ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

How can I use a GPS sensor for obtaining ground-truth on a mobile robot? [closed]

asked 2014-05-12 08:03:04 -0600

Matias gravatar image

I have a GP-635T sensor which I am reading using the nmea_navsat_driver package (which provides the nmea_serial_driver node). The sensor does not seem to work as easily as the GPS sensor on a smartphone: it takes a long-time to lock-in and gets lots very easily. At the moment, I haven't been able to obtain very good and stable readings such as those obtained on smartphones. How does a GPS sensor need to be used to get more stable and precise readings? Do I need to fuse this with other information such as odometers or an IMU, using EKF or similar? Is it feasible to use this type of GPS sensors for outdoor ground-truth?

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by Matias
close date 2015-06-26 17:27:52.185652

1 Answer

Sort by ยป oldest newest most voted

answered 2014-05-12 08:14:20 -0600

ahendrix gravatar image

A single GPS is not suitable for ground truth outdoors. Even the very best autonomous systems will have a position error of around 1m.

If you really need accurate ground truth, you may want to look into a differential GPS system.

edit flag offensive delete link more


the position errors are quite large and erratic. I'm not sure if it is due to this particular sensor, or the fact that GPS sensors are generally fused with other inertial sensors, as I suspect is the case for smartphones. What would be the correct way to use a gps sensor in ROS?

Matias gravatar image Matias  ( 2014-05-19 05:09:03 -0600 )edit

Question Tools


Asked: 2014-05-12 08:03:04 -0600

Seen: 548 times

Last updated: May 12 '14