;

  29.   Prepare_Data(&Accel_Xtemp,&Accel_Ytemp,&Accel_Ztemp,&Accel_Xtempout ,&Accel_Ytempout,&Accel_Ztempout);

  30.         Accel.X = Accel_Xtempout;     ////8192
  31.         Accel.Y = Accel_Ytempout;
  32.   Accel.Z = Accel_Ztempout;        
  33.        
  34.         Gyro.X = (Gyro_X - Gyro_Offset.X) ;
  35.         Gyro.Y = (Gyro_Y - Gyro_Offset.Y) ;
  36.         Gyro.Z = (Gyro_Z - Gyro_Offset.Z) ;

  37. }


  38. u8 Gyro_calibration(void)
  39. {
  40. //u16 i;
  41. //short  Gyro_X,Gyro_Y,Gyro_Z;
  42. static long int GyroofsetX=0,GyroofsetY=0,GyroofsetZ=0;
  43. static u16 i= 0;
  44. //        for(i=0;i<1000;i++)
  45. //         {
  46. // MPU_Get_Gyroscope(&Gyro_X,&Gyro_Y,&Gyro_Z);        //得到陀螺仪数据
  47.     GyroofsetX+=Gyro_X;
  48.                 GyroofsetY+=Gyro_Y;
  49.                 GyroofsetZ+=Gyro_Z;
  50.           i++;
  51. //                delay_ms(5);
  52. //   }
  53. //       
  54.         if(i==1000)
  55.         {
  56.    i=0;
  57.    Gyro_Offset.X = GyroofsetX/1000;
  58.    Gyro_Offset.Y = GyroofsetY/1000;
  59.    Gyro_Offset.Z = GyroofsetZ/1000;
  60.          GyroofsetX=GyroofsetY=GyroofsetZ=0;
  61.                 return 1;
  62.         }

  63.         return 0;
  64. }


  65. void Prepare_Data(int *accx,int *accy,int *accz,int *accoutx,int *accouty,int *accoutz)
  66. {  
  67.         static uint8_t         filter_cnt=0;
  68.         static int16_t        ACC_X_BUF[FILTER_NUM],ACC_Y_BUF[FILTER_NUM],ACC_Z_BUF[FILTER_NUM];
  69.         int32_t temp1=0,temp2=0,temp3=0;
  70.         uint8_t i;       
  71.         ACC_X_BUF[filter_cnt] = *accx;
  72.         ACC_Y_BUF[filter_cnt] = *accy;
  73.         ACC_Z_BUF[filter_cnt] = *accz;
  74.         for(i=0;i<FILTER_NUM;i++)//滑动平滑滤波
  75.         {
  76.                 temp1 += ACC_X_BUF[i];
  77.                 temp2 += ACC_Y_BUF[i];
  78.                 temp3 += ACC_Z_BUF[i];
  79.         }
  80.         *accoutx = temp1 / FILTER_NUM;
  81.         *accouty = temp2 / FILTER_NUM;
  82.         *accoutz = temp3 / FILTER_NUM;
  83.         filter_cnt++;
  84.         if(filter_cnt==FILTER_NUM)        filter_cnt=0;

  85. }












  86. float q0 = 1.0f, q1 = 0.0f, q2 = 0.0f, q3 = 0.0f;    // quaternion of sensor frame relative to earth frame
  87. static float rMat[3][3];

  88. float dcmKiGain = 0.005;
  89. float dcmKpGain = 2.0;

  90. static float invSqrt(float x)
  91. {
  92.     return 1.0f / sqrtf(x);
  93. }

  94. void imuMahonyAHRSupdate(float dt, float gx, float gy, float gz,
  95.                                                                                                         bool useAcc, float ax, float ay, float az,
  96.                                                                                                         bool useMag, float mx, float my, float mz,
  97.                                                                                                         bool useYaw, float yawError)
  98. {
  99.     static float integralFBx = 0.0f,  integralFBy = 0.0f, integralFBz = 0.0f;    // integral error terms scaled by Ki
  100.     float recipNorm;
  101.     float hx, hy, bx;
  102.     float ex = 0, ey = 0, ez = 0;
  103.     float qa, qb, qc;
  104.    
  105.     // Calculate general spin rate (rad/s)
  106.     float spin_rate = sqrtf(sq(gx) + sq(gy) + sq(gz));
  107.           float ez_ef;
  108.     // Use raw heading error (from GPS or whatever else)
  109.     if (useYaw) {
  110.         while (yawError >  M_PIf) yawError -= (2.0f * M_PIf);
  111.         while (yawError < -M_PIf) yawError += (2.0f * M_PIf);

  112.         ez += sin_approx(yawError / 2.0f);
  113.     }

  114.     // Use measured magnetic field vector  
  115.     recipNorm = sq(mx) + sq(my) + sq(mz);
  116.     if (useMag && recipNorm > 0.01f)
  117.                         {
  118.         // Normalise magnetometer measurement ?????????
  119.         recipNorm = invSqrt(recipNorm);
  120.         mx *= recipNorm;
  121.         my *= recipNorm;
  122.         mz *= recipNorm;
  123.                                 // For magnetometer correction we make an assumption that magnetic field is perpendicular to gravity (ignore Z-component in EF).
  124.                                 // This way magnetic field will only affect heading and wont mess roll/pitch angles

