Open Access Research on Pneumatic Servo Control for Double-Cylinder Collaborative Loading Based on Neural Network

Size: px

Start display at page:

Download "Open Access Research on Pneumatic Servo Control for Double-Cylinder Collaborative Loading Based on Neural Network"

Todd Gray
5 years ago
Views:

1 Sed Orders for Rerits to 51 The Oe Electrical & Electroic Egieerig Joural, 014, 8, Oe Access Research o Peumatic Servo Cotrol for Double-Cylider Collaborative Loadig Based o Neural Network Xu Jiaqiag * Deartmet of Mechaical ad Electrical Egieerig, Beijig Vocatioal College of Labor ad Social Security, ad Beijig Exerimetal Techical School, No.5, Huixidog Street, Chaoyag District, Beijig, 10009, Chia Abstract: The eumatic cotrol system for double-cylider collaborative loadig is extesively alied i the idustry. However, comared to the hydraulic system, it is difficult for the eumatic system for double-cylider collaborative loadig to get satisfactory effect due to strog o-liearity ad low rigidity. To imrove robustess of the double-cylider collaborative loadig system ad icrease its cotrol recisio, the double-looed PID hybrid cotroller ad doublevariable eural etwork PDI cotroller are desiged. The comarative test is erformed for them i research. The research results idicate that the eutral etwork cotrol is strogly adative to the ukow, ucertai ad o-liear features of the cotrolled objects, ca effectively overcome ufavorable iflueces of the o-liear factors o the system, ad imrove the cotrol quality. Keywords: Peumatic, double cylider, collaborative loadig, servo cotrol, PID, eural etwork. 1. INTRODUCTION As oe maer of the eumatic force servo cotrol, the double-cylider eumatic collaborative loadig is extesively alied. Geerally the double-cylider collaborative loadig system idicates the ositio loadig system i most cases. The load motio is out of cotrol i this system. The force is uiquely cotrolled. The focus is to solve the multi-residual force elimiatio roblem i research, so this roblem belogs to sigle-variable cotrol system. The double-cylider collaborative loadig studied i this aer is differet from the above ositio loadig system ad belogs to the double-cylider motio collaborative loadig cotrol system. Whe this system works, the motio seed of the load ad clamig force of double cyliders o the load should be cotrolled simultaeously whe the system is workig, amely the load force ad motio seed are cotrolled. This system belogs to double-variable cotrol system. The left ad right cylider will move while imosig the force o the load i cotrol. The load force should chage accordig to the give law ad the load should move alog the give seed curve. It is comlex to cotrol the double-cylider motio collaborative loadig system ad the mai roblem is to solve the mutual coulig roblem of double variats ad ifluece of differet o-liear cotrol factors of the system o cotrol.. ESTABLISHMENT OF DOUBLE-CYLINDER LOADING SYSTEM.1. System Scheme The block diagram of the double-cylider motio collaborative loadig exerimetal system is show as the Fig. (1). This system is the eumatic servo cotrol system for the double closed-loo valve-cotrolled cylider ad a idustrial comuter is used to cotrol ad moitor the eumatic loadig i real time. The system icludes the load force cotrol closed-loo circuit ad seed cotrol closed-loo circuit. The cotrol comoets iclude the roortioal ressure valve ad roortioal directio vale. The executive comoets iclude the loadig cylider ad mai eumatic cylider. The load is clamed by two cyliders. The force sesor is istalled o the load to detect the loadig force o the load. The seed sesor is istalled o the master cylider to detect the motio seed of the cylider isto rod, amely load motio seed. The cotrol variat icludes the loadig force F of the load ad motio seed V of the load. Whe the system is oeratig, the comuter will erform double-cylider collaborative cotrol over the master cylider ad loadig cylider. The seed cotrol force should be coordiated with the force cotrol. The master cylider should be coordiated with the loadig cylider. The motio law of the cotrolled objects should be esured ad the force law of the cotrolled object should be also esured, so the double-variat cotrol of the load force ad motio is realized. The loadig cylider is used to comlete the force cotrol exerimet ad rovide the exteral force load for ositio cotrol ad seed cotrol / Betham Oe

