Study On Mars Atmospheric Entry Trajectory Optimization And Guidance

As the nearest planet from Earth, Mars has the similar nature environment with earth. So, Marsexploration is becoming a hot topic of deep space exploration activities in recent years. Mars is veryfar from Earth and there is large uncertainty in the Mars’ environment, which lead to the Marsexploration missions difficult to achieve. Therefore, the progress of trajectory optimization andguidance method study has an extremely significance. The main results obtained in this dissertationare summarized as follows.The issue of Mars entry trajectory optimization is addressed using three optimization methodsrespectively. Firstly, Gauss pseudo-spectral method is adopted to discrete the Mars entry trajectoryoptimization problem into a nonlinear programming problem. The random vectors are taken as theinitial guess of the optimization variables. The optimal entry trajectory can be obtained by use ofsequential quadratic programming (SQP). Secondly, the control variables are discredited using thecoding technology, and Hybrid algorithm simplex method and Genetic Algorithm are adopted to solvethis problem for comparison. Then, Particle Swarm Optimization is used to get another optimaltrajectory. Lastly, considering all of simulation data, the advantages and disadvantages of these threeoptimization methods are comprehensively analyzed.Because Mars entry vehicles have the initial and system model errors when they arrive at theinterface of Martian atmosphere, active entry guidance is necessary. Firstly, a predictor-correctorguidance method is developed to solve the Mar’s entry guidance problem with different simulationconditions. Then, a multiple sliding mode guidance methods based on RBF Neural Networks isaddressed. The state error is taken as the first sliding mode surface to ensure the tracking errorsapproach to zero in a finial time. The purpose of the second sliding mode is to track the virtual controllaw and approximate the system uncertain parameters using RBF Neural Networks. Finally, thecharacteristics of this guidance method is analyzed by the simulation results.