I tried using MPU-6050 to get yaw pitch roll .but even after using Jeff Rowberg's I2Cdev library I am getting unstable yaw.so I tried quaternions .but even those were unstable as for no movement the were giving different values.l look up for other sensors .there I found two-1.MPU-9150 with obsecure dmp documentation.2-BNO055-ACCLEROMETER ,GYRO,MAGNETOMETER+CORTEX M0  Rolling out quaternions or Euler's as requested.so I want to Kno how reliable is BNO055 vs MPU-6050.If anyone has used it can you give me some insight
Thanks in advance