2 Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume F V Proortioal Y d + f() e Cotroller ressure valve - Y Proortioal Master Y d + g() e Cotroller Directioal valve cylide - r Y Seed Feedback Fore sesor Loadig cylider Fig. (1). Electric Servo Cotrol System of Double Closed-loo Valve-cotrolled Cylider. Fig. (). Structure Schematic of Double-cylider Collaborative Force Servo Loadig Exerimetal System. exerimet. The master cylider ca be used to comlete ositio cotrol ad seed cotrol exerimet ad assist the loadig cylider to comlete the force loadig cotrol exerimet... System Structure The structure of the double-cylider motio collaborative loadig exerimetal system [1] is show as the Fig. (). This system icludes the eumatic device ad detectio cotrol device. The eumatic device icludes the roortioal directioal valve, roortioal ressure valve, rod cylider (loadig cylider), rod-less cylider (master cylider) ad differet accessories. The detectio cotrol device icludes the idustry cotrol comuter, shift sesor, force sesor, sigal coditioig circuit ad cotrol software. The cylider i the seed cotrol circuit (called as master cylider) should be low-frictio rod-less cylider, which features high stability, big rigidity, strog bearig caability, low frictio ad symmetric two-cavity feature ad is suitable for research o seed ad ositio servo cotrol. The DGP PPV-A-B from Germay FES- TO cylider is selected i this system. The cylider diameter is 40mm ad the stroke is 600mm. The loadig cylider ad load cylider (amely rodless cylider) will collaboratively move i double-cylider collaborative loadig cotrol, so the structural arameters of the loadig cylider should be similar to it of the master cylider. Sigle isto rod double-actio cylider roduced by Jaa SMC is selected i this system accordig to the structure requiremet. The model is DNC PPV-A. The cylider strake is 600mm ad the diameter is 50mm. The cotrol valve i the seed closed-loo cotrol circuit should be a eumatic servo valve with the low ower cosumtio, high dyamic resose frequecy ad electric feedback fuctio. The MPYE roortioal directioal valve from Festo is referred i this system accordig to the erformace requiremets. The cotrol valve i the load force closed-loo cotrol circuit should be a roortioal ressure valve with better dyamic erformace, embedded feed-back circuit ad better liearity. The VEP311-1 electric roortioal

3 514 The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 Xu Jiaqiag Fig. (3). Equivalet Flow Grah of Proortioal Directioal Valve-cotrolled Cylider Model. ressure valve roduced by Jaa SMC is selected i this system..3. Software Desig [] The Lab Widows/CVI software develomet latform is used to develo the system cotrol rogram uder Widows eviromet. This software ca erform data collectio, data aalysis, data dislay ad real-time cotrol for the system. The corresodig cotrol fuctios are divided ito differet modules i develomet of the cotrol rogram, e.g. collectio rogram, data rocessig rogram, dislay rogram, outut cotrol rogram, huma-machie cotact rogram, ad the the rogram modules are develoed. The double-thread structure is used i rogram desig. The mai thread ca realize arameter setu, cotrol method selectio, OS commad ad sigal source geeratio fuctios. The secodary thread ca realize data collectio, digital filterig, cotrol algorithm ad cotrol outut fuctios. 3. MATHEMATIC MODEL OF SYSTEM[3] E.g. for the motio of the system load from left to right, the double-cylider collaborative loadig system is dyamically aalyzed to establish a mathematical model. Show as the Fig. (), the master cylider (rod-less cylider) is assumed as cylider A ad the load cylider (rod cylider) is assumed as the cylider B Dyamic Descritio of Cylider A The equivalet flow grah of the roortioal directioal valve-cotrolled cylider model is show as the Fig. (3). Show as the Fig. (3), the mass flow flowig ito valve s two cavities is described as follows: q mi q mis! q mie (1) Mass flow equatio at the valve ort: q miq C d A q! H RT H! f " $ # D H ' & () Dyamic equatio of valve ort:! i V i + i! V i k km i RT i Based o the equatio (1) ad (3), the ressure chage equatio of two cavities of cylider A is obtaied as follows: ) " A qis (u)! s! f i, + $ krc d T '. i + # s &. + RT s ( T. ( k i V! i + " s! A T qie (u)! i! f e. + $ i # ' i &.! i * - (4) V i Pisto force equatio of cylider A: M a!!x ( 1! )A! B a!x! f a sg!x! F a (5) 3.. Dyamic Descritio of Cylider B Force equilibrium equatio of Sool i Proortioal ressure valve: m V!!x V + B V!x V + K S x V + f V sg!x V K ub u b! f A f (6) Positio force equilibrium equatio of cylider B: M b!!x F b + A 3 3! A e ( A 3! A 4 )! B b!x! f b sg!x (7) Cylider ressure equatio of right cavity of cylider B: k! 4 V (RT 4 A S! A x 4 b System State Equatio [4] 1) State variable: (3) C d A q RT " z ( )! 4 A 4!x) (8) x 1 x, x!x, x 3 x V, x 4!x V, x 5 1, x 6, x 7 4 (9) ) System state equatio:!x 1 x (10)

