外文翻譯--越野車(chē)可控懸架系統(tǒng)綜述（節(jié)選）

1、<p>　　2600單詞，12900英文字符,3900漢字</p><p>　　畢業(yè)設(shè)計(jì)（論文）附錄資料</p><p>　　學(xué) 院 汽車(chē)與交通學(xué)院 </p><p>　　專(zhuān) 業(yè) 交通運(yùn)輸（汽車(chē)電子技術(shù)與檢測(cè)診斷）</p><p>　　班 級(jí) </p>&l

2、t;p>　　學(xué) 號(hào) </p><p>　　姓 名 </p><p>　　指導(dǎo)教師 </p><p>　　2013年 1月 6 日</p><p><b>　　一、英文翻譯</b></p>&

3、lt;p>　　Survey of Controllable Suspension System for Off-road Vehicles </p><p>　　Abstract：The controllable suspension system can improve the performances of off-road vehicles both on road and cross- countr

4、y．So far，four controllable suspensions，that is，body height control，active，semi-active and slow-active suspensions，have been developed．For off-road vehicles，the slow-active suspension and the semi-active suspension which

5、have controllable stiffness，damping and body height are more appropriate to use．For many years，some control methodologies for controllable suspension system</p><p>　　Key Words：control theory；survey；controlla

6、ble suspension；off-road vehicle </p><p>　　The suspension system is an important part of the vehicle， which influences riding comfort and handle-stability greatly．Since the first pair of leaf spring was used

7、in carriage，suspension systems were uncontrollable for a long time and named as passive suspension system．A typical passive suspension system consists of springs，dampers and some control arms， which are once invariable d

8、esigned．But，the roads for vehicle running are various．For the vehicle’s suspension system，different kinds of roads</p><p>　　1 Categories of Controllable Suspension System</p><p>　　The controlla

9、ble suspension system is a general reference of the non-passive suspension system，the antonym of passive suspension should be active suspension But，in the suspension control，the active suspension is a special type of con

10、trollable suspensions．Today，the controllable suspensions can be divided into four categories according to the controlled objects and structures．They are body height control， active，semi-active and slow-active suspensions

11、．</p><p>　　The quarter-vehicle vibration models of these different controllable suspension systems are shown in Fig．1． </p><p>　　1．1 Body Height Control Suspension System </p><p>　　

12、The body height control suspension system is the pioneer of controllable suspension systems．The first body height control suspension system was used in CitrOen DS19 launched in 1955 and made in France．It adopts four cont

13、rollable air springs and can adjust obviously that the active suspension system improves the body height according to run condition and load to guarantee proper wheel travels[1] .</p><p>　　This kind of syste

14、m is the simplest controllable suspension system，usually found in luxurious buses and pickup trucks．A bus equipped with this system can keep the body height unchanging whether it is empty or full．The body height control

15、suspensions are also designed for the pickup truck’s real suspensions，it can keep the body horizontal，whether it links a trailer or not． </p><p>　　The main structural difference between the body height contr

16、ol suspension and the passive suspension is that the former has a height control system additionally，which includes body height sensors，height drives and a controller，as shown in Fig．1(a)． </p><p>　　It aims

17、at controlling the body height only， though the suspension’s stiffness and damping also change in the control process．The inputs mainly include the speed of vehicle and the distance between body and wheel，which are colle

18、cted by speed and displacement sensors．The proper body height data shall be calculated based on a certain control strategy by the body height control system and output to the executing mechanism． </p><p>　　1

19、．2 Active Suspension System </p><p>　　The concept of active suspension was presented early in 1954[2]．Thompson，in 1960’s，consummated its basic structure and control law and proved that so-called ful1-active

20、suspension system could improve the performances of vehicle effectively．Since 1980’s， the research achievements on active suspension had been put into use. Some testing vehicles were built[3] . The experiment for these v

