2
\$\begingroup\$

I want to navigate the RC car from point A to point B. The gps module is used to find the current location coordinates i.e point A. The user enters the point B coordinates based on point A. I am wondering what algorithm i should use to move the rc car i.e this algorithm should tell the robot to move forward or turn right until it reaches point B.

I've been looking at kalman filtering algorithm. Is there a another algorithm out there which i can use.

Environment: The robot should navigate indoors. I have to use the GPS because it is a requirement. The GPS is the only navigation sensor present on the robot.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ What are the typical length-scales we're talking about? Indoor-as-in-small-living-room or indoor-as-in-aircraft-hangar? \$\endgroup\$ – drxzcl Feb 22 '11 at 9:18
  • 1
    \$\begingroup\$ Are there any obstacles? Do you know where they are? \$\endgroup\$ – drxzcl Feb 22 '11 at 9:18
  • 1
    \$\begingroup\$ Do you have a time-limit? Do you have any incentive to be quick? \$\endgroup\$ – drxzcl Feb 22 '11 at 9:23
  • \$\begingroup\$ @Ranieri- I don't have a time limit or obstacles to worry about. The typical length scale is a small living room or hallway. \$\endgroup\$ – Praveen Feb 22 '11 at 16:10
  • 1
    \$\begingroup\$ You might as well just move randomly (e.g. get a Roomba as @W5VO suggests) and leave the required GPS module unpowered. It will be equally useful on or off as the accuracy will be insufficient to know that you've arrived at any given point with ~5-10m indoors. \$\endgroup\$ – Nick T Feb 22 '11 at 21:49
2
\$\begingroup\$

GPS indoors is a terrible idea. Your accuracy will be so poor, and being indoors implies you would need something fairly accurate because the scale of your setup will be relatively small.

Additionally, even if you used it outside, some form of dead-reckoning (e.g. compass) would be of enormous use so you could determine which direction you were pointed. It would not strictly be essential, only if you wanted to approach this in an efficient manner.

Using some sort of filter with GPS is a good idea, but if you are indoors, the position it will provide will not simply be randomly distributed around the true position, but rather around some chaotic mix of positions of the nearest windows based on what satellites you can get a fix on.

\$\endgroup\$
4
  • \$\begingroup\$ Nick, i agree with you that using GPS indoors is a terrible idea. Do u know of an algorithm that will help the rc car or a robot to move from point A to point B autonomously? \$\endgroup\$ – Praveen Feb 22 '11 at 2:54
  • 2
    \$\begingroup\$ @Praveen: 1. Determine vehicle orientation, 2. Determine angle and range to target, 3. Travel to target. \$\endgroup\$ – Nick T Feb 22 '11 at 3:09
  • 2
    \$\begingroup\$ @Praveen 1. Put GPS module on Roomba. 2. Tell Roomba to clean your indoor room. 3. Have microcontroller read GPS and kill Roomba power when within X feet of target. 4. Profit! \$\endgroup\$ – W5VO Feb 22 '11 at 6:18
  • \$\begingroup\$ Assuming that there are no obstacles of course! \$\endgroup\$ – BG100 Feb 22 '11 at 7:56
1
\$\begingroup\$

GPS does not work indoors.

Here's the code to my GPS guided RC car that I did a while back. It uses a proportional control loop (the P part of PID Controller):

https://github.com/tapatio/gpsrobot/blob/master/code/target/gpsrobot.c

Oops, just realized this is an ancient thread. I'll post my solution anyhow.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.