The descent image motion estimation system (DIMES) is a system of hardware and software, designed for original use in estimating the horizontal velocity of a spacecraft descending toward a landing on Mars. The estimated horizontal velocity is used in generating rocket-firing commands to reduce the horizontal velocity as part of an overall control scheme to minimize the landing impact. DIMES can also be used for estimating the horizontal velocity of a remotely controlled or autonomous aircraft for purposes of navigation and control.

The DIMES Algorithm fuses images of terrain below with IMU data and radar-altimeter readings toestimate horizontal velocity.
DIMES was developed by the Mars Exploration Rover (MER) Project and was used successfully by the MER Entry Descent and Landing (EDL) system during both landings. In the original spacecraft application, when the need to determine horizontal velocity was discovered, it was too late to install traditional horizontal-velocity-measuring radar in the spacecraft. DIMES was conceived as a means of estimating the horizontal velocity by augmenting data acquired by sensors already installed in the spacecraft with data from an easy to accommodate descent imager.

The DIMES sensors include a descent imager (an electronic camera that acquires images of the approaching terrain), a radar altimeter, and an inertial measurement unit (IMU). The DIMES flight software implements an algorithm for combining measurement data from the aforementioned sensors to estimate horizontal velocity.

The input required by the DIMES software includes three descent images. For each descent image, the software also requires the following elements of the state of the landing spacecraft at times when the images were acquired: the attitude of the spacecraft relative to the surface, the horizontal velocity estimated by the IMU, and the altitude. Using this state information, the software warps each image to the ground plane, then computes horizontal displacements between successive images by use of image correlation, applied to two locations in each of the first and second images and two locations in each of the second and third images. This process yields four image-based estimates of horizontal velocity. These estimates are compared to each other for consistency. As a further consistency check, accelerations are computed from differences of these velocities and these accelerations are compared with accelerations as measured by the IMU.

The results of the consistency checks are used, along with image-correlation metrics to decide whether the estimate of horizontal velocity is correct. In the original application, if the estimate of velocity is determined to be correct, it is sent to the rocket-firing subsystem; if the estimate of horizontal velocity is found to be incorrect, then the rocket-firing subsystem is commanded to proceed without a DIMES velocity estimate.

This work was done by Andrew Johnson, Yang Cheng, Reg Willson, Jay Goguen, Alejandro San Martin, Chris Leger, and Larry Matthies of Caltech for NASA's Jet Propulsion Laboratory.

The software used in this innovation is available for commercial licensing. Please contact Karina Edmonds of the California Institute of Technology at (626) 395-2322. Refer to NPO-40920.