21、ehicles showed obviously that the active suspension system improves vehicle's performance greatly.</p><p>　　The active suspension system consists of sensors,controllers and force actuators，as shown in Fi

22、g．1（b）[4]. And，for driving force actuators，an additional power is necessary．</p><p>　　It adopts the force actuator to replace the customary spring and absorber．The force actuators can be controlled to produc

23、e appropriate forces to support the body，whenever the vehicle runs in any road． </p><p>　　The body and wheel sensors are used to measure the accelerations of the body and wheel and provide these data to the

24、controller．The latter processes these data and outputs some instructions to the force actuator according to predefined control strategy which determines the quality of the active suspension system．</p><p>　　

25、Although the active suspension system has been presented for more than fifty years，it hasn’t been largely commercialized yet up to now．Of course，the technical and economic reasons coexist．</p><p>　　In techno

26、logy，today’s active suspension systems can work well in low frequency band but not in high frequency，since the force actuators seems too stiff to control[5]． </p><p>　　Although some active suspension systems

27、 can operate well up to 70 Hz[6]，they will consume energy very much[7]． One of the methods to reduce the power consumption uses springs and dampers in parallel with the actuators．In addition，it also improves the security

28、 of the active suspension system．But,as a negative result,the system response will be decreased．</p><p>　　In economy，building and operating an active suspension system costs too much．It fatally limits the ac

29、tive suspension systems to be extended． </p><p>　　1．3 Semi-active Suspension System </p><p>　　The semi-active suspension system was presented later but applied earlier to the vehicles than activ

30、e suspension system．The controllable suspension system with adjustable stiffness and damping was introduced in early 1970’s．It almost does not consume energy，since the force actuators which need too much energy are elimi

31、nated．So，it is called as semiactive or no power active suspension system．</p><p>　　The common semiactive suspension systems only control the damping of suspension actively，and some senior semiactive suspensi

32、on systems also control the stiffness．In fact，a semi-active suspension system is just a passive suspension system with controllable damping and stiffness，as shown in Fig．1 (c)．So ，its performances are still not as good a

33、s the active suspension system．</p><p>　　A famous control model of the semiactive suspension system was so-called Sky-Hook Damping Control proposed by Karnopp in 1973[8]．In this model，a supposed inertial dam

34、per，called as Sky-Hook damper，is set between a sprung mass and a virtual fixed Sky-Hook．The force of the Sky-Hook damper proportional to the relative speed of the sprung mass to the Sky-Hook can reduce the jounce of the

35、vehicles．</p><p>　　For the Sky-Hook and Sky-Hook damper are both inexistence in real vehicles，an controlled adjustable damper is set to replace the passive damper between the sprung mass and unsprung mass in

36、 real Sky-Hook model．Theoretically，the damping force should change continuously and in real-time[9]，but it is usually hard to be done in real vehicles．</p><p>　　A control model for the semi-active suspension

37、 systems with discrete adjustable damping and stiffness was proposed by Margolis in 1975[10] ．Several dampers or springs are paralleled，thus，if one or more of them are shut off，the damping or stiffness of the suspension

38、system will change discretely．For it is easy to achieve，the Margolis model have been used in so me cars．</p><p>　　Some semi-active suspension systems are even added the body height control to improve the veh

39、icle performances．This controllable height semi-active suspension system has been used in some luxurious car and SUV recently．Some of them are named as the active suspension by their manufacturer，but they are still diffe

40、rent from the real active suspension．</p><p>　　1．4 Slow-active Suspension System </p><p>　　The slow-active suspension system is presented later but more remarkably．Its essential structure

41、can be regarded as a series of an active suspension system and a passive suspension system，as shown in Fig．1 (d)．As the passive suspension system can isolates high-frequency vibration well，the active suspension system ca

42、n only isolate low-frequency vibration．The force actuators only work in the low-frequency band， power consumption are reduce evidently．Theoretically，the slow-active suspension system </p><p>　　To improve per

