Go Down

Topic: Help me with tracking hand position (Read 18 times) previous topic - next topic


Hey I want to build a 3dof robotic arm which will be controlled by hand gestures. This is my fyp. The three dof's are 2 linear movements (horizontal , vertical) and an angular movement(yaw). That's how the hand moves and the robotic arm exactly replicates it.

The problem is that I first thought that I will use accelerometer to track the linear position of my hand (integrating the acceleration data twice), but when i deeply researched into that area(on the net, ofcourse), I was shocked to realize that we cant get position data out of an accelerometer because of the noise issues, and that even after 5-10 seconds of operation, the resultant position is several metres drifted from the actual position.

Now I just want to know that is there any technique which I can use instead of the accelerometer to get linear position, but I want a sensor that is a single unit (for example that I can wear on my hand and no other device is required to track it, as IR tracking).



Perhaps what you need to research are waldos. Exactly mimicking your hand motion without using physical sensors, such as flex sensors and potentiometers, and physical linkage between the sensors, has not been proven to be possible in labs with more money, time, and skill than you (currently) possess.

If you do figure out a way, you'll become very rich.


Hey I want to build a 3dof robotic arm which will be controlled by hand gestures.
There are a number of projects on youtube that might be of help for your project.

Consider the daffodil. And while you're doing that, I'll be over here, looking through your stuff.   8)

Go Up