CMU-CS-97-100
Computer Science Department
School of Computer Science, Carnegie Mellon University



CMU-CS-97-100

STRIPE: Remote Driving Using Limited Image Data

Jennifer S. Kay

January 1997

Ph.D. Thesis

Unavailable Electronically


Keywords: Artificial intelligence, human computer interaction, mobile robots, semi-autonomous vehicles, unmanned ground vehicles, vehicle teleoperation, polyhedral-Earth reprojection, low-bandwidth teleoperation, high-latency teleoperation, NavLab, STRIPE


Driving a vehicle, either directly or remotely, is an inherently visual task. When heavy fog limits visibility, we reduce our car's speed to a slow crawl, even along very familiar roads. In teleoperation systems, an operator's view is limited to images provided by one or more cameras mounted on the remote vehicle. Traditional methods of vehicle teleoperations require that a real time stream of images is transmitted from the vehicle camera to the operator control station, and the operator steers the vehicle accordingly. For this type of teleoperation, the transmission link between the vehicle and operator workstation must be very high bandwidth (because of the high volume of images required) and very low latency (because delayed images can cause operators to steer incorrectly).

In many situations, such a high-bandwidth, low-latency communication link is unavailable or even technically impossible to provide. Supervised TeleRobotics using Incremental Polyhedral Earth geometry, or STRIPE, is a teleoperation system for a robot vehicle that allows a human operator to accurately control the remote vehicle across very low bandwidth communication links, and communication links with large delays.

In STRIPE, a single image from a camera mounted on the vehicle is transmitted to the operator workstation. The operator uses a mouse to pick a series of "waypoints" in the image that defines a path that the vehicle should follow. These 2D waypoints are then transmitted back to the vehicle, where they are used to compute the appropriate steering commands while the next image is being transmitted. STRIPE requires no advance knowledge of the terrain to be traversed, and can be used by novice operators with only minimal training.

STRIPE is a unique combination of computer and human control. The computer must determine the 3D world path designated by the 2D waypoints and then accurately control the vehicle over rugged terrain. The human issues involve accurate path selection, and the prevention of disorientation, a common problem across all types of teleoperation systems. STRIPE is the only semi-autonomous teleoperation system that can accurately follow paths designated in monocular images on varying terrain. The thesis describes the STRIPE algorithm for tracking points using the incremental geometry model, insight into the design and redesign of the interface, an analysis of the effects of potential errors, details of the user studies, and hints on how to improve both the algorithm and interface for future designs.

226 pages


Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by reports@cs.cmu.edu