43、formances of the slow-active suspension system，the springs and dampers in the system should be controllable．This slow-active suspension system can be regarded as a series of an active suspension and a semi-active suspens

44、ion．</p><p>　　As an unavoidable result，the control system and mechanical structure are more complex． </p><p>　　The performances of the slow-active suspension system are almos

45、t as goo d as the full-active suspension system，and the power consumption is fairly less．Its prospect will be very wel1．</p><p>　　2 Control Methodologies for Controllable Suspension Systems </p><

46、p>　　The control theories for controllable suspension systems grow along with the development of modem control theory．Recently，the typical control strategies include LQG (linear-quadratic-Gaussian)optimal control，model

47、 reference adaptive／self-tuning control， preview control，fuzzy control，neural network control，etc． </p><p>　　2.1 LQG Optimal Control Strategy LQG</p><p>　　For the linear vibration model of the

48、 active suspension systems，the control attempts to minimize the integrated weight of body vertical acceleration，wheel dynamic load and wheel dynamic travel．The objective function of the control system is quadratic．And th

49、e disturbance input from road is a stochastic process that can be deal with as a Gaussian white noise． Thus，the suspension control problem can be regarded as a typical LQG optimal control[13] ． </p><p>　　Ac

50、cording to LQG optimal control strategy，the optimal control force Uo can be defined a U。=KX． (1) In the formula，K is a control feedback gain matrix．X is a state variable matrix e

51、xpressed as </p><p>　　where Z2一Z1 is the wheel dynamic travel，Z2 is the sprung mass speed，Z1一q is the wheel dynamic compression， Z1 is the unsprung mass speed． 。 </p><p>　　2．2 Fuzzy Control St

52、rategy </p><p>　　Differing from the traditional mathematic and control theory，the fuzzy control strategy is based on fuzzy set theory created by an American cybernetics expert，Zadeh，in 1965．It uses lan

53、guage analytical rule like IF-THEN．This comprehensible rule can be easily transformed into an acceptable control algorithm by computer． The fuzzy control has been demonstrated to be able to control the complex or n

54、onlinear systems efficiently． Because the vehicle suspension system is complex，time-varying and </p><p>　　In Fig．5，S is a value predefined by the system， y is the data measured by sensors，е和Δе are the crisp

55、values of system error and its change rate，E and EC are the fuzzy set of е和Δе，U is the fuzzy output of fuzzy inference and u is the crisp output after defuzzifying U ． Fuzzy logic has been seldom used in the off-roa

56、d vehicles．Although related papers on fuzzy control of suspension systems have increased in recent years，most of them stay in theoretical research and simulation．The reason is tha</p><p>　　2．3 Neural Network

57、 Control Strategy </p><p>　　NN( neural network)control has become a new control methodologies，after the research of ANN (artificial neural network) made great Progress in 1980’s．NN is a nonlinear and par

58、allel calculator made of a lot of neural element and their joints．The weights of the different joints are different in NN system.Adjusting the joint weights in NN system is a process called as“study”． </p><p&g

59、t;　　NN has two main study algorithms．one of them is directed study algorithm and another is undirected study algorithm．For example，back-propagation algorithm shown in Fig．6 is a typical direct study algorithm in NN contr

60、ol[20]． The application of NN to vehicle suspension control has a favourable prospect，since NN has powerful study ability and is good at solving the nonlinear and uncertain system problems． Research shows that the

61、active suspension systems using NN control strategy has better p</p><p>　　But，NN likes a black box，its study process is impossible to be observed，and its outputs are hard to be explained．As a result，the reli

62、ability and accept-ability of NN control are weakened． </p><p>　　NN control strategy applied to the real off-road vehicle suspension systems has not been reported yet． However：along with the development of

63、 the NN control methodologies and relative hardware products，the NN controlled suspension systems for the off-road vehicles will be designed． </p><p>　　2．4 Other Suspension Control Strategies </p>

