Just a heads up, Invensense IMU3000 chip is now available for purchase. It's a 3-axis gyro that's designed to be paired with a digital 3-axis accelerometer, and does sensor fusion on chip, removing the need for kalman filtering on an external MCU. Outputs are, besides gyro and accelerometer readings, temperature, attitude (as a quaternion), and lineair acceleration. 

They sell a breakout board for $60 that comes with an accelerometer, so it's essentially a full IMU with handy I2C interface. I'm looking to get one but FedEx to the Netherlands isn't cheap, so if anyone is willing to share shipping costs let me know :)

Views: 2273

Reply to This

Replies to This Discussion

Hi there, I'd like to buy a couple IMU3000 and a couple of ITG-3200 .. unfortunately I can't find them anywhere in Europe.

Yes, the shipment cost is simply crazy! I live in Italy but I think it would be possible to share costs anyway. Please contact me: my mail is fvaresano at yahoo dot it.

Looking forward hearing from you.

Fabio Varesano
Looks nice, but I wonder what exactly their sensor fusion does? In some parts of the datasheet , they say they output "3 axis gyroscope, 3 axis accelerometer and temperature data".

And how would they compensate drifts without a magnetometer?
I'm living in France and I'm working on a sailboat autopilot. I think that this breakboard with IMU3000 is the best IMU for this price we can find at this time so if it's not to late I'm ready to share the shipping costs with you. Let me know if you are always OK.

The datasheet is a bit vague, but afaict there's a FIFO that can be read through I2C. The output from the sensor fusion can be read from there.

Drift is compensated for by using an external digital 3-axis accelerometer.
pitch and roll can be drift compensated with a 3-axis accelerometer in static mode but for heading it's impossible without magnetometer or GPS.
Ah, I see what you mean. I'm not sure if it's possible to feed magnetometer data back into the estimation process. I'll have to check the data sheet and documentation. I'd personally use it with a GPS, so it's not an issue for me.
My application (sailboat autopilot) like yours (drone I suggest) needs a drift compensated IMU in dynamic mode and over longterm. So we have to fusion the data from this IMU with magnetometer or GPS or the twice. A solution may be the EKF.
The Kalman filter can be divided in 2 parts: time update (predict) wich can be done by our IMU at high rate and measurement update (correct) when measures( magnetometer, GPS or others) are avalaibles.
But I have some understandings problems with the DMP inside the IMU 3000, maybe could you help me.
How to obtain fusion data ( like rotaion matrix or quaternions) from the IMU board without the ARM interface board ?
Is it possible to update theese fusion data in the DMP ?
I don't think it's possible to feed data back into the sensor fusion of the chip. I guess you could get heading estimates from the IMU3000 and write a software kalman filter to fuse it with a compass?

Afaics interfacing with the IMU board would just be a matter of connecting to the i2c pins and providing it with power.
Come on, 89,- euro?!

I'd be interested in sharing shipping with you, as the last part from Italy to the Netherlands is dirt cheap. I'd just need an IMU3000 board, PM me if you're interested. I have PayPal. I guess Rommel might also be interested, so you could try giving him a PM as well.
I think you are right and I'm beginning to understand what about this mess of DMP.
But we have to be aware for our applications in dynamic situations. If pitch and roll are updated with accelerometers measurements, every acceleration other than gravity will cause erroneous updatings. So the best we can do with this IMU board for dynamic application is to get only raw data and not fusion data.

Reply to Discussion


© 2016   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service