MPU6050 Stability

Do you know the MPU6050 the Gyro/Accelerometer module ..etc
Well I got an idea and some materials around and I want to build a "camera stabilizer"
or how else they call it,

I'm thinking of a simple one like that, using few servo motors..etc

The problem is that I got 5% experience with MPU6050 I only tested it as a module to see if it's working and done some small projects leds..etc

If I use the output data of the MPU can i use them directly to program my stabilizer, what about the vibrations that may cause the MPU having big errors on the output data(?). Is the any library or formulas or something that i can use making the stability at least 80% accurate and with less errors at vibrations(?)..etc

Edit:
Talking about stability and vibrations i want to mount that on my bike.

http://www.x-io.co.uk/res/doc/madgwick_internal_report.pdf

A Google search for Madgwick Arduino turned up a few references. I would expect some to include code already written.

What do you know about Quaternions? You need some fairly advanced maths to get this kind of stabiliser or "gimbal" to work.

Fortunately some better mathematicians than I have put together libraries and working Arduino programs. Some of it is even published on the internet. Use their code and ideas to make your own project great.

http://www.x-io.co.uk/res/doc/madgwick_internal_report.pdf

A Google search for Madgwick Arduino turned up a few references. I would expect some to include code already written.

What do you know about Quaternions? You need some fairly advanced maths to get this kind of stabiliser or "gimbal" to work.

Fortunately some better mathematicians than I have put together libraries and working Arduino programs. Some of it is even published on the internet. Use their code and ideas to make your own project great.

kinda a lot of work to build such small project, hmm
So the other day I've seen no one is posting here well before you posted of curse i done a research on google and found out about Kalman filter for MPU6050 I knew about it before but never used it, so i found a library that already had a example of MPU6050 + Kalman fiter, the library as known named kalman.h

Here is the actual code

MPU6050:

#include <Wire.h>
#include <Kalman.h> // Source: https://github.com/TKJElectronics/KalmanFilter

#define RESTRICT_PITCH // Comment out to restrict roll to ±90deg instead - please read: http://www.freescale.com/files/sensors/doc/app_note/AN3461.pdf

Kalman kalmanX; // Create the Kalman instances
Kalman kalmanY;

/* IMU Data */
double accX, accY, accZ;
double gyroX, gyroY, gyroZ;
int16_t tempRaw;

double gyroXangle, gyroYangle; // Angle calculate using the gyro only
double compAngleX, compAngleY; // Calculated angle using a complementary filter
double kalAngleX, kalAngleY; // Calculated angle using a Kalman filter

uint32_t timer;
uint8_t i2cData[14]; // Buffer for I2C data

// TODO: Make calibration routine

void setup() {
  Serial.begin(9600);
  Wire.begin();
#if ARDUINO >= 157
  Wire.setClock(400000UL); // Set I2C frequency to 400kHz
#else
  TWBR = ((F_CPU / 400000UL) - 16) / 2; // Set I2C frequency to 400kHz
#endif

  i2cData[0] = 7; // Set the sample rate to 1000Hz - 8kHz/(7+1) = 1000Hz
  i2cData[1] = 0x00; // Disable FSYNC and set 260 Hz Acc filtering, 256 Hz Gyro filtering, 8 KHz sampling
  i2cData[2] = 0x00; // Set Gyro Full Scale Range to ±250deg/s
  i2cData[3] = 0x00; // Set Accelerometer Full Scale Range to ±2g
  while (i2cWrite(0x19, i2cData, 4, false)); // Write to all four registers at once
  while (i2cWrite(0x6B, 0x01, true)); // PLL with X axis gyroscope reference and disable sleep mode

  while (i2cRead(0x75, i2cData, 1));
  if (i2cData[0] != 0x68) { // Read "WHO_AM_I" register
    Serial.print(F("Error reading sensor"));
    while (1);
  }

  delay(100); // Wait for sensor to stabilize

  /* Set kalman and gyro starting angle */
  while (i2cRead(0x3B, i2cData, 6));
  accX = (i2cData[0] << 8) | i2cData[1];
  accY = (i2cData[2] << 8) | i2cData[3];
  accZ = (i2cData[4] << 8) | i2cData[5];