64、<p>　　Some other control strategies can also be used for vehicle suspension， such as traditional PID control[22] ， variable structure sliding-mode control[23] ， random sub-optimal control[24] ， H∞and μ (norm and str

65、ucture singular value)control[ 25-26]，etc． Recently，the control methodologies for vehicle suspension systems trend to combine two original control methods and create a new control method，such as adaptive fuzzy cont

66、rol，NN fuzzy control，fuzzy PID control，NN PID control，sliding-mode PID </p><p><b>　　二、中文翻譯</b></p><p>　　越野車(chē)可控懸架系統(tǒng)綜述</p><p>　　摘要：可控懸架可以使越野車(chē)輛兼顧越野行駛和道路行駛.可控懸架可分為車(chē)高控制、主動(dòng)、半主動(dòng)

67、和慢主動(dòng)懸架.對(duì)于越野車(chē)輛而言,慢主動(dòng)懸架最適合實(shí)際應(yīng)用.可控懸架控制方法隨著現(xiàn)代控制理論的發(fā)展而發(fā)展,并出現(xiàn)了一些由多種原有控制理論相互結(jié)合而成的新控制方法.目前,在軍用和民用越野車(chē)輛上,可控懸架的研究與開(kāi)發(fā)仍處于試驗(yàn)階段。</p><p>　　關(guān)鍵詞：控制理論 ； 可控懸架 ； 越野車(chē)輛 ；主動(dòng)懸架 </p><p>　　懸掛系統(tǒng)是汽車(chē)的重要組成部分， 其對(duì)平順性（乘坐舒適性）和安全性

68、（操縱穩(wěn)定性） 影響很大。 由于第一對(duì)鋼板彈簧是用于車(chē)箱 ， 懸架系統(tǒng)很長(zhǎng)一段時(shí)間內(nèi)是不可控制的而命名為被動(dòng)懸架系統(tǒng)。傳統(tǒng)的被動(dòng)懸架主要由彈性元件、阻尼元件和控制元件組成 ， 這種設(shè)計(jì)一直不變。 但是，車(chē)輛運(yùn)行的道路有所不同。 相對(duì)不同類(lèi)型的道路車(chē)輛懸架的設(shè)計(jì)有所不同。從而，不同的懸掛參數(shù)，如剛度，阻尼和車(chē)身高度，設(shè)計(jì)時(shí)應(yīng)盡量減少來(lái)自地面的影響和身體的震動(dòng)。然而，裝配著被動(dòng)懸架系統(tǒng)的傳統(tǒng)車(chē)輛不可能在不同的道路上都良好工作。對(duì)于越野車(chē)，多

69、變的道路與固定懸架之間的沖突比道路車(chē)輛更為突出。越野車(chē)大部分仍采用被動(dòng)式懸掛系統(tǒng)，具有較高剛度、阻尼和車(chē)身，能克服崎嶇道路。然而，這些車(chē)輛的乘坐舒適性和操作穩(wěn)定性比道路車(chē)輛的差?；谶@些原因，可控懸架系統(tǒng)應(yīng)是車(chē)輛的最佳選擇，特別是對(duì)于越野車(chē)。</p><p>　　1 可控懸架系統(tǒng)的分類(lèi) </p><p>　　可控懸架系統(tǒng)可做為非被動(dòng)懸架系統(tǒng)的一個(gè)參考，主動(dòng)懸架與被動(dòng)懸架相反。 <

70、;/p><p>　　但是，在懸架控制上，主動(dòng)懸架是一種特殊類(lèi)型可控懸架。現(xiàn)今，可控懸架根據(jù)控制對(duì)象和結(jié)構(gòu)可分為四類(lèi)：身體的高度控制，主動(dòng)懸架，半主動(dòng)和慢主動(dòng)懸架。</p><p>　　四種不同可控懸架系統(tǒng)的車(chē)輛振動(dòng)模型如圖1所示。</p><p>　　1．1 車(chē)身高度控制懸掛系統(tǒng) </p><p>　　車(chē)身高度控制懸掛系統(tǒng)是可控懸架系統(tǒng)

