Research On Coopertive Navigation For The Long-range Rover Localization On Mars

Mars probing has very important scientific value, so the exploration of Mars isan significant part of universe exploration. Mars rover is a kind of mobile robot,while it also can travel on the Mars and carry out exploration mission. For theindependent navigation system is a very major part of Mars. This article we studiedthe independent navigation for the long-range rover localization on Mars.At first, we built the fidelity LIDAR model, then we introduce a method thatthrough coordinate transformation of the original terrain data, re-sampling andbilinear interpolation to build a uniform grid elevation map. When gridding nearestneighbor interpolation is used and features can be detected using morphologicaldilation. At last features detected from the global and local maps are matched basedon the DARCES (Data-Aligned Rigidity-Constrained Exhaustive Search).Secondly, we have presented a stereo vision based motion estimation techniquefor Mars rover autonomous navigation. With these techniques, the Mars rover canestimate the position and atititude accurately, also can estimate the motion initiallyand obtain the maxis of the neighboring state of motion.Finally, initialized with pose estimates and knowing correspondences betweenorbital elevation and local features, the algorithm refines the rover’poses. Odometryand orientation measurements can be introduced into MOGA(Multi-frameOdometry-compensated Global Alignment) to further refine the alignment. Then thealgorithm can be accurately and autonomously localized a Mars rover over long-range traverses.