4 Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume Fig. (4). Outut Equatio Sub-system Simulatio Module Diagram.!x!x 3 x 4 (x 5! x 6 )A! ( A 4 x 7! A 3 3 ) + e ( A 4! A 3 )!(B a + B b )x! ( f a + f b )sg x M a + M b (11) (1)!x 4 1 m V (K ub!u b " P f! A f " B V x 4 " K S x 3 " f V sg x 4 ) (13) x 5 krc d A R Ts!x 6 q1s x 51 e ( ua) s f A ( u ) 1 1 f ka x q e s x 5 1 V10 + ASa + Ax1 krc da R / T S [A qs (u a ) s! f ( x 6 5 ) " A qe (u)!! f ( e x 6 ) + ka x ] V A S a " A P x 1 k!x 7 V A S! A x 4 b 4 1 3) Outut: v, F or F a b 1 (14) (15) (RT 4 C db A q RT " z ( )! 4 A 4 x ) (16) v x!x (17) F a (P 1! P )A P! B a x! f a sg x! M a!x (18) F b M b!x! A A 4 x 7! e ( A 3! A 4 ) + B b x + f b sg x (19) 4) Iut u ad a ub 4. SIMULATION RESULTS AND ANALYSIS [5] To aalyze the system features, the double-cylider collaborative loadig system is simulated uder the static state is simulated. The, 1 ad block are established 4 accordig to the system state equatio (14), (15) ad (16) i the system simulatio. The iuts iclude the valve oeess ad cylider shift. The oututs iclude, 1 ad. Based o the equatio (13), the system acceleratio is obtaied. After oe-order itegral ad two-order 4 itegral, the system seed ad shift ca be obtaied. After the iitial values of seve state variats are set, the state variat value ca be comuted at ay time i the motio simulatio. Based o the equatio (18), the simulatio for the load force F ca be erformed. The simulatio module diagram is show as the Fig. (4). The simulatio curve of the system oe-loo cotrol ad closed-loo cotrol is show as the Fig. (5 ad 6). The trasmissio medium of the eumatic cotrol system is the comressed air ad its rigidity ad stability is worse much tha it of the hydraulic medium, so the closedloo cotrol is used ad a roer cotrol method is used to imrove the system resose. It is difficult to model a oliear system, so the moder cotrol theory is difficult to get satisfactory effect. We roose the eural etwork

5 516 The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 Xu Jiaqiag Fig. (5). 400N Ste Resose Simulatio Curve for Oe-loo Cotrol. Fig. (6). 400N Ste Resose Simulatio Curve for Close-loo Cotrol. cotrol strategy to study the eumatic servo cotrol system for double-cylider collaborative loadig. 5. DESIGN OF DOUBLE-CYLINDER PNEUMAT- IC FORCE SERVO CONTROLLER BASED ON NEURAL NETWORK PID 5.1. Neural Network Double-Variat Cotrol Structure [6] To reach better cotrol effect, the double-cylider eumatic force servo system is cotrolled by usig the cotroller based o the eural etwork ad the eural etwork double-variable cotrol structure is show as the f Fig. (7). Assumig that the cotrolled object of the doublecylider collaborative loadig system is P, which is the double-iut double-outut (DIDO) system with cotrolled variat force ad seed, the cotroller NFC icludes two PID eural sub-etworks. Three euros at the hidde layer of the sub-etwork are roortio P, itegral I ad differetial D uit. Assumig that two etworks are F ad V, each sub-etwork icludes two iuts ad oe outut. Oe iut of the sub-etwork F corresods to the give value r1 of cotrolled variat force. Aother iut corresods to the system outut y 1 Oe iut of the sub-etwork v corresods to the give value r of the cotrolled variat seed. Aother iut corresods to the system outut y, u1, ad u are two etwork oututs v 1, ad v are the disturbace imosed o the system outut. This PID eural etwork is a three-layer forward feedback etwork cosistig of the rocessig uits with the geeral sigmoid fuctio feature ad rovides ay oliear maig caability from iuts to oututs i L