71、的先驅(qū)。第一個(gè)車(chē)身高度控制懸掛系統(tǒng)是雪鐵龍于1955年在法國(guó)制造并在DS19上使用。它采用了四個(gè)可控制和調(diào)整的空氣彈簧，主動(dòng)懸架系統(tǒng)明顯的提高了行駛條件，且根據(jù)負(fù)載來(lái)保證車(chē)輪機(jī)構(gòu)保持適當(dāng)?shù)母叨取?lt;/p><p>　　這種系統(tǒng)是最簡(jiǎn)單的控制懸掛系統(tǒng)，通常用在豪華車(chē)和皮卡。公共汽車(chē)裝這種系統(tǒng)可以保持身體高度不變，無(wú)論空車(chē)還是滿載。車(chē)身高度控制懸掛也是為皮卡的真正懸架而設(shè)計(jì)，它可以保持身體的水平，無(wú)論連結(jié)拖架與否。&l

72、t;/p><p>　　主要機(jī)構(gòu)之間的高度控制懸架和被動(dòng)懸架結(jié)構(gòu)上的差異是，前者具有高度控制系統(tǒng)外，其中包括身高傳感器，驅(qū)動(dòng)器和控制器的高度，如在圖1（a）所示。</p><p>　　其目的只是控制車(chē)高，盡管懸架的剛度和阻尼在控制過(guò)程也改變。這些數(shù)據(jù)主要包括車(chē)輛的速度和車(chē)身與車(chē)輪之間的距離，由速度和位移傳感器收集。正確的身高數(shù)據(jù)的計(jì)算方法應(yīng)根據(jù)一定的控制模式，身高控制系統(tǒng)和輸出執(zhí)行機(jī)構(gòu)。<

73、/p><p>　　1．2 主動(dòng)懸架系統(tǒng) </p><p>　　早在1954年Thompson便提出主動(dòng)懸架概念，完善其基本結(jié)構(gòu)和控制，并證明了全主動(dòng)懸掛系統(tǒng)可以提高車(chē)輛有效性能。自1980年代以來(lái)，主動(dòng)懸架的研究成果便投入使用。并制造了一些測(cè)試車(chē)輛。這些車(chē)輛的試驗(yàn)表明，主動(dòng)懸架系統(tǒng)明顯地提高汽車(chē)的性能。</p><p>　　主動(dòng)懸架系統(tǒng)包括傳感器，控制器和驅(qū)

74、動(dòng)執(zhí)行器， 如圖1（b）所示。另外，驅(qū)動(dòng)執(zhí)行器需要一個(gè)額外的動(dòng)力。 它用驅(qū)動(dòng)執(zhí)行器取代傳統(tǒng)彈簧和減震器。當(dāng)車(chē)輛在任何道路上行駛時(shí)驅(qū)動(dòng)執(zhí)行器可產(chǎn)生一定的力來(lái)支撐車(chē)身。</p><p>　　車(chē)身和車(chē)輪傳感器用來(lái)測(cè)量車(chē)身和車(chē)輪的加速度，并提供數(shù)據(jù)到控制器。傳感器處理這些數(shù)據(jù)后輸出一些指令給執(zhí)行器，由已確定的控制策略決定主動(dòng)懸架系統(tǒng)的質(zhì)量。 </p><p>　　雖然主動(dòng)懸架系統(tǒng)已經(jīng)出現(xiàn)

