Study On The Deep Space Autonomous Navigation Method And Its Application In Approaching The Small Celestial Bodies

With the implementation of Mars exploration and re-entry Moon and such exploration missions, small celestial bodies exploration becomes the hot trend of new century deep space exploration. The autonomous navigation technology becomes a regarded key technology in the deep exploration fields. With the supports of the Tenth Five-Year 863 Program (Autonomy Technology of Deep Space Exploration and Its Simulation and Demonstration System, the technology of the small celestial exploration and it experiment system). This dissertation deeply studies the autonomous navigation system design theory and builds an autonomous approach navigation real-time hardware-in-the-loop simulation system. The main contents of this dissertation are as follows.First, a process of the autonomous navigation scheme design is presented. It includes the process of the deepspace exploration mission; the selection method of the navigation celestial bodies; the photo strategy and the image processing method; the navigation algorithm design. The orbit dynamics models during the departure, cruise, approaching, circling and landing phases for small celestial body exploration mission are established. The dynamics around the libration points is established too. This paper promotes the position plane method and analyzes the orbit identification based on the measurement model. A new algorithm for autonomous navigation of estimation the little celestial body character based on the single feature point is presented.Second, this paper analyzes the performance of the navigation system using the observability theory, the nonlinearity theory and the stability of the extended Kalman filter theory. Here sums up the observable analysis methods and establishs the inner connection among the severl observability analysis methods. The simplification method of the nonlinearity analysis is proved by a new thought. Based on the stability theory, a numerical algorithm is presented to get the bounds of the initial error and the measurement noise.Then using the above three theories, the performance of the navigation systems are analyzed, which are described by orbit elements and position-velocity respectively. A new criterion is presented to judge whether the stochastic system model simplification is reasonable based on the Kalman filter. Afterwards, importing the Bplane reference frame, a current method for the approach navigation is building on the frame. Then the paper gives an algorithm to select the navigation celestial bodies. Using the observability theory, the advantage of the redundance system is analyzed.The precision of the navigation system is analyzed. To get the future precision, the paper introduces the error transfer theory. Based on the navigation system which estimated the position of a vehicle from the range measurements obtained using three beacons in the same plane, the navigation error bounds were researched. It included the bounds under the bias noise and random measurement noise. At last, importing the error ellipse theory, a method is designed to analyze the navigation accuracy in the B plane.Finally, this paper brings forward an autonomous navigation system scheme for approach mission based on the above research results. And the function, the work flow, and the work mode of autonomous navigation system software are analyzed. The autonomous navigation semi-physical simulation system is built up based on the software including Matlab/Simulink/dSPACE, and the hardware including the star sense, the navigation camera and the asteroid geomorphy simulator. The autonomous navigation system scheme is validated by using the numerical simulation and real-time simulation.