BMX160 evaluation test

[ShaKhu]

Hello everyone,

I am evaluating BMX160 IMU sensor for use in our products. The specifications of the BMX160 are highly competitive and it is a leading candidate among IMUs that I am evaluating. I have a shuttle board for the same BMX160 sensor for the purpose of evaluation; however, I am having difficulty in evaluating it's performance.

My evaluation test comprises of measuring displacements using BMX160, that is if I move BMX160 then it should tell me how much it has moved let's say in 2D plane, I assume that it is lying flat on the table. In other words, I am trying to recreate trajectory of IMU movement on 2D plane. I will be highly thankful if you can kindly send me or point me towards sample source code/existing software packages/drivers where I can use BMX160 to recreat its trajectory in 2D space.

Please, feel free to ask for further details. Your prompt response will be highly appreciated.

I look forward to your positive reply.

Thanks & Regards,

Khu

[BSTRobin]

Hello ShaKhu,
BMX160 output sensor raw data, you need run your algorithm for IMU movement on 2D plane.

[ShaKhu]

Hello there!

Thanks a million for your prompt response, I was not expecting any response to begin with. Anyway, I have written an I2C driver which can read BMX160 raw data. I have also converted the raw data from IMU coordinate system to Real world coordinate system. As a matter of fact, I have tried to solve the problem using following approaches:

1) Using the data from accelerometer only. I use the direction of g vector to determine the orientation of the IMU and then do all the calculations to determine the trajectory in 2D space.

2) Using the data from gyroscope to determine the orientaion of IMU. I determine the Alpha, Beta and Gamma rotation angles along x, y and z axis respectively and then do all the Jazz once again over the data of acceleromeer to determine the trajectory in 2D.

3) Using omega arithmatic. Okay, I still need to complete some work over here. This approach is still work in progress.

Anyway, can you just let me know whether I can capture hand writting using BMX160. Again, this is going to be embrassing as I don't expect anyone to do my work. But, can you kindly, let me know the "recommended" steps to get the job done as I am out of ideas. You can instruct me the recommended approach for example: Filter the data using Kalman filter, use omega arithmatic, etc. Thats all I need, although I still need to put my head around about how to go about testing my implemetation.

Thanks again for taking time out of your busy schedule and replying me back!

Regards,
Khu

[BSTRobin]

Hello ShaKhu,

BSXlite is an official algorithm library for your reference. Basic on sensor RAW data, BSX is a complete 9-axis fusion solution, provides orientation information in form of quaternion or Euler angles.
https://www.bosch-sensortec.com/software-tools/software/sensor-fusion-software/

For more in-depth algorithms, if you have a specific project, you can contact the agent in your location to let the agent evaluate your project.
https://www.bosch-sensortec.com/about-us/contact/our-distributors-and-sales-representatives/

[ShaKhu]

Thanks, a lot for the information. I really appreciate that you have tried to help me out. Thanks a lot for your time.

I am not using ARM and hence BSXLite is of limited use to me. I guess, I can evaluate IMU on an ARM board. Moreover, please, correct me if I am wrong, as far as I know the BSXLite library does not support BMX160 hence BSXLite is of no use to me. As the library is not open source hence I can not even modify it for BMX160.

Anyway, thanks a lot for the pointers.

I am from India, there are no BOSCH agents listed on the web page for here. When I contacted "Bosch Sensortec Customer Support" via contact@bosch-sensortec.com, they pointed me here to get the assistance. We are living in difficult times, I wonder whom should I contact to sort out the matters.

I don't want to reject a perfectly good IMU from a very reputable organization because of my inability to use it.

Thanks again for all your help.

Thanks,

ShaKhu