75、了50多年，但至今仍未實(shí)現(xiàn)量產(chǎn)。當(dāng)然，技術(shù)和經(jīng)濟(jì)原因都有。 </p><p>　　技術(shù)方面，如今的主動(dòng)懸架系統(tǒng)可以在低頻段很好的工作但高頻不能，因?yàn)轵?qū)動(dòng)執(zhí)行器太穩(wěn)固而難以控制。雖然一些主動(dòng)懸架系統(tǒng)可以很好地運(yùn)轉(zhuǎn)到70赫茲，但消耗很多能源。一個(gè)降低功率消耗和阻尼的方法是采用平行彈簧和阻尼器。此外，還提高了主動(dòng)懸架系統(tǒng)的安全。但，不足的是，該系統(tǒng)的響應(yīng)將下降。在經(jīng)濟(jì)，裝配主動(dòng)懸架系統(tǒng)的成本過(guò)大，極大地限制了主動(dòng)懸

76、架系統(tǒng)的推廣。</p><p>　　1．3 半主動(dòng)懸架系統(tǒng) </p><p>　　與之前的主動(dòng)懸架車(chē)輛相比，半主動(dòng)懸架系統(tǒng)的應(yīng)用提出了更高要求。在70年代初可調(diào)剛度和可調(diào)阻尼引入了可控懸架系統(tǒng)。</p><p>　　它幾乎不消耗能源，因?yàn)榇蓑?qū)動(dòng)執(zhí)行器需要太多的能源已被淘汰。因此，它被稱為是半主動(dòng)或無(wú)動(dòng)力主動(dòng)懸架系統(tǒng)。</p><p&g

77、t;　　常見(jiàn)的半主動(dòng)懸架系統(tǒng)僅控制主懸架阻尼，而一些高級(jí)半主動(dòng)懸掛系統(tǒng)同時(shí)控制剛度。實(shí)際上，一個(gè)半主動(dòng)懸架系統(tǒng)僅是一個(gè)被動(dòng)懸架系統(tǒng)與可控阻尼和剛度的結(jié)合，如圖1（c）所示。所以，它的性能不如主動(dòng)懸架系統(tǒng)的好。</p><p>　　Karnopp在1973年提出了著名的半主動(dòng)懸架系統(tǒng)的控制模型，所謂的天棚阻尼控制模型。在此模型中，一個(gè)假定的慣性阻尼器，叫作天棚阻尼器。它設(shè)置在一個(gè)彈簧質(zhì)量和一個(gè)虛擬的固定天棚之間。與

78、天棚的彈簧質(zhì)量的相對(duì)速度成正比的天棚阻尼器的力量可以減少車(chē)輛的顛簸。</p><p>　　由于天棚與天棚阻尼器都不存在真正的車(chē)輛中，因此，在真正的天棚模式1中，可以用一個(gè)可控制可調(diào)節(jié)阻尼器代替彈簧質(zhì)量和非彈簧阻尼器之間的被動(dòng)阻尼器。從理論上來(lái)說(shuō)，阻尼是連續(xù)不斷變化的，但這同時(shí)很難在真正的車(chē)輛中實(shí)現(xiàn)。</p><p>　　1975年，馬戈利斯提出了一種有離散可調(diào)阻尼和剛度的半主動(dòng)懸架系統(tǒng)的控

79、制模式。幾個(gè)阻尼器或彈簧并聯(lián)，這樣，如果它們中有一個(gè)或者幾個(gè)關(guān)閉了，懸架系統(tǒng)的阻尼和剛度將離散變化。由于這樣很容易實(shí)現(xiàn)，馬戈利斯模型已被應(yīng)用于一些汽車(chē)。 </p><p>　　一些半主動(dòng)懸架系統(tǒng)更是增加了車(chē)身高度控制，以提高車(chē)輛性能。最近這個(gè)高度可控半主動(dòng)懸掛系統(tǒng)已經(jīng)被用于一些豪華車(chē)和SUV。其中一些被它們的制造商命名為主動(dòng)懸架，但是它們?nèi)匀慌c真正的主動(dòng)懸架有所不同。 </p><p&

