Research On Autonomous Formation Reconfiguration Techniques For Distributed Satellite Systems

The autonomous control of Distributed Satellite Systems (DSS) is an inevitable trend of space technology, and the autonomous formation reconfiguration is one of the key techniques of autonomous control. An in-depth study on autonomous formation reconfiguration techniques was carried out by theoretical analysis and simulation based on methods and tools from many disciplines. Some valuable conclusions were drawn and shown as follows.1. Under the instruction of Complex Adaptive Systems (CAS) theory, basic frameworks of Agent based DSS autonomous control were put forward. A formalized description of Agent based autonomous DSS was given, which is the theoretical foundation of further study. A simulation based analysis and design method of autonomous DSS was raised, which can offer instructions to autonomous DSS design.2. Design of the autonomous formation reconfiguration system was investigated. Organizational concepts were introduced and a structure-oriented organizational analysis and design method of autonomous DSS was presented. Then an Agent organization of distributed-central structure was constructed for autonomous formation reconfiguration. Agent classes, amounts and capabilities in the organization were given and the positions allocation coordination mechanism was discussed.3. A series of planning and control methods for the autonomous formation reconfiguration of DSS were studied. A path planning method without radial control was put forward, and it is effective on the problems of reconfiguration path planning and positions optimization when radial thruster fails. Thereafter, two reconfiguration schemes for DSS with one satellite fault were presented and solved by the path planning methods with and without radial control. As for the position allocation problem of complex DSS, auctioneering algorithm was selected for autonomous formation reconfiguration based on analysis of several algorithms. Besides the above planning methods, some control schemes were established. A spiral reconfiguration control strategy was raised. It can avoid the optimization cost and fulfill the demands of autonomy. Based on the spiral control method, a stable fuel-time optimal rendezvous orbit of an inspector satellite to its master was given, which can gain trade-off between fuel and time, thus satisfy the demands of the space mission. In addition, Fuzzy Logic Control (FLC) was used in formation reconfiguration. A fuzzy controller was designed for reconfiguration control. The control scheme is robust because it does not depend on the system model. Then a hybrid control strategy of FLC and Linear Quadratic Regulator (LQR) was put forward. This method can make use of the runtime information of DSS, thus it can fulfill the demands of autonomous control. Moreover, a fuzzy LQR control scheme without radial control was established, which provides a new way for formation reconfiguration control when the radial thrust fails.4. Simulation was carried out on DSS autonomous formation reconfiguration. First, a BDICC (Belief, Desire, Intention, Capability, Commitment) Agent model was established which is fit for role-playing. Then Backup Normal Form (BNF) descriptions of the BDICC hybrid Agent were given and Knowledge Query and Manipulation Language (KQML) based communication was discussed. At last, simulation instances were given based on the reconfiguration algorithms and autonomy methods in this thesis. Simulation results show the validity of reconfiguration algorithms and the feasibility of Agent-based autonomy methods.