  125.                                 // (hx; hy; 0) - measured mag field vector in EF (assuming Z-component is zero)
  126.                                 // (bx; 0; 0) - reference mag field vector heading due North in EF (assuming Z-component is zero)
  127.         hx = rMat[0][0] * mx + rMat[0][1] * my + rMat[0][2] * mz;
  128.         hy = rMat[1][0] * mx + rMat[1][1] * my + rMat[1][2] * mz;
  129.         bx = sqrtf(hx * hx + hy * hy);

  130.         // magnetometer error is cross product between estimated magnetic north and measured magnetic north (calculated in EF)
  131.         ez_ef = -(hy * bx);

  132.         // Rotate mag error vector back to BF and accumulate
  133.         ex -= rMat[2][0] * ez_ef;
  134.         ey -= rMat[2][1] * ez_ef;
  135.         ez += rMat[2][2] * ez_ef;
  136.     }

  137.     // Use measured acceleration vector ?????????
  138.     recipNorm = sq(ax) + sq(ay) + sq(az);
  139.     if (useAcc && recipNorm > 0.01f)
  140.                         {
  141.         // Normalise accelerometer measurement ?????????
  142.         recipNorm = invSqrt(recipNorm);
  143.         ax *= recipNorm;
  144.         ay *= recipNorm;
  145.         az *= recipNorm;

  146.         // Error is sum of cross product between estimated direction and measured direction of gravity
  147.         ex += (ay * rMat[2][2] - az * rMat[2][1]);
  148.         ey += (az * rMat[2][0] - ax * rMat[2][2]);
  149.         ez += (ax * rMat[2][1] - ay * rMat[2][0]);
  150.     }

  151.     // Compute and apply integral feedback if enabled
  152.     if(dcmKiGain > 0.0f) {   //imuRuntimeConfig->dcm_ki
  153.         // Stop integrating if spinning beyond the certain limit????,??????????
  154.         if (spin_rate < DEGREES_TO_RADIANS(SPIN_RATE_LIMIT)) {
  155.        //     float dcmKiGain = dcm_ki;
  156.             integralFBx += dcmKiGain * ex * dt;    // integral error scaled by Ki
  157.             integralFBy += dcmKiGain * ey * dt;
  158.             integralFBz += dcmKiGain * ez * dt;
  159.         }
  160.     }
  161.     else {
  162.         integralFBx = 0.0f;    // prevent integral windup
  163.         integralFBy = 0.0f;
  164.         integralFBz = 0.0f;
  165.         }

  166.     // Calculate kP gain. If we are acquiring initial attitude (not armed and within 20 sec from powerup) scale the kP to converge faster
  167.     //   dcmKpGain =  dcm_kp ;//* imuGetPGainScaleFactor();

  168.     // Apply proportional and integral feedback??????
  169.     gx += dcmKpGain * ex + integralFBx;
  170.     gy += dcmKpGain * ey + integralFBy;
  171.     gz += dcmKpGain * ez + integralFBz;

  172.     // Integrate rate of change of quaternion??????
  173.     gx *= (0.5f * dt);
  174.     gy *= (0.5f * dt);
  175.     gz *= (0.5f * dt);

  176.     qa = q0;
  177.     qb = q1;
  178.     qc = q2;
  179.     q0 += (-qb * gx - qc * gy - q3 * gz);
  180.     q1 += (qa * gx + qc * gz - q3 * gy);
  181.     q2 += (qa * gy - qb * gz + q3 * gx);
  182.     q3 += (qa * gz + qb * gy - qc * gx);
  183.     // Normalise quaternion
  184.     recipNorm = invSqrt(sq(q0) + sq(q1) + sq(q2) + sq(q3));
  185.     q0 *= recipNorm;
  186.     q1 *= recipNorm;
  187.     q2 *= recipNorm;
  188.     q3 *= recipNorm;

  189.     // Pre-compute rotation matrix from quaternion
  190.     imuComputeRotationMatrix();
  191. #if 0         
  192. //                 Angle.Y = asin(-2 * q1 * q3  + 2 * q0* q2)* 57.3 ; // pitch  
  193. //          Angle.X = atan2( 2 * q2 * q3 + 2 * q0 * q1, -2 * q1 * q1 - 2 * q2* q2 + 1)* 57.3; // roll
  194. //          Angle.Z = atan2(2 * q1 * q2  + 2 * q0 * q3, -2 * q2*q2 - 2 * q3* q3 + 1)* 57.3; // yaw
  195. #else         
  196.          imuUpdateEulerAngles(&Angle.X,&Angle.Y,&Angle.Z);
  197. #endif

  198. }

  199. void imuComputeRotationMatrix(void)
  200. {
  201.     float q1q1 = sq(q1);
  202.     float q2q2 = sq(q2);
  203.     float q3q3 = sq(q3);

  204.     float q0q1 = q0 * q1;
  205.     float q0q2 = q0 * q2;
  206. ……………………

  207. …………限于本文篇幅 余下代码请从51黑下载附件…………
复制代码

所有资料51hei提供下载:
imu.zip (3.66 KB, 下载次数: 176)



作者: home8976    时间: 2018-8-22 14:33
过来看看
作者: lyjpla    时间: 2019-10-4 21:45
谢谢!非常有参考价值

作者: 枫叶-嵌入式    时间: 2019-10-6 11:25
这个资料很不错,正需要
作者: 18519187648    时间: 2022-9-17 17:09
你这怎么用啊,一些参数都没写说明
作者: lm7298    时间: 2022-9-19 15:35
资料很不错,正好需要,感谢




欢迎光临 (http://www.51hei.com/bbs/) Powered by Discuz! X3.1