Society of Robots - Robot Forum

Mechanics and Construction => Mechanics and Construction => Topic started by: [email protected] on January 26, 2013, 10:44:14 AM

Title: quadcopter construction and design
Post by: [email protected] on January 26, 2013, 10:44:14 AM
hey folks i am planing to make a quadcopter......
what would be the  best material for the main chasis....?
which sensor should be used for the stability ??? a gyro or a accelometer
and i would need the kalman filter
any input or good tutorials? please post it.....
Title: Re: quadcopter construction and design
Post by: Gertlex on January 27, 2013, 10:32:53 AM
Start reading here :)
http://diydrones.com/ (http://diydrones.com/)
Title: Re: quadcopter construction and design
Post by: jwatte on January 27, 2013, 01:37:54 PM
hey folks i am planing to make a quadcopter......

Sounds like fun.

Quote
what would be the  best material for the main chasis....?

Carbon fiber. Aluminum or Titanium could be OK, too.

Quote
which sensor should be used for the stability ??? a gyro or a accelometer

You need both, and ideally also a magnetometer. Which means you need a 9-degree IMU.

Quote
and i would need the kalman filter

Yes.

Quote
any input or good tutorials? please post it.....

diydrones.com
Title: Re: quadcopter construction and design
Post by: [email protected] on January 30, 2013, 07:20:21 AM
thanxx !!!!! for the reply.....
can anyone tell me about what's a kalman filter have no idea about it???? i know it requires a lot of math....
is there an other alternative for kalman filter ???
Title: Re: quadcopter construction and design
Post by: jkerns on January 30, 2013, 09:35:06 AM
A Kalman filter is a way to take noisy measurments (accelerations, gyro rate of change) and create estimate of what the actual "state" of your system is (attitude, speed). 

Math is a fact of life. Newton (and others) didn't invent calculus just to make life hard for students - they invented it because that is the tool that is needed to describe how the physical world operates.

You don't need to use a kalman filter, but then you would need to design an algorithm to translate from the sensor inputs to the actual state of your quadcopter. If you know the magnitude of the noise the Kalman filter will give you the "optimal" estimate of those states. For other methods you will have to sort out the gains of your estimator on your own.
Title: Re: quadcopter construction and design
Post by: [email protected] on January 30, 2013, 10:48:20 AM
there are some algorithms like averaging sensor output etc.....
can i use them???
like this one http://bleaklow.com/2012/06/20/sensor_smoothing_and_optimised_maths_on_the_arduino.html (http://bleaklow.com/2012/06/20/sensor_smoothing_and_optimised_maths_on_the_arduino.html)

OR this one http://blog.thomnichols.org/2011/08/smoothing-sensor-data-with-a-low-pass-filter (http://blog.thomnichols.org/2011/08/smoothing-sensor-data-with-a-low-pass-filter)

OR http://playground.arduino.cc/main/smooth (http://playground.arduino.cc/main/smooth)
Title: Re: quadcopter construction and design
Post by: jkerns on January 30, 2013, 11:45:25 AM
The task isn't so much to average or smooth the sensor readings. The task is to determine the quadcopter orientation in space based on the sensor inputs.

Your inputs are accelerations and rotations in three axes, you need to find attitude (in the pitch and roll axes) damp any unwanted rotation (yaw axis) and control the speed in at least one direction.
Title: Re: quadcopter construction and design
Post by: [email protected] on January 31, 2013, 02:26:15 AM
once again thanxx for the reply....  :D
can ull suggest me any other algorithm or should i learn kalman filter....????
Title: Re: quadcopter construction and design
Post by: [email protected] on February 02, 2013, 06:00:12 AM
found this awesome series of tutorials for the frame construction .....
Basic Quadcopter Tutorial - Chapter 1 (http://www.youtube.com/watch?v=7DHoLYf5pMA#ws)