Research On Navigation And Localization Of CE-3Lunar Rover

In the second stage of Chinese lunar exploration, the rover which was carried by a lunar lander had explored on the moon and got the topography and geological structure images on the detection zone. Then it sampled and analysed the rock and mineral component on the study area. Lots of valid data of the scientific engineering had been received. The precision navigation and localization technology could ensure the efficiency of scientific detection by the CE-3lunar rover with its life cycle and determine the relative distance among the target points and the lunar rover, so that it could choose the reliable planning path, which was supported by the detection of the target points. To solve the urgent high precision navigation and localization problems of the lunar rover in these stages, we will focuse on the research of algorithms about the lunar rover data processing in this paper, including the following contents:(1) A localization algorithm is proposed that is based on the weighted least squares theory with the stereoscopic images received by the lunar rover.According to the weight matrix of the observation data based on the different resolution of the stereoscopic images of the lunar rover, the weighted total least squares theory used in the relative orientation and Procrustes algorithm used in the absolute orientation are formulated, then the procedure is described. At last, the experiments of the CE-3lunar rover are made to prove the practicable which make use of the weighted total least squares theory based on the stereoscopic images received by the lunar rover. The relative positioning accuracy is4.41%with the algorithm, which corrects the error of telemetry data in the CE-3project and proves the best moving distance of the lunar rover and distribution principle of the tie points, at the same time, It solves the problem of the navigation and localization of the rover on the orbit.(2) The exact systematic error calibration algorithm and the process of the telemetry position and attitude data of the lunar rover with the stereo images are proposed in order to obtain hign-precision localization information of the target points. The process includes creating three-dimensional models, building and calculating the network bands of the stereo images by the weighted least squares adjustment, statistical analysis of the constant error of the telemetry position and attitude data. The new systematic error calibration algorithm is tested on the ground simulation experiment of the first Chinese lunar rover and the lunar surface environment, and can be recommended to expand the application fields of the systematic error calibration algorithms. In the indoor experimental field, the system error of feature points can be eliminated by the algorithm in the forward direction with the compensation of the system error of the telemetry data, reaching the accurate level of198.3mm on the three-dimension space. The system error of the telemetry data can be also separated by the algorithm on y coordinate direction of the rover.(3) An exact lunar rover localization algorithm in bundle adjustment used the Dog Leg algorithm and the Levenberg-Marquardt algorithm is proposed in order to obtain the hign-precision location information, especially lacking of adequate control information on the lunar surface environment. According to the bundle block adjustment math models with the Stereoscopic images, the appropriate damping strategy and the trust region strategy are given, so that the positioning of the rover can be calculated. The relative positioning accuracy is4.32%by the algorithm which mutually verifies the farword localization algorithm of the stereoscopic images of the lunar rover and corrects the error of telemetry data with the CE-3project, and solves the convergence problem of the bundle adjust model with the situation of large error of the telemetry data, and overcomes the dependence of the original data to the the unknown parameters by the least squares theory, and expands the idea of the lunar rover localization algorithms based on the bundle adjustment.(4) A telemetry data-supported by stereoscopic images combined the bundle block adjustment algorithm is proposed to get the localization of the lunar rover. Rely on the spatial orientation model of the rover’s telemetry data and the combined block adjustment math models with stereoscopic images, the bundle block adjustment is formulated with the rover’s telemetry data-supported by stereoscopic images, including the system error of the image points, the offset of the position and the rotation angle between rover and the cameras. Then the weight matrix of the observation with the distance value is put forward. Some conclusions are drawn with the experiment results of the CE-3lunar rover. In the outdoor experimental field, the relative positioning accuracy is3.96%by the algorithm which corrects the error of telemetry data in the CE-3project, and reflects the internal stability of positioning and attitude to the recording instruments on the lunar rover. On this basis, the above navigation and localization algorithms of lunar rover are compared and analysed in the paper.In this paper, the research result has been proved on the indoor and outdoor experimental fields of the second stage of CE-3project, and also has been applied to the remote operating system of CE-3, which provides the basic data for the terrain recovery and the path planning of the CE-3lunar rover.