Rock Segmentation in the Navigation Vision of the Planetary Rovers

Visual navigation is an essential part of planetary rover autonomy. Rock segmentation emerged as an important interdisciplinary topic among image processing, robotics, and mathematical modeling. Rock segmentation is a challenging topic for rover autonomy because of the high computational consumption, real-time requirement, and annotation difficulty. This research proposes a rock segmentation framework and a rock segmentation network (NI-U-Net++) to aid with the visual navigation of rovers. The framework consists of two stages: the pre-training process and the transfer-training process. The pre-training process applies the synthetic algorithm to generate the synthetic images; then, it uses the generated images to pre-train NI-U-Net++. The synthetic algorithm increases the size of the image dataset and provides pixel-level masks—both of which are challenges with machine learning tasks. The pre-training process accomplishes the state-of-the-art compared with the related studies, which achieved an accuracy, intersection over union (IoU), Dice score, and root mean squared error (RMSE) of 99.41%, 0.8991, 0.9459, and 0.0775, respectively. The transfer-training process fine-tunes the pre-trained NI-U-Net++ using the real-life images, which achieved an accuracy, IoU, Dice score, and RMSE of 99.58%, 0.7476, 0.8556, and 0.0557, respectively. Finally, the transfer-trained NI-U-Net++ is integrated into a planetary rover navigation vision and achieves a real-time performance of 32.57 frames per second (or the inference time is 0.0307 s per frame). The framework only manually annotates about 8% (183 images) of the 2250 images in the navigation vision, which is a labor-saving solution for rock segmentation tasks. The proposed rock segmentation framework and NI-U-Net++ improve the performance of the state-of-the-art models. The synthetic algorithm improves the process of creating valid data for the challenge of rock segmentation. All source codes, datasets, and trained models of this research are openly available in Cranfield Online Research Data (CORD).

Dimensional optimization of rocker-bogie suspension for planetary rover based on kinetostatics and terramechanics

Dimensional optimization is important for planetary rovers to reach good performance, such as high mobility, stability, and low energy consumption. The paper presents a dimensional optimization for a planetary rover with rocker-bogie suspension. During the optimization process, the influence of dimensions on the actuation requirements is studied based on kinetostatics and terramechanics. The objective function is built considering the average driving torque requirements in the most common type of windblown terrain in Mars called megaripples. The optimal dimension design is reached through the genetic algorithm, and the influences of dimensional parameters on rover performances are studied by drawing performance atlases. This work realizes the consideration of energy consumption in the design phase of a planetary rover. Finally, the results guide the design of a rover prototype and are validated by a series of experiments.

Enhanced Lunar Topographic Mapping Using Multiple Stereo Images Taken by Yutu-2 Rover with Changing Illumination Conditions

In a planetary-rover exploration mission, stereovision-based 3D reconstruction has been widely applied to topographic mapping of the planetary surface using stereo cameras onboard the rover. In this study, we propose an enhanced topographic mapping method based on multiple stereo images taken at the same rover location with changing illumination conditions. Key steps of the method include dense matching of stereo images, 3D point-cloud generation, point-cloud co-registration, and fusion. The final point cloud has more complete coverage and more details of the terrain than that conventionally generated from a single stereo pair. The effectiveness of the proposed method is verified by experiments using the Yutu-2 rover, in which two data sets were acquired by the navigation cameras at two locations and under changing illumination conditions. This method, which does not involve complex operations, has great potential for application in planetary-rover and lander missions.