Pose And Inertial Parameters Estimation Of An Uncooperative Tumbling Space Target

With the exploring of the outer space,the space environment becmoes more and more packed.Some man-made spacecrafts cannot provide further services due to their mal-functions or running out of the fuel.However,these kinds of spacecrafts still occupied the space resources.In addition,since these kinds of spacecrafts cannot provide their information and tumbling freely,they are called the uncooperative space targets.Since these uncooperative space targets have wasted the precious space resources and threatened the environment that other good satellites operated in,it is very important to remove them from the outer space.In order to eliminate these uncooperative space targets,having good estimations of their pose and inertial parameters is very essential.Only when having good prior knowledge of the uncooperative space targets,the following missions can be accomplished.Therefore,doing research on the pose and inertial parameters of the uncooperative space target has vital important theoretical and practical meaning for the future on-orbit service missions.Taking the parameters estimation of uncooperative space targets as researching background,the pose and inertial parameters estimation of an uncooperative space target is studied.The main research contents and research results are as follows:(1)This thesis studied the modeling technique that considers the attitude and orbital models as a whole.First of all,considering the translational and rotational coupled effects,this paper derived the dual quaternions(DQ)based kinematics and dynamics of an uncooperative space target by modeling the parameters in DQ form.Then,in order to avoid the high dimensional problems,this thesis considered the traditional DQ form by utilizing its vector parts to form the dual vector quaternions(DVQ)of each parameter and derived the DVQ based kinematics and dynamics of the uncooperative space target.Under this circumstance,the dual operator which exists in the DQ form models are eliminated.The DVQ based model set up a solid theoretical foundation of the whole thesis.(2)The system observability of the DVQ based model and measurement model are analyzed.For the measuring process,the vision-based sensors are selected to provide the relative pose measurements.For the modeling process,the DVQ based system kinematic and dynamic modeling is considered.Depending on the above-mentioned models,the observabilities of the system are analyzed which has made a solid foundation of the further researches by reveling whether the system is observable,what is the degree of the observability and which linear formation of the states are the mostly observable.These analyses provide solid theoretical foundations for the further researches.(3)The parameters estimating algorithm of the uncooperative space target under high non-linear measurements model are studied.Considering the DVQ based system models and the measurements models,the regularized Newton-Gauss iteration algorithm combined with the Kalman Filtering technique are proposed to make reliable estimations of the uncooperative space target.By utilizing this algorithm,the parameters of the uncooperative space target can be determined accurately.Comparing with the existing algorithms,the proposed one can have faster converging time and higher estimating accuracy.(4)The pose and inertial parameters estimating algorithm under short-term measurement failures is studied.Combining the DVQ based models and measurement model,the innovation based robust Kalman Filter is designed.By utilizing this algorithm,the pose and inertial parameters of the uncooperative space target can be estimated.Meanwhile,another algorithm which depends on two formation flying small satellites is proposed.Under this circumstance,a Federal Kalman Filter technique is utilized to design the robust parameter estimating algorithm when short-term measurements failures occur.Compared with other algorithms,the proposed one can make accurate estimations under short-term measurements failures with lower computational cost.As a result,this algorithm can be accomplished on the platform of the small satellites.(5)The pose and inertial parameters estimating algorithm under long-term measurement failures is studied.Considering the measuring sensors can be affected by the complex lighting environment and sensors failures during measuring the uncooperative space target,the parameters estimating algorithm under long-term measurement failures is proposed.Considering the traditional Kalman Filter based parameter estimating algorithm rely heavily on the measurements,this thesis proposed an artificial neural network(ANN)based parameter estimating algorithm to make reliable parameter estimations of the uncooperative space target under long-term measurement failures.The ANNs designed by this thesis can accomplish the task and have much better results when compared with the traditional Kalman Filter based parameter estimating algorithms.