Arduino sketches for BNO080 9DOF with sensor fusion. The _2 version uses the standard wire library necessary for M0/ M4 etc controllers. The nrf52 version works on Adafruit nrf52 feather and displays data on Adafruit Bluefruit LE app. The _1 version is not up to date. So take it for reference only.
The BNO080 by http://hillcrestlabs.com/ is a successor of the BNO055 from Bosch Sensortec www.bosch-sensortec.com. It offers more features and a significantly improved performance however with a quite different interface. Breakout boards are now also available http://hillcrestlabs.com/store/fsm300/ . Hillcrestlabs has issued a code using a ST Nucleo eval board. https://github.com/hcrest/bno080-nucleo-demo. There are also universal breakouts from Sparkfun https://www.sparkfun.com/products/14686 and Adafruit: https://www.adafruit.com/product/4754
The BNO080 generates high performance quaternions, linear acceleration etc: at an extraordinary high data rate (up to 400Hz). Therefore the device is best suited for robotics and other IMU applications where is desirable to use also Atmega 328 controllers like Arduino Nano and others.
This sketch provides a basic functionality for testing and further developments. The BNO080 provides more robust quaternions than the BNO055 and with improved accuracy. The calibration algorithm can be controlled and results can be stored easily in the flash. The tare function is very useful in case of relative movements.
Note that for a correct heading a proper magnetometer function is needed. Thus the sensor should be >1m away from iron masses (a vehicle´s roof). There is real time heading estimation accuracy reporting available.
It is recommended to degauss the board at least once to ensure that the sensor and adjacent components (crystal, resistors , capacitors) does not have a significant remanent magnetic field.
Degaussing can be done by a diminishing AC magentic field. You may use the coil of a magnetic valve for that purpose. Use such a coil with iron core otherwise its getting very hot. (Be aware of working with dangerous voltages!) switch on the coil at a distance (> 0.5m) , then move over the breakout, move away again and switch the coil off.) https://en.wikipedia.org/wiki/Degaussing
You may do this also with screw drivers, tweezers etc - old wisdom from magnetic tape recorders
My best practice is Use rotation vector 0x05 with 400Hz update rate. Degauss the board. Be careful with indoor use. In setup call save_periodic_DCD(); and set cal to ME_cal(1,1,1,0); This achieves best cal 3 status. Calibrate the board including iron parts in the vicinity. Temperature should be as constant as possible. Ideal would be a kind of a thermostat housing (not tested) It is important to understand that compensating gyro drift (temperature etc) requires some slow continuous movements of the sensor like wearing on the body and walking around. Having it at rest with changing temperature gives poor results