ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
1 | initial version |
A possible slight difference could be that ego-motion is more about the estimation of the twist (linear and rotational velocities, both 2D or 3D cases), and odometry is the estimation of the path. In some cases, like in wheel odometry, this path estimation is computed by means of time-integration of an estimated twist. But in other cases, like classic visual odometry or scan matching, the poses are directly estimated, without performing any time-integration. Anyway, I don't think that a strong consensus exist within the community about that, so you should take into account the context.
2 | No.2 Revision |
A possible slight difference could be that ego-motion is more about the estimation of the twist (linear and rotational velocities, both 2D or 3D cases), and odometry is the estimation of the path. In some cases, like in wheel odometry, this path estimation is computed by means of time-integration of an estimated twist. But in other cases, like classic visual odometry or scan matching, the poses are directly estimated, without performing any time-integration.
time-integration.
Anyway, I don't think that a strong consensus exist within the community about that, so you should take into account the context. context.