Research On The Independent Navigation And Filtering Technology Of Deep Space Interplanetary Detector

After entering the 21 st century,countries in the world have further accelerated their exploration in the field of deep space exploration.The number and variety of deep space probes are also increasing rapidly at the same time.While the existence of the “earth-like” feature and the discovery that there used to be rivers on the mars have made the mars exploration the mainstream of the current deep space exploration.Because of the deep space mission of the probe is time-consuming and far from the earth,and the environment is complex and changeable at the same time,the effective autonomous navigation and control as one of the key technologies in the field of deep space detection is particularly important.In addition,the effective application of filtering technology in various navigation methods can greatly improve the precision and autonomy of navigation.Based on deep space exploration of the national 973 Program,the thesis mainly discusses the cruise period of the mars probe.Firstly,this thesis studies the current autonomous navigation schemes and filtering methods.Then this thesis proposes an appropriate autonomous navigation method by analyzing the characteristics of the cruising period.Meanwhile filtering methods are also improved.In the cruise period,the asteroids imaging can be approximated as a point light source,which is the main navigation objects that can observed.Therefore,this thesis uses the observation of the asteroid to obtain the corresponding navigation information,and proposes an autonomous navigation system.In the scheme,the state equation of the four-body model as well as the measurement equation based on the angle of the asteroid is established to complete the CNS/SINS integrated autonomous navigation system.Then the thesis constructs a navigation star database and establishes a set of evaluation index to realize the simulation verification of the navigation scheme.The models of the navigation system and observation are both nonlinear systems,so particle filter that are suitable for nonlinear system should be used for more accurate navigation estimation.Firstly,the resampling technique of particle filter is improved by proposing two stratified resampling methods based on the reward-punishment strategy and the particle optimization strategy.Based on layered particle set,the two resampling methods fully consider the influence of the small weight particles and deal with the particles in different ways,which effectively contain the shortage of particles caused by resampling.The filtering precision is improved while maintaining real-time.Secondly,a one-dimensional nonlinear tracking model is used to simulate the improved particle filter and verify the feasibility of the improved methods.Thirdly,the improved resampling scheme is combined with EKF and UKF to further improve the filtering precision.Finally,the above navigation scheme is simulated and verified by the example of the mars reconnaissance mission in 2018.And the improved EKPF and UPF methods are also applied to the navigation system.At the same time,the experimental data are imported into STK software for visual verification.Then the experimental results show that the proposed navigation system and improved filtering methods have better navigation performance.