| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
hi im using an asctec quadrotor, and i would like to map the environment, the only odometry data that i can provide is the velocity, can someone please suggest which form of mapping is the most suitable ? im using a hokuyo urg lidar as my sensor
There are two open source options available that I know of and that should work 'out of the box' after setup of the system (robot model, tf frames etc.) :
The mapping system developed within the CityFlyer project from CCNY uses a scanmatching approach for generating a odometry estimate and feeds this to gmapping. Described here.
hector_slam uses fast scan alignment at 40Hz directly and does not require odometry data. Both available tutorials should help with robot setup. Described in this paper. (Note I'm involved here, shameless plug and all :) )
Both approaches have been tested successfully with datasets recorded on quadrotors.
yea im trying to use hector mapping, but im having problems with the bag file. I can get the map data with the sample bag files put up but cant get anything from the bagfiles i create. im subscribing to /scan /tf and /asctec/imucalcdata , the tf frames are just basic ones that just describe the location of the lidar with respect to the center of gravity of the quadrotor. can u please tell me if im missing anything im using the hokuyo urg 04lx
Mapping using only a 2D range sensor on a 6DOF vehicle is quite tricky. If you can keep the altitude fairly fixed (or are willing to pretend that the world is uniform along the z-axis, which is fairly true above desk height), you can probably use gmapping. If your quadrotor has an accelerometer, you can use that to figure out which way "up" is, and then project the 2D scan onto the X-Y plane, and pass it to gmapping.
Asked: 2011-11-14 23:27:58 -0500
Seen: 3,651 times
Last updated: Feb 02 '12
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
