The data of MPU6050 and QMC5883L can be read, and the data fusion of both can be realized, In github, I also provide a QMC5883L calibration value calculation script, which also contains the calibration method,Currently, only Chinese and English are supported.
you can see example sketch, It includes how to read the data and how to get the merged data.
In this library, the data of MPU6050 and QMC5883L can be read separately or at the same time. However, only the same I2C interface can be used when the two sensors are read at the same time. If the data of only one sensor needs to be read, you can set it by using the begin() method
You can change the parameters in the begin method to determine the sensor to be initialized. The parameters are: initialize both sensors, initialize only QMC5883L, and initialize only MPU6050. Note that you only need to call the begin method once in your project.
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Wire.begin();
imu.begin(CHOOSE_ALL); // Both sensors are initialized
//imu.begin(CHOOSE_MPU6050); // Only MPU6050 is initialized
//imu.begin(CHOOSE_QMC5883L); // Only QMC5883L is initialized
}If both sensors are successfully initialized, the following output is displayed on the serial monitor
***********************************************************
QMC_Init!!!
MPU_Init!!!
***********************************************************If the sensor fails to be initialized, the following output is displayed on the serial port monitor
*****************************************************************
QMC_Init false!!! Please check whether the connection is correct
MPU_Init false!!! Please check whether the connection is correct
*****************************************************************Note that if only one sensor is initialized individually, only information about the corresponding sensor will be output ! ! !
QMC5883L should be calibrated before use, otherwise correct readings cannot be obtained. The obtained Offest and Scale should be recorded and filled in the QMC5883L_SetOffsets method and the QMC5883L_SetScales method
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Wire.begin();
imu.begin(CHOOSE_QMC5883L); // Only QMC5883L is initialized
imu.QMC5883L_SetOffsets(***,***,***);
imu.QMC5883L_SetScales(*,*,*);
}The different placement angles of QMC5883L relative to the magnetic field will also affect the final output Angle, and the corresponding deviation can be set in the Heading_Offset method
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Wire.begin();
imu.Heading_Offset(360); // Set Heading offest
imu.begin(CHOOSE_QMC5883L); // Only QMC5883L is initialized
}The calibration of QMC5883L can be done by calling this function, after the program runs, please follow the comments below, please note that the calibration of QMC5883L does not need to run every time the code is executed, so you do not need to write this function in your code!
/**************************************************************************************************
Author: Vegetable_SYC
Address: https://github.com/Vegetable-SYC/IMU_Fusion_SYC
Version: v1.1.6
QMC5883L calibration steps:
Step 1: Start by running the following code directly and viewing serial monitor
step 2: You need to keep turning QMC5883L until the progress bar is full
Step 3: Enter the value obtained in step1 into imu.QMC5883L_SetOffsets and imu.QMC5883L_SetScales
Step 4: Annotation imu.QMC5883L_Calibration();
Step 5: Uncomment the function imu.Calculate and Serial.println(imu.getHeading())
Step 6: Upload the modified program to get the correct Angle value
**************************************************************************************************/
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
Serial.begin(9600);
Serial.print("\n");
Wire.begin();
// After calibration set the QMC5883L calibration value
// imu.QMC5883L_SetOffsets(-117, -634, 0);
// imu.QMC5883L_SetScales(1.00, 0.91, 0);
imu.begin(CHOOSE_QMC5883L); // Select QMC5883L
}
void loop() {
imu.QMC5883L_Calibration();
// imu.Calculate();
// Serial.println(imu.getHeading());
}The gyroscope calibration of the MPU6050 is automatic. After calling the MPU6050_CalcGyroOffsets method, the MPU6050 only needs to be placed on the horizontal surface. During the calibration, the MPU6050 should not be moved
Please note that the automatic calibration procedure needs to be executed after the MPU6050 is initialized, and the execution order cannot be changed, otherwise an error may occur!!!!
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Wire.begin();
imu.begin(CHOOSE_MPU6050); // Only MPU6050 is initialized
imu.MPU6050_CalcGyroOffsets(); // MPU6050 calibration
}The serial port monitor will output the following information during calibration
***********************************************************
MPU6050 is being calibrated
Uneven placement can cause data errors
Calibration complete!
X_offest : 6.14
Y_offest : 1.02
Z_offest : 2.11
***********************************************************If you know the deviation of the MPU6050 gyroscope, you can choose to manually set the deviation, so that the device initialization time will be greatly reduced.
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Wire.begin();
imu.begin(CHOOSE_MPU6050); // Only MPU6050 is initialized
imu.MPU6050_SetGyroOffsets(***,***,***);// MPU6050 calibration value
}Attention please!! Before performing data fusion, ensure that the MPU6050 and QMC5883L readings are correct; otherwise, the fused Angle value will be incorrect.
A simple data fusion scheme is provided in the library, which can effectively fuse the data of the two sensors, and the combination of the two can effectively suppress the zero drift of the MPU6050 and enhance the accuracy of the Angle of motion.
In the actual Angle fusion, I found that there would be some burrs when the Angle changed, so I had to use the filtering algorithm. Although the median filter could filter well, when the Angle data changed between 0 and 360, the Angle would jitter greatly due to the large change amplitude. Therefore, I adopted an incremental data processing method. The parameter of the DataFusion () function represents the smoothing factor alpha, which ranges from 0 to 1. If no parameter is written, the default value is 0.1. The smaller the value of the smoothing factor, the smoother the output of the Angle, but the slower the response. The larger the value, the more accurate the output of the Angle and the faster the response.
#include "IMU_Fusion_SYC.h"
IMU imu(Wire);
void setup() {
// put your setup code here, to run once:
Serial.begin(9600);// Set baud rate
Wire.begin();
imu.QMC5883L_SetOffsets(***,***,***);
imu.QMC5883L_SetScales(*,*,*);
imu.Heading_Offset(360); // Set offest
imu.begin(CHOOSE_ALL); // IMU initialization
imu.MPU6050_CalcGyroOffsets();// MPU6050 calibration
}
void loop() {
// put your main code here, to run repeatedly:
imu.Calculate(); // Calculating data
Serial.println("");
Serial.print("Data_Fusion:");
Serial.println(imu.Data_Fusion(0.1)); // Output fusion Angle
}MIT
Vegetable-SYC