Research On Task Planning Techniques For Imaging Observation Satellites In Collaboration With Signal Detection

According to the nearly30years’ case analysis of local warfare, imagingobservation satellites exhibit characteristics of being the first to be used in every warand being concentrated to the full use. Signal detection gets the information ofelectromagnetic signals and detects the radio signals with long duration and wide range.Signal detection can get the time-efficient intelligence. Although there are abundant anddiverse imaging observation satellites and signal detection resources, the united andcomprehensive application framework is yet to come to coordinate the two systems tofulfill the task of earth observing. This paper focused on the problems of task planningfor imaging observation satellites in collaboration with signal detection, the main worksare listed as the following statements:(1) The mathematical model of constrained multi-objective optimization wasconstructed by researches on the problems of task planning for imaging observation.The complex constraints and relations in these problems are analyzed to search for thework patterns of imaging observation satellites under the guidance of signal detectionsystems. It is of great theoretical and practical significance to build the mathematicalmodel of task planning problem for imaging observation satellites and construct therelated optimization algorithm. Through quantitative analysis and formal representationof task planning problem for imaging observation satellites，the mathematical model ofconstrained multi-objective optimization was founded.Meta-heuristic algorithms are optimization algorithms closely related with themulti-objective combinatorial optimization problems. Scatter search and CellularAutomata are two typical meta-heuristic algorithms. Two multi-objective optimizationalgorithms were designed in this paper by combining these two typical algorithms withgenetic algorithm to solve task planning problems for imaging observation satellites.They are multi-objective scatter search algorithm and multi-objective cellular geneticalgorithm.(2) The scheme of task planning for imaging observation satellites was establishedbased on the multi-objective scatter search algorithm, which came from the researcheson the combination method of the scatter search algorithm and the genetic algorithms.Through analysis of constraint characteristics and modeling of the imaging observationsatellites task planning problem, a novel multi-objective scatter search algorithm isproposed to find the pareto-optimal set of solutions. Three new components, includingan adaptive probability mutation operator based searching strategy, aconstrained-dominance comparator based on number of the constraint violations and asolution combination method based on dual crossover operators are incorporated intothe standard Archive-Based hYbrid Scatter Search (AbYSS) algorithm. Experimental results demonstrate the proposed scatter search algorithm is valid and effective.(3) The optimization solution to the problem of task planning for imagingobservation satellites was built up on the multi-objective cellular genetic algorithms,which originated in the researches on the mixed model of the cellular automataalgorithm and the genetic algorithms. The proposed multi-objective cellular geneticalgorithm relates genetic operators of genetic algorithm to the local rules of cellularautomata. Using basic genetic operators as local rules of cellular automata plays a rolein increasing the possibility of multi-objective cellular genetic algorithm generatingdiverse solutions, which improves the dynamic performance of this system. Theevolution rules of life game and the disaster feedback mechanism are fused togetherwell. The sudden disaster is used to deal with the central cells in inactive state under thecontrol of evolution rules. The feedback process is incorporated into the disastermechanism to reuse the non-dominated solutions obtained currently by the searchsystem. Thus, the performance of task planning system for imaging observationsatellites based on this multi-objective cellular genetic algorithm has been improved.The results of simulation experiments demonstrate that task planning for imagingobservation satellites based on the proposed multi-objective cellular genetic algorithm isvalid and effective.