  // Source: http://www.freescale.com/files/sensors/doc/app_note/AN3461.pdf eq. 25 and eq. 26
  // atan2 outputs the value of -π to π (radians) - see http://en.wikipedia.org/wiki/Atan2
  // It is then converted from radians to degrees
#ifdef RESTRICT_PITCH // Eq. 25 and 26
  double roll  = atan2(accY, accZ) * RAD_TO_DEG;
  double pitch = atan(-accX / sqrt(accY * accY + accZ * accZ)) * RAD_TO_DEG;
#else // Eq. 28 and 29
  double roll  = atan(accY / sqrt(accX * accX + accZ * accZ)) * RAD_TO_DEG;
  double pitch = atan2(-accX, accZ) * RAD_TO_DEG;
#endif

  kalmanX.setAngle(roll); // Set starting angle
  kalmanY.setAngle(pitch);
  gyroXangle = roll;
  gyroYangle = pitch;
  compAngleX = roll;
  compAngleY = pitch;

  timer = micros();
}

void loop() {
  /* Update all the values */
  while (i2cRead(0x3B, i2cData, 14));
  accX = ((i2cData[0] << 8) | i2cData[1]);
  accY = ((i2cData[2] << 8) | i2cData[3]);
  accZ = ((i2cData[4] << 8) | i2cData[5]);
  tempRaw = (i2cData[6] << 8) | i2cData[7];
  gyroX = (i2cData[8] << 8) | i2cData[9];
  gyroY = (i2cData[10] << 8) | i2cData[11];
  gyroZ = (i2cData[12] << 8) | i2cData[13];

  double dt = (double)(micros() - timer) / 1000000; // Calculate delta time
  timer = micros();

  // Source: http://www.freescale.com/files/sensors/doc/app_note/AN3461.pdf eq. 25 and eq. 26
  // atan2 outputs the value of -π to π (radians) - see http://en.wikipedia.org/wiki/Atan2
  // It is then converted from radians to degrees
#ifdef RESTRICT_PITCH // Eq. 25 and 26
  double roll  = atan2(accY, accZ) * RAD_TO_DEG;
  double pitch = atan(-accX / sqrt(accY * accY + accZ * accZ)) * RAD_TO_DEG;
#else // Eq. 28 and 29
  double roll  = atan(accY / sqrt(accX * accX + accZ * accZ)) * RAD_TO_DEG;
  double pitch = atan2(-accX, accZ) * RAD_TO_DEG;
#endif

  double gyroXrate = gyroX / 131.0; // Convert to deg/s
  double gyroYrate = gyroY / 131.0; // Convert to deg/s

#ifdef RESTRICT_PITCH
  // This fixes the transition problem when the accelerometer angle jumps between -180 and 180 degrees
  if ((roll < -90 && kalAngleX > 90) || (roll > 90 && kalAngleX < -90)) {
    kalmanX.setAngle(roll);
    compAngleX = roll;
    kalAngleX = roll;
    gyroXangle = roll;
  } else
    kalAngleX = kalmanX.getAngle(roll, gyroXrate, dt); // Calculate the angle using a Kalman filter

  if (abs(kalAngleX) > 90)
    gyroYrate = -gyroYrate; // Invert rate, so it fits the restriced accelerometer reading
  kalAngleY = kalmanY.getAngle(pitch, gyroYrate, dt);
#else
  // This fixes the transition problem when the accelerometer angle jumps between -180 and 180 degrees
  if ((pitch < -90 && kalAngleY > 90) || (pitch > 90 && kalAngleY < -90)) {
    kalmanY.setAngle(pitch);
    compAngleY = pitch;
    kalAngleY = pitch;
    gyroYangle = pitch;
  } else
    kalAngleY = kalmanY.getAngle(pitch, gyroYrate, dt); // Calculate the angle using a Kalman filter

