【摘要】：捷聯(lián)慣性導(dǎo)航系統(tǒng)(SINS,Strapdown Inertial Navigation System)作為自主式水下潛器(AUV,Autonomous Underwater Vehicle)的主要導(dǎo)航方式,在沒有有效輔助的情況下,由于誤差積累引起的捷聯(lián)慣導(dǎo)系統(tǒng)發(fā)散問題,多采用多普勒測(cè)速儀(DVL,Doppler velocity Log)對(duì)其漂移進(jìn)行限制。然而,探測(cè)方法較為粗糙,水下地形復(fù)雜等問題使得某些情況下DVL的探測(cè)范圍無法到達(dá)海底,降低了 SINS/DVL組合導(dǎo)航模式的可行性。在DVL失效,無法得到準(zhǔn)確量測(cè)時(shí),捷聯(lián)慣導(dǎo)系統(tǒng)誤差迅速增大,導(dǎo)航精度大大降低。同時(shí),系統(tǒng)模型失真,噪聲統(tǒng)計(jì)特性不確定,將會(huì)導(dǎo)致卡爾曼濾波精度降低,嚴(yán)重時(shí)會(huì)出現(xiàn)濾波發(fā)散的情況。因此,需要一個(gè)能有效降低漂移的導(dǎo)航方式對(duì)SINS的速度信息和位置信息進(jìn)行校正,并且需要魯棒性較好的濾波器對(duì)該組合導(dǎo)航系統(tǒng)進(jìn)行狀態(tài)估計(jì)。針對(duì)以上問題,本文提出了采用描述AUV運(yùn)動(dòng)的數(shù)學(xué)模型輔助捷聯(lián)慣導(dǎo)的組合導(dǎo)航方法,并且選用漸消記憶卡爾曼濾波和H∞濾波對(duì)模型輔助的組合導(dǎo)航系統(tǒng)進(jìn)行狀態(tài)估計(jì)。本文詳細(xì)介紹和深入研究了以下內(nèi)容:首先,本文分析了傳統(tǒng)的組合導(dǎo)航方式和模型輔助的組合導(dǎo)航方式之間的區(qū)別;介紹了捷聯(lián)慣導(dǎo)的原理、機(jī)械編排、進(jìn)行了誤差分析,并給出了捷聯(lián)慣導(dǎo)的誤差方程。其次,本文根據(jù)AUV運(yùn)動(dòng)的模型及海流對(duì)運(yùn)動(dòng)模型的影響,建立在海流影響下的AUV運(yùn)動(dòng)的數(shù)學(xué)模型,利用其三自由度模型、合外力及力矩的數(shù)據(jù)解算得到AUV的位置信息和速度信息。然后,結(jié)合對(duì)漸消記憶濾波的深入研究提出了改進(jìn)的漸消記憶卡爾曼濾波算法,并將其應(yīng)用到模型輔助組合導(dǎo)航系統(tǒng)中,在模型準(zhǔn)確和不準(zhǔn)確的情況下分別進(jìn)行了勻速直線運(yùn)動(dòng)和變速運(yùn)動(dòng)的仿真,仿真結(jié)果表明漸消記憶濾波算法可以改善模型輔助捷聯(lián)慣性組合導(dǎo)航系統(tǒng)精度,且在模型不準(zhǔn)確時(shí)抑制卡爾曼濾波發(fā)散。最后,采用了魯性更好的H∞濾波算法對(duì)模型輔助捷聯(lián)慣導(dǎo)組合導(dǎo)航系統(tǒng)進(jìn)行狀態(tài)估計(jì),在模型準(zhǔn)確和不準(zhǔn)確的情況下分別進(jìn)行了勻速直線運(yùn)動(dòng)仿真和變速仿真,仿真結(jié)果表明H∞濾波算法不但可以改善模型輔助捷聯(lián)慣導(dǎo)組合導(dǎo)航系統(tǒng)精度還可以提高系統(tǒng)魯棒性,且在模型不準(zhǔn)確時(shí)可以抑制卡爾曼濾波的發(fā)散。本文的研究結(jié)果表明改進(jìn)的漸消記憶卡爾曼濾波在AUV模型輔助捷聯(lián)慣導(dǎo)組合導(dǎo)航系統(tǒng)中的應(yīng)用可以有效的抑制SINS發(fā)散,提高組合導(dǎo)航精度。H∞濾波在模型輔助組合導(dǎo)航系統(tǒng)中的應(yīng)用能夠有效提高組合導(dǎo)航系統(tǒng)精度和魯棒性。該組合導(dǎo)航系統(tǒng)可以作為DVL工作失效時(shí)的備份導(dǎo)航系統(tǒng),且這兩種濾波方式能夠有效抑制模型不準(zhǔn)確的情況下卡爾曼濾波的發(fā)散問題。
[Abstract]:As the main navigation mode of autonomous Underwater Vehicle), sins Strapdown Inertial Navigation System) is a problem of divergence of sins caused by error accumulation in the absence of effective assistance. Doppler velocimeter (DVL) is often used to limit the drift. However, some problems such as rough detection method and complex underwater terrain make the detection range of DVL can not reach the bottom of the sea under some circumstances, which reduces the feasibility of SINS/DVL integrated navigation mode. When DVL fails and cannot be measured accurately, the strapdown inertial navigation system error increases rapidly and the navigation accuracy decreases greatly. At the same time, the distortion of the system model and the uncertainty of the statistical characteristics of noise will lead to the reduction of Kalman filtering accuracy and the occurrence of filtering divergence in serious cases. Therefore, a navigation method that can effectively reduce drift is needed to correct the velocity and position information of SINS, and a robust filter is needed to estimate the state of the integrated navigation system. In order to solve the above problems, this paper proposes a mathematical model to describe the motion of AUV in sins integrated navigation, and uses fading memory Kalman filter and H 鈭，

本文編號(hào)：2194010