MPU-6050

I while back I did some work on a self-balancing robot using and Arduino Uno and the InvenSense MPU-6050 6DOF sensor. Using the sensor is easy, thanks to Jeff Rowberg’s I2Cdev library and sample code. If you look around line 200 of the MPU6050_DMP6 example arduino sketch that comes with the library you will see the following: // supply your own gyro offsets here, scaled for min sensitivity mpu.setXGyroOffset(220); mpu.setYGyroOffset(76); mpu.setZGyroOffset(-85); mpu.setZAccelOffset(1788); // 1688 factory default for my test chip I did some further reading on the I2CDev forum and found several threads on calibrating the MPU-6050 sensor and determining the optimal offsets. These, apparently, are specific to your device, as well to the exact orientation of the module, once it is installed. I found a significant improvement using the script by Luis Rodenas attached to this forum thread. Below is the Arduino sketch version 1.1 (the most current at the time of this post). For best results, mount your module and run the script, while you have it stable in the position you will use it. For example, on a self balancing robot, have the robot upright in the optimal balanced position and keep it steady while the script completes and you see output with the suggested offsets. You then need to copy these offsets and overwrite the defaults in the sample code (around line 200, as mentioned above). ...

I spent some time reading on accelerometers, gyros, sensor fusion, PID, optimized PWM motor control. I tinkered with the components I have for a while and then procrastinated for even longer. Finally, I can say that I have a prototype of a self balancing bot that shows promise. More fine-tuning is needed, and I plan to add a Bluetooth module to get PID controller data wirelessly, possibly to adjust the PID coefficients in real time and, hopefully to steer the bot remotely someday. ...