Shandor Dektor

Gradient Based Terrain Relative Navigation (TRN) Current Terrain Relative Naviation techniques rely on correlating measurements of altitude against an stored elevation map.  Gradient based TRN extends this ability to vehicles without altitude measurements.  With gradient TRN, the fundamental measurement is the local slope.  As the robot drives along, it measures roll and pitch angles, which can be correlated against a differentiated elevation map. Gradient TRN opens up TRN to novel environments:The Moon + Mars!  GPS denied environments (forests, areas with active GPS jamming)Underwater canyons! Gradient TRN, however, poses greater challenges than altitude TRN:

• Lower Signal to Noise Ratio

Differentiating an elevation map amplifies errors present in the original DEM.  Higher SNR creates many challenges with both altitude and gradient TRN - it likely contributes to the well documented problem of 'false fixes' - overconfidence in a wrong estimate.  Current TRN algorithms are susceptible to false fixes in low information/low SNR environments.  The focus of my research is on solving this problem.

• Vehicle Odometry

The Rover position estimate is highly dependent on the quality of the vehicle odometry.  Rover odometry is typically derived from kalman filtering IMU, compass, and odometry.  This produces an optimal estimate of the vehicle state to be used in conjuction with Terrain Correlation. On of the benefits of working with rovers is I get to test the algorithms on our ATRV-Jr!  It's a great excuse to head out of lab for a nice walk - if you're interested in my research, you're welcome to join me!