Research On Position Algorithms And Performance Analysis For Cooperative Navigation Of Multi-AUVs

As the booster of the ocean exploration for human,autonomous underwater vehicles(AUV)are increasingly used for a variety of underwater operations.However,limited by the special underwater environment,the traditional GPS navigation can not be applied for AUVs,which makes the underwater navigation for AUVs extremely difficult.At the same time,with the multi-robot system research has becoming the hot spot in the world,multi-AUVs system,due to its many advantages over the single platform,have draw increasing attentions.Therefore,this paper studied the naviagation approches for the member of a group of AUVs,which exchanges its own navigation information with one another to improve their individual position estimates.First,we give a brief introduction for some of the key technologies which is needed for the cooperative navigation system for AUVs,including the navigation sensors which is used for the external measurement,the under water communication technology which is need for shairing the navigation information among AUVs and of course we also introduced the traditional underwater navigation techniques which is commonly used for a single AUV.Then,we analyzed the principle of cooperative navigation for MAUVs.we also studied the basic structure of MAUVs and proposed an improved structure.In the meantime we establisehed a basic nonlinear mathematical model for the underwater movement of AUVs and depend on this we extended our research of several cooperative navigation algorithms based on different no-linear approximation methods.For Taylor series approximation of nonlinear systems,we introduced the basic principles and proposed the EKF-based cooperative navigation algorithm.To make full use of the range measurement and reduce the approximation error,we improved this algorithm and proposed the IEKF-based cooperative navigation algorithm.For UT approximation of nonlinear systems,we studied the basic principles and analyzes its theoretical accuracy,According to the mathematical model of cooperative navigation system for MAUVs,we proposed the UKF-based cooperative navigation algorithm.For Stirling polynomial approximation of nonlinear systems,we introduces its basic principles and analyzes the theoretical accuracy,then apply it for cooperative navigation system and get the CDKF based cooperative navigation algorithm.For the Monte Carlo approximation of the nonlinear system,we also analyzes the basic principles and gives the cooperative navigation algorithm based on particle filter.By combining the UKF algorithm with particle filter,we proposed a new cooperativenavigation algorhm based on improved particle filter.For all the algorithm above,to verify the effectiveness of them,simulation experiments were carried out for each of them,comparing with the traditional dead reckoning algorithm.And the analyse of the performance of each algorithm has also been presented.Finally,to further analyze the performance of the cooperative position algorithm discussed above,we compaired the position error of the algorithms based on different approximation methods,according to the result of simulation experiment.Meanwhile,throught the simulation experiments,we studied the effections that the measurement noise and update frequency have on the accuracy of cooperative navigation algorithms and figure out how these factors put their impact on the performace of navigative algorithm.