Inertial Navigation for Low Speed Vehicles

"Will 2-axis accelerometer i.e. x and y and 1 gyro will be enough?" Enough for what?

An inexpensive inertial sensor would be a glass or water.

Can you shed a little more light on your project? Size, operating environment (is this for a Martian rover) might be useful.

@Doorman

Okay details about this is :-

I am working on a final year project "inertial Navigation system for low speed vehicles". Few weeks ago I knew nothing about Inertial navigation. I just have to focus on algorithm so i dont want to waste my money on expensive sensors as precision is not my requirement for now. I think 2 axis accelerometer will be enough for me so I selected adxl202/212, but for GYROs, my supervisor told me ADXRS150 is OK but 0,05 deg/sec or 1800 deg /hr is too much drift. So i need GYROS with 5-10deg/hr drift. I hope I can get some help from here.

You are working in X and Y axis only, and this is an earthbound low speed vehicle. Even 5°-10° per hour will result in a large Abbe error.

Are you building the inertial measurement unit? You tell us you are working the algorithm (a Kalman filter?) but you are worried about the expense of components.

Yes, you got everything right. I Am worried about the expense of the components, I am building a IMU Which can be used alternate to GPS OR with GPS and I will work on Kalman Filter

WHat will you suggest now ?

What sort of magnetometer for static calibration? Vector? Fluxgate compass? If you are operating in one plane, using GPS for dynamic calibration and running a Kalman filter, why use a gyroscope?

If you are operating in one plane. Flat. Left and right hand turns only. No hills.

What distance and obstacles does this have to deal with? If you will live with a gyroscope that has 10°/hour drift, it seems it will be of little value.

Just a fluxgate compass to sense the direction.

I am using a gyroscope because its a requirement from a senior management. attaching GPS was my idea, that needs to kept aside for the moment.

Yes I will operate in one plane. I will give a demo in some robotic car so that will be a one plane, Obstacles. . Speedbreakers, little bumps maybe.

Yeah, I was sort of surprised to see you are allowed a GPS device. I am not surprised to see a gyroscope listed as a required element. Okay, you hadn't stated that, so now we know. Do you see how important communication can be?

I help a fellow with radio control airplanes. The Analog ADXRS450 is what we almost always use for them. These withstand a lot of vibration from a plane and might be more robust than you really need. Digi-Key sells this for about $50.00 USD each.

Yes I guess so and thank you for the help. Btw should I buy the kits for the accelerometer and gyros which would be costly i guess OR should I buy ICS and pass them through butterworth filter following with amplifiers into the controllers?

8051 OR DSPs OR FPGA's?

Yes, I am building IMU, Kalman Filter.

Inertial Sensors are not easily available in my country. So, if you can tell me the price of these sensors or alternate. I will Use Magnetometer for static callibration aswell as Gps for Dynamic callibration.

Go here:Search GlobalSpec and search for "low cost sensors".

You'll have to make your own decisions about suitability since we have no idea of the requirements.

Are you a student?

Yes I am a student. I just have to focus on algorithm so i need low cost sensors. I dont have to acquire precision.

ANd btw I meant 5-10deg/hr not sec .. Sorry

5/10 deg/hr not sec. Sorry

Funny, I did similar as a final year project many years ago, but used no sensors!
The vehicle (3 wheeled) was set up so that it could only execute 90 degree turns and it counted wheel rotatations. It was only suppesed to navigate to an x y coordinate set in on switches, it was all discrete logic and counters.
Yeah it wasn't very accurate, but that wasn't the point, it worked!
Gotta be easier now with a micro.
Del

His algorithm will be doing the counting and logic instead. Just a different way to generate numbers and process them into feedback is all.

My RC helicopter has a gyro, no accelerometers. I have no idea what its error progression is. Flight times aren't long enough to matter really.

I hope it goes easy for me. For now it doesn't look that easy

RC gyros starting at $11.95

Could you exploit the data from the accelerometers in your smart phone?

Some smart phones even have discreet gyros as well (so I've read)

Is there any imposed restriction on the form factor or packaging of these sensors?

On board calibration as well.....

Could you not write an app for a suitably equipped schmuttfone?