  if (abs(kalAngleY) > 90)
    gyroXrate = -gyroXrate; // Invert rate, so it fits the restriced accelerometer reading
  kalAngleX = kalmanX.getAngle(roll, gyroXrate, dt); // Calculate the angle using a Kalman filter
#endif

  gyroXangle += gyroXrate * dt; // Calculate gyro angle without any filter
  gyroYangle += gyroYrate * dt;
  //gyroXangle += kalmanX.getRate() * dt; // Calculate gyro angle using the unbiased rate
  //gyroYangle += kalmanY.getRate() * dt;

  compAngleX = 0.93 * (compAngleX + gyroXrate * dt) + 0.07 * roll; // Calculate the angle using a Complimentary filter
  compAngleY = 0.93 * (compAngleY + gyroYrate * dt) + 0.07 * pitch;

  // Reset the gyro angle when it has drifted too much
  if (gyroXangle < -180 || gyroXangle > 180)
    gyroXangle = kalAngleX;
  if (gyroYangle < -180 || gyroYangle > 180)
    gyroYangle = kalAngleY;

  /* Print Data */
#if 0 // Set to 1 to activate
  Serial.print(accX); Serial.print("\t");
  Serial.print(accY); Serial.print("\t");
  Serial.print(accZ); Serial.print("\t");

  Serial.print(gyroX); Serial.print("\t");
  Serial.print(gyroY); Serial.print("\t");
  Serial.print(gyroZ); Serial.print("\t");

  Serial.print("\t");
#endif

  Serial.print(roll); Serial.print("\t");
  Serial.print(gyroXangle); Serial.print("\t");
  Serial.print(compAngleX); Serial.print("\t");
  Serial.print(kalAngleX); Serial.print("\t");

  Serial.print("\t");

  Serial.print(pitch); Serial.print("\t");
  Serial.print(gyroYangle); Serial.print("\t");
  Serial.print(compAngleY); Serial.print("\t");
  Serial.print(kalAngleY); Serial.print("\t");

#if 0 // Set to 1 to print the temperature
  Serial.print("\t");

  double temperature = (double)tempRaw / 340.0 + 36.53;
  Serial.print(temperature); Serial.print("\t");
#endif

  Serial.print("\r\n");
  delay(2);
}

and it comes up with this as well

I2C file as well (couldn't post the code because i passed the max length of 9000 characters)

What I'm thinking to do it to delete the unnecessary things and add my servos and of curse try to
map the servos with the data from the MPU6050.

What do you think, will that work?

check Jeff Rowberg lib.
about mpu6050

ı ve worked with mpu6050 for long time. i2cdevlib is the best documentation you can find.

Also, look at digital motion processing(dmp) in mpu6050 lib.

check Jeff Rowberg lib.
about mpu6050

ı ve worked with mpu6050 for long time. i2cdevlib is the best documentation you can find.

Also, look at digital motion processing(dmp) in mpu6050 lib.

So what do you suggest to use i2cdevlib or kalman filter lib? i have both.
I think kalman filter may work a bit more stable you know because of the filter but I'm not sure
never tested in a project.

here is test video

i2c is better than kalman in my opinion. easy to use and noiseless

i2c is better than kalman in my opinion. easy to use and noiseless

Nice video, so u telling me that DMP may do the job for me right? I will see what im missing lib's. tomorrow and do some tests.

mechatron17

My friend :smiley: thanks for the video the best DMP for MPU6050 ever, really stable. That's amazing but it takes 19kb of code :smiley: ow well gonna try to reduce the code as much as possible.

ı have done nothing. Thanks to Jeff Rowberg :))

btw you re welcome.

ı have done nothing. Thanks to Jeff Rowberg :))

btw you re welcome.

Thanks Jeff Rowberg as well, the best code ever, perfect for projects.

kalman code doesn't work, check again: undeclared functions and variables

Did you a question?

Make a new thread with your question.