The Vectornav VN-100 IMU (they call it an Attitude Heading Reference System, for some reason) is out and it looks like the best device on the market in this price range. Here's the description from the site.
The VN-100 combines a 3-axis accelerometer, 3-axis gyro, 3-axis magnetometer, and a high performance processor onto a single surface mountable chip-sized module to create a high performance orientation sensor. Fully calibrated for bias, gain, and misalignment, the VN-100 accurately calculates orientation over the entire 360° range at 200 Hz. Filtered orientation data and inertial measurements are assessible via either a SPI or RS-232 serial interface.
With its small size, high performance, and low cost, the VN-100 has numerous potential applications. A development kit is available for the VN-100 which comes with a sensor pre-installed and gives easy access to all of the sensor's features.
Features and Benefits:
Complete Sensor Package
3-axis accelerometer
3-axis gyro
3-axis magnetometer
Multiple Output Types
Heading , Pitch, Roll
Quaternion / DCM
Acceleration, Angular Rates, Magnetic
High Precision
Heading accuracy < 1.0 (static)
Pitch/Roll accuracy < 0.5 (static)
< 3 deg dynamic
Fast Onboard Processor
Extended Kalman Filter
200 Hz update rate
Surface Mount Package
Hand solderable
30 pads
Ultra-Compact Size
22 x 24 x 3 mm
3 grams
Fully Calibrated
Scale Factor and Gain
Axis misalignment < 0.05 deg
Hard / Soft Iron Compensation
Digital Interface
Serial UART up to 921600 bps
SPI Interface up to 18 MHz
Low Cost
$400 - 500
*Price depends on quantity
I wasn't aware that the DevBoard was using the velocity vector from the GPS. If thats the case then the DevBoard would be at times an AHRS since it has at least two vectors. I'm not to familiar with the performance of the DevBoard in terms of attitude determination, but if you have a aircraft that maintains a healthy velocity vector at all times then the DevBoard should give you decent results. Problems will arise however as the velocity becomes lower, since uncertainty in the velocity vector now translates into a much larger uncertainty in the direction of the measured velocity vector. This also assumes that the velocity vector remains in a known orientation with respect to the aircraft's body reference frame. This may not be the case in conditions of severe sideslip, or aerodynamic perturbations. The magnetometer gives you a second inertially fixed reference vector to measure in the body frame. Since this vector isn't a function of vehicle motion, the AHRS will maintain its attitude accuracy at low velocities, and under conditions with severe sideslip or translational motion as in the case of a helicopter. Using the magnetic field isn't without its problems though, as it can be easily distorted by nearby ferrous materials. Since the planes you guys are flying are small and likely use electric motors, there is a good chance that magnets will be close enough to the mag sensors to cause a significant disturbance. Regardless of whether you use one of our sensors or possibly the DevBoard with an added magnetometer, make sure you perform a hard and soft iron calibration. This will allow to mathematically remove the effects of nearby magnetic materials from the attitude solution. It effectively maps a distorted magnetic field back into a clean spherical one. The ultimate solution for a UAV would be to combine the idea of using the velocity vector along with the magnetic vector. The more vectors that you can use to determine your attitude the more accurate your attitude solution. This is effectively the principle behind the expensive star trackers on satellites used for attitude determination. Unfortunately for us satellites have access to alot more reference vectors than we do here on earth.
Chris, to make tracking of gyros drift and solve AHRS you need two vectors in diferent directions, one from accelerometer and another from a magnetometer or any other sensor than give you heading vector (Bill Premerlani uses GPS)
you have not enough information in one only vector (only accelerometer)
Thanks for your explanation of the difference between and IMU and AHRS, but I think by that definition the UAV DevBoard is an AHRS, too. It also has an attitude computer solving the attitude equations (using the DCM, rather than a Kalman filter, but the result is the same).
What do the magnetometers contribute that can not be extracted from gyros and accelerometers themselves?.
very impressive, the size especially. It's very tempting especially because all the hardwork of combining the sensor inputs is done for you. It's expensive but not terrible compared to say the sparkfun 6DOF that I use (http://www.sparkfun.com/commerce/product_info.php?products_id=8454).
Instead of that big dev board, I wish it was offered on a simple breakout board maybe with just a voltage regulator, a mini-usb connection and the necessary pins for the RS232 connection, etc. Do all that for $500 and I'm in!
Hi Michele,
Then VN-100 would very work well for a balance robot. One of the difficulties in building a robot of this type is dealing with the gyro drift. With the VN-100, the onboard kalman filter stabilizes the gyro to give you a drift-free angular rate solution. You can run the angular rate output of the sensor along with the pitch angle straight into your control loop. In the coming months we will be posting example code for various projects to help people become familiar with how to use our sensor. A balancing robot sounds like a great candidate for a one of these example apps. If you guys think of any other apps that you would like to see examples for let me know.
Hey guys, I just noticed that our product was posted on your website. I have seen some confusion as to the difference between an AHRS and an IMU. An IMU consists of an orthogonal instrument cluster of 3-axis accelerometer and gyros. The UAV DevBoard is a great example of an IMU. An AHRS is an IMU combined with an attitude computer that takes the filtered measurements from the IMU and solves the attitude equations. Different algorithms exist to solve the attitude equations, requiring a minimum of two vectors, gravity and magnetic. In this case we use a quaternion based Extended Kalman Filter. The quaternion based algorithms give better performance due to their absence of any mathematical singularities, and the Kalman filter eliminates the drift in the gyros.
You can checkout a demonstration of the VN-100 at www.tinyurl.com/vectornav
Looks just like what we are all using anyway, just in a small package and with a STM32 processor onboard. I would have expected more than 200Hz though.
The price would be good is it was open source, otherwise it's about $200-300 too expensive.
Comments
you have not enough information in one only vector (only accelerometer)
jlcortex
Thanks for your explanation of the difference between and IMU and AHRS, but I think by that definition the UAV DevBoard is an AHRS, too. It also has an attitude computer solving the attitude equations (using the DCM, rather than a Kalman filter, but the result is the same).
What do the magnetometers contribute that can not be extracted from gyros and accelerometers themselves?.
jlcortexnmine.com
Instead of that big dev board, I wish it was offered on a simple breakout board maybe with just a voltage regulator, a mini-usb connection and the necessary pins for the RS232 connection, etc. Do all that for $500 and I'm in!
Then VN-100 would very work well for a balance robot. One of the difficulties in building a robot of this type is dealing with the gyro drift. With the VN-100, the onboard kalman filter stabilizes the gyro to give you a drift-free angular rate solution. You can run the angular rate output of the sensor along with the pitch angle straight into your control loop. In the coming months we will be posting example code for various projects to help people become familiar with how to use our sensor. A balancing robot sounds like a great candidate for a one of these example apps. If you guys think of any other apps that you would like to see examples for let me know.
Can also be used to experiment of balance robot (type sagway)?
You can checkout a demonstration of the VN-100 at www.tinyurl.com/vectornav
The price would be good is it was open source, otherwise it's about $200-300 too expensive.
-
1
-
2
-
3
-
4
of 4 Next