Hello all. I have a project in mind, that is basically a wirelessly controlled robot that is driven by tilting a controller. Imagine a circular controller that you simply tilt in the direction you want the robot to move. I was thinking about using a 5 or 6 DOF IMU, but if I wanted to only use a dedicated accelerometer or a gyro, which one would be better suited for my project? It should be able to detect rotation of the controller around the x and y axes. It would also be nice if it could compensate for any lateral movement caused by the operator walking around behind the robot. Any thoughts?
If you want to detect tilt then a gyro would seem the sensible choice, but you'd need to compensate for drift somehow.
I have had pretty good tilt monitoring using a 2-axis accelerometer.
You measure up to +/- 1g when tilting by the pull of the earth's gravity.
Averaging results reduces shake and vibration.
Or how about my first idea for using an IMU, would that greatly increase complexity or are IMUs reasonable to handle once you get the basics down?
With an accelerometer you have pitch and roll.
With an IMU (basically an accelerometer and gyro combined) you get pitch (front-to-back tilt), roll (left-to-right tilt) and yaw (lateral twist).
Do you need yaw?
No, I think not. If this makes any sense, it just needs to have the basic operation of something like that game, Labryinth, for iOS devices, which simply detects tilt of the device. I would also prefer for the tilt detection to be precise, so that I can discern between a slight tilt to the right and a hard tilt to the right, so that the robot can be driven faster or slower in a certain direction. So basically, tilt around the X and Y axes. I have never used accelerometers or gyros before, so the simpler the better. However, if an IMU would be better in the long run and offers more options, I'm up for the challenge.
An accelerometer that detects a little over 1g would be ideal. You will be measuring between 0 and +/-1g in any direction for tilt.
You can get SPI ones for around $5 or so.
So, if I am buying an accelerometer, the lower the rating for grams is, the more precise is it at lower amounts of tilt? So a 1g accelerometer will be able to detect finer tilt than a 3g one?
So, if I am buying an accelerometer, the lower the rating for grams is
No, grams is a measure of mass.
'g' is the acceleration due to gravity, which is what gives mass weight.
A wireless wiii nunchuck seems to be ideal for your project.
You then have the option to either use the joystick or built in 3D accelerometers.
Or actually, I've looked at some tutorials and code for the ADXL345 and it is indeed quite complex from an accelerometer beginner's standpoint, so I'm considering the ADXL335, which is analog and is quite simple. Would the ADXL335 be precise enough to determine orientation in my project?