Autonomous driving
Hello,
I am working on an advanced project. I have a small RC car with integrated Arduino Uno R4, MPU 6050, Ultrasonic sensor that is mounted to a servo at the front of the car, and two optical encoders for the left and right wheels. The goal of the project is for the car to autonomously drive around the house and map its environment using the ultrasonic sensor. All of the various sensor data is transmitted live over the Wi-Fi connection.
The biggest challenge we are having is Determining the cars heading and its x and y position. We use the Gyro of the MPU6050 to update the heading when the car turns left or right. It starts off actually working pretty good. We program the car so that it can only make 90° turns and a generally achieves 90° +/ -2°.
My basic question is has anyone integrated the MP 6050 into an actual advanced project like this? And if so, would love to better understand how they solved the various noise, calibration, accuracy, challenges. I have tried complimentary filtering, Kalman filtering, fusing with the optical encoder, pretty much everything, but always the same result...after several turns, it is off by 10 degrees or more. Strategies, or projects that have solved this, would love to have links to them.
That is pretty typical for gyro-only navigation with sensors like the obsolete, long discontinued MPU-6050, but you can do better with a modern gyro, and much better with a 9DOF sensor that gives a North reference for absolute yaw angles. Both sensors must be carefully calibrated after mounting in the model.
What you have now is probably some sort of clone, which may not even meet the specs of the original MPU-6050.
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.