80、gt;　　1．4 慢主動(dòng)懸架系統(tǒng) </p><p>　　慢主動(dòng)懸架系統(tǒng)較晚提出，但更引人關(guān)注。它的基本結(jié)構(gòu)可以看作是一系列主動(dòng)懸架系統(tǒng)和被動(dòng)懸架系統(tǒng)，如圖1（d）所示。被動(dòng)懸架系統(tǒng)能很好地隔離高頻震動(dòng)，而主動(dòng)懸架系統(tǒng)只能隔離低頻震動(dòng)。 驅(qū)動(dòng)執(zhí)行器只能在低頻傳動(dòng)帶工作，功率消耗明顯降低。理論上，慢主動(dòng)懸架系統(tǒng)的反應(yīng)速度仍然比真正的主動(dòng)懸架系統(tǒng)慢得多，這就是它被這樣命名的原因。但也發(fā)現(xiàn)它其他一些名稱，如

81、窄帶主動(dòng)式懸掛系統(tǒng)或有限帶式懸架系統(tǒng)。與此相反的，真正的主動(dòng)懸架系統(tǒng)通常稱為全主動(dòng)式懸掛系統(tǒng)或帶寬式主動(dòng)懸架系統(tǒng)。 </p><p>　　為了提高慢主動(dòng)懸架系統(tǒng)的性能，系統(tǒng)中的彈簧和阻尼器應(yīng)該是可控制的。這種慢主動(dòng)懸架系統(tǒng)可以當(dāng)作一系列主動(dòng)懸架和半主動(dòng)懸架。</p><p>　　作為一個(gè)不可避免的結(jié)果，控制系統(tǒng)和機(jī)械結(jié)構(gòu)更為復(fù)雜。 </p>

82、;<p>　　慢主動(dòng)懸架系統(tǒng)的性能幾乎和全主動(dòng)懸架系統(tǒng)的性能一樣好，而且電力耗費(fèi)也相當(dāng)少。它的前景將非常好。</p><p>　　2 可控懸架系統(tǒng)的控制方法 </p><p>　　可控懸架系統(tǒng)的控制理論與現(xiàn)代控制理論的同時(shí)發(fā)展。最近，典型控制策略包括LQG（線性二次高斯）最優(yōu)控制，模型參考自適應(yīng)/自校正控制，預(yù)見(jiàn)控制，模糊控制，神經(jīng)網(wǎng)絡(luò)控制等 </p><

83、;p>　　2.1 LQG最優(yōu)控制策略 </p><p>　　為了主動(dòng)懸架系統(tǒng)線性振動(dòng)模型，控制器試圖最小化身體垂直加速度，車(chē)輪動(dòng)載荷和車(chē)輪跑偏。 該控制系統(tǒng)的目標(biāo)函數(shù)為二次函數(shù)。 來(lái)自道路的不確定干擾可以作為處理高斯白噪聲。因此，懸架控制問(wèn)題可以當(dāng)作一個(gè)典型的LQG最優(yōu)控制。 據(jù)LQG最優(yōu)控制策略，最優(yōu)控制力Uo可以被定義為 U。=K

84、X． (1) 在公式中，K是一個(gè)控制反饋增益矩陣。X是一個(gè)狀態(tài)變量，表示為 </p><p>　　其中Z2一Z1是車(chē)輪偏移量，Z2是簧載質(zhì)量的速度，Z1一q是車(chē)輪壓縮量，Z1是簧下質(zhì)量速度。 </p><p>　　2．2 模糊控制策略 </p><p>　　不同于傳統(tǒng)的數(shù)學(xué)和控制理論， 模糊控制策略是美國(guó)控制論專(zhuān)家Za

