The ultimate goal is to create an extraterrestrial unmanned system for subsurface mapping and exploration. Neural networks are to be used to recognize anomalies in the profiles that correspond to potentially exploitable subsurface features. The ground penetrating radar (GPR) techniques are likewise identical. Hence, the preliminary research focus on GPR systems will be directly applicable to seismic systems once such systems can be designed for continuous operation. The original GPR profile may be very complex due to electrical behavior of the background, targets, and antennas, much as the seismic record is made complex by multiple reflections, ghosting, and ringing. Because the format of the GPR data is similar to the format of seismic data, seismic processing software may be applied to GPR data to help enhance the data. A neural network may then be trained to more accurately identify anomalies from the processed record than from the original record.
Continuous monitoring of the lunar or Martian subsurface using on-board pattern recognition and neural processing of Rover geophysical data
1990-01-01
Miscellaneous
No indication
English
Searching for Subsurface Lunar Water Ice Using a Nuclear-Powered Rover
British Library Conference Proceedings | 2005
|Autonomous Martian flying rover
NTRS | 1990