Rover locations

M20 MER MSL

Because Mars does not have a GPS satellite constellation to help pinpoint locations, science teams track position in relative terms as the rover moves and making corrections along the way.

Site, drive, and pose

Rover positions along its traverse are identified by a system of “site”, “drive”, and “pose” counters, or indices, that define a location hierarchy. A new site is declared when the science team decides a location has scientific significance or there is a tactical or operation need. That resets the drive and pose values to zero, for example, site 4, drive 0, pose 0. When the drive system determines it's time for a new drive location within the current site, the drive counter is incremented, and the pose counter is reset to 0. The values for site almost always increment by 1. Drive values rarely do.

Site	The Site index defines which instance of the Site frame is relevant a particular rover motion counter (RMC). Whenever the Site frame is incremented, all the other RMC values are set to 0. Declaring a new Site frame thus resets all motions and creates a new local area in which to work. Unlike the other indices, there is no meaning to odd or even values. Also, unlike other indices, Sites start with 1 (the landing site), not 0.
Drive	The Drive index (sometimes incorrectly called Position) increments whenever the rover drives or otherwise moves its wheels (e.g., trenching or steering). The value is odd while the wheels are moving, and even while they are not. Incrementing Drive sets all the subsequent indices (i.e., all except Site) to 0. Exception: if a mechanism index is odd, indicating it is moving during the drive, then that index is not reset to 0. This should be an unusual case, however. The Drive index is the one of highest interest for most localization activities.
Pose	The Pose index indicates a change in the rover’s knowledge of its position or orientation when it is not actually driving. Unlike the other indices, it does not indicate that anything moved; only that the pose knowledge has changed. This may be due to running visodom (visual odometry), doing a sun find, reading the IMUs (Inertial Measurement Units), or receiving an explicit ground command to update the pose knowledge. Note that the rover might have moved slightly, mostly in orientation, due to being pushed by the arm (e.g., during a drill preload). There is unfortunately no reliable way to distinguish this from an attitude update, although the arm index will change during arm movements. So, a pose change does not indicate arm-induced motion, but lack of an arm update does rule it out.

Site/drive/pose locations are often shown in the Notebook without using the words drive and pose. For example, you might see Site 8 / 610 / 4 meaning site 8, drive 610, pose 4.

The ROVER_MOTION_COUNTER value in data product labels contains the site, drive, and pose information for a given product.

Note that science team documents sometimes refers to "site and position" rather than "site and drive". Yes, it can be confusing.

Counting wheel turns and mapping locations

Wheel turns are counted as the rover moves to determine distance. Knowing the Because of slip and skid along the surface during the drive, visual odometry image pairs are collected and internally processed to make adjusts on the go. Knowing drive direction and the distances between the drive positions, we can start to plot locations on a map. After all, we know the map position of landing.

Further corrections are made by hand via adjustments to site/drive locations by comparing the HiRISE basemap and visuals in the collected images. Locations of objects within collected images are calculated in Rover XYZ space (origin in a fixed space relative to the rover body, +X in front of the rover, +Z down into the planet). By knowing the object’s location within a given image, the camera’s position and design (via a camera model), the orientation of the acquired image (via a set of orientation vectors), and the pixel location of the object within the image can be calculated.

This interwoven tapestry of positions and locations can be used to create a robust drive traverse map and multi-image mosaics, connect coincident data from multiple instruments, and find images of a given target taken from other rover positions.

Getting more information

Rover localization data are found from links in the Resources section in the AN.