Computer generated images for path following for an unmanned vehicle

A method of path following for an unmanned vehicle using previously taken video sequences has been shown to be viable. This system has now been extended to use computer generated video sequences to control the path-following capabilities of the full-size vehicle. The system operates by `matching´ live incoming video images of the unmanned vehicle´s environment with previously obtained snapshots or computer generated images. There is a detectable correlation between the incoming video and previously obtained video sequences provided that the vehicle is `roughly´ aligned with the previous path. The accuracy at present is not particularly high but is acceptable for the control of the unmanned vehicle. The accuracy can be improved by continuously tracking the correlation between video snapshots over a period of time and by taking the original video images close together