85、deh于1965年根據(jù)模糊論創(chuàng)建的理論。 它使用假設(shè)的語(yǔ)言分析規(guī)則。 這種易理解的規(guī)則可以轉(zhuǎn)化成計(jì)算機(jī)可以接受的控制算法。 模糊控制已被證明能夠有效控制復(fù)雜的或非線性的系統(tǒng)。 因?yàn)檐?chē)輛懸掛系統(tǒng)是復(fù)雜的，變化的和非線性的系統(tǒng)， 模糊控制應(yīng)用到這種系統(tǒng)顯然比傳統(tǒng)的控制更適當(dāng)?shù)摹?但是，除模糊控制的優(yōu)點(diǎn)外，還應(yīng)看到其規(guī)則的不足。 而在一般情況下，沒(méi)有大量的計(jì)算和實(shí)驗(yàn)?zāi)：刂埔?guī)則難以理解。模糊控制的實(shí)現(xiàn)可以分為三個(gè)步驟。 首先，精

86、確的輸入被模糊化， 第二，使用模糊的推斷得出模糊的結(jié)果， 第三，將結(jié)果清晰輸出，如圖5所示。 </p><p>　　在圖5，S是由系統(tǒng)預(yù)定義的值， y是傳感器測(cè)量的數(shù)據(jù)， е和Δе是系統(tǒng)誤差及其變化率的精確值， E和EC是模糊集е和Δе， U為模糊輸出，u為清晰輸出。 模糊邏輯已經(jīng)很少使用在越野車(chē)上。 雖然與懸掛系統(tǒng)模糊控制相關(guān)論文在近年有所增加， 但大多數(shù)停留在理論研究和模擬上。 原因是，獲取模糊規(guī)

87、則是非常困難的，需要大量的計(jì)算和實(shí)驗(yàn)驗(yàn)證。 </p><p>　　2．3 神經(jīng)網(wǎng)絡(luò)控制策略 </p><p>　　NN（神經(jīng)網(wǎng)絡(luò)）控制已成為一個(gè)新的控制方法，在1980年接著ANN（人工神經(jīng)網(wǎng)絡(luò)）的研究后有了巨大進(jìn)步 。 神經(jīng)網(wǎng)絡(luò)是一種由很多神經(jīng)元和接頭構(gòu)成的非線性并行計(jì)算機(jī)。 在神經(jīng)網(wǎng)絡(luò)系統(tǒng)不同重量折接頭有所不同。在NN系統(tǒng)中調(diào)整接頭重量的過(guò)程叫“學(xué)習(xí)”。 神經(jīng)網(wǎng)絡(luò)有

88、兩個(gè)主要的研究算法。 其中之一是針對(duì)研究算法，另一個(gè)是無(wú)向研究算法。例如，在圖6所示，在神經(jīng)網(wǎng)絡(luò)控制中反向傳播算法是一個(gè)典型的的直接研究算法。 神經(jīng)網(wǎng)絡(luò)在汽車(chē)懸架控制上的應(yīng)用具有良好的前景，因?yàn)樯窠?jīng)網(wǎng)絡(luò)的智能化，且能很好解決非線性和不確定系統(tǒng)的問(wèn)題。研究表明，采用神經(jīng)網(wǎng)絡(luò)控制策略的主動(dòng)懸架系統(tǒng)比使用傳統(tǒng)LQG最優(yōu)控制策略的好。 </p><p>　　但是，神經(jīng)網(wǎng)絡(luò)像一個(gè)黑盒子，不可能看到它的動(dòng)作過(guò)

89、程，其輸出也很難解釋。 因此，神經(jīng)網(wǎng)絡(luò)控制的可靠性和可接受性被削弱。 </p><p>　　尚未有把神經(jīng)網(wǎng)絡(luò)控制策略應(yīng)用到實(shí)際越野車(chē)懸掛系統(tǒng)的報(bào)道。 然而：隨著神經(jīng)網(wǎng)絡(luò)控制的方法和相關(guān)硬件產(chǎn)品的開(kāi)發(fā)， 越野車(chē)使用的神經(jīng)網(wǎng)絡(luò)控制懸掛系統(tǒng)將被設(shè)計(jì)出來(lái)。 </p><p>　　2．4 其他懸架控制策略 </p><p>　　其他一些控制策略也可以用于汽車(chē)懸架，