6 Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume Fig. (7). Neural Network Double-variat Cotrol Structure. meaig. Whe BP algorithm is used for learig ad traiig, the maig from iuts to oututs of the cotrol system (icludig cotrolled object) is comleted accordig to the requiremets of the rule fuctio E! E mi. The PID eural etwork makes the decoulig cotrol erformace reach the requiremet i the weight adjustmet traiig ad learig. 5.. Cotroller Algorithm [7] 1) Iut ad outut a. Iut ad outut of the eutros at the iut layer: R [r 1, y 1,r, y ] [r sj ] (0) s: sub-etwork SN, s1, : j: SN of the eutros at sub-etwork iut layer, j 1, b. Iut ad outut of hidde eutros Iut of i th ode of s sub-etwork hidde layer: x si (k)! 1 w sij r sj (k), i 1,,3 (1) j1 Outut qsi ( k ) of P, I ad D ode at s sub-etwork hidde layer. # x s1 (k),!1 " x s1 (k) " 1 q s1 (k) $ 1, x s1 (k) > 1!1, x s1 (k) <!1 & () # q s (k!1) + x s (k),!1" q s (k) " 1 q s (k) $ 1, q s > 1!1, q s <!1 & # x s3 (k)! x s3 (k!1),!1 " q s3 (k) " 1 q s3 (k) $ 1, q s3 > 1!1, q s3 <!1 & c. Iut ad outut of eutros at outut layer (3) (4) The iut is the weighted sum of oututs of odes at the hidde layer of two sub-etworks: 3 x O 1 (k)!! w s1i q si (k) (5) s1 i1 3 x O (k)!! w si q si (k) (6) s1 i1 w shi : weight from the hidde ode I to outut ode h i s th sub-etwork. d. Network outut: u 1 (k) x 1 O (k) (7) u (k) x O (k) (8)

7 518 The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 Xu Jiaqiag Fig. (8). Force PID Cotrol Ste Resose Curve. ) Learig algorithm The PID eural etwork cotroller ad multi-variat object are regarded as the geeral etwork. The back roagatio (BP) learig algorithm is used for olie real-time traiig. The rule fuctio is: 1 J! E 1! 1 #r $ (k) " y (k) & 1! e (k) (9) : SN of the iut variat, which is same as s, 1, a. Weight value adjustmet algorithm from hidde layer to outut layer after k-ste traiig w shi (k +1) w shi (k)! " shi #J # w shi ' w shi + "! h (k)q si (k) (30) ' δ h 1 ( k) e y ( k) y ( k 1) ( k) u ( k 1) u ( k ) h h (31) b. Weight value adjustmet algorithm from iut layer to hidde layer after k-ste traiig 1 w sij (k +1) 1 w sij (k)! " sij #J # 1 w sij w sij + ""! shi (k)r sj (k) (3) 1 h1 '! shi (k)! h (k)( w shi )sg q (k) " q (k "1) si si x si (k) " x si (k "1) c. Secific imlemetatio stes of algorithm (33) a) Idetify the iut/outut variats ad iitialize the weights: b) Comute the outut u ( k ) ad u ( ) 1 k of NNC accordig to the iut give value r ( k) ad outut samled value y ( k ) c) Comute the chage Δ w of the weight value accordig to BP back roagatio algorithm ad chage the sij weight values of the eutros. d) Retur to the ste b) ad erform it. 6. EXPERIMENTAL RESULTS AND ANALYSIS [8] The comarative exerimet results of PID doublevariat cotrol ad eural etwork PID double-variat cotrol are show as follows: 6.1. Comarative Exerimet of Ste Iut Sigals Exerimetal Coditios The iut sigal is the ste sigal ad the give force F00N ad give seed V40mm/s Exerimetal Results 1) The ste resose curve of PID cotrol is show as the Fig. (8 ad 9): ) The ste resose curve of the eural etwork PID cotroller is show as the Fig. (10 ad 11) Exerimetal Aalysis 1) Comarative aalysis of the ste resose curve i the Fig. (8 ad 10) a. Aalysis o overshoot MP: The Fig. (8) shows the PID cotrol ste resose curve. From the coordiate e o the curve, the overshoot eak is 0.5N, the stable value is 00N, ad the overshoot value is 10. The Fig. (10) idicates the ste resose curve of eural etwork PID cotrol. Its overshot eak is 0 ad the resose has o overshoot. b. Aalysis o resose time t s. The Fig. (8) shows the PID cotrol ste resose curve. From the coordiate of oit e, the resose time is.4s. The Fig. (10) idicates the ste resose curve of eural etwork PID cotrol. From the coordiate of oit e, the resose time is 1.5s.

Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 519 Fig. (9). Seed PID Cotrol Ste Resose Curve. Fig. (10).

From the coordiate of the oit a ad c. the ascedig time is 0.4s. The Fig. (10) shows the ste resose curve of eural etwork PID cotrol. From the coordiate of the oit a ad c, the ascedig time is 0.7s. d.

From the coordiate of oit b, the delay is 0.6s. ) Comarative aalysis o seed ste resose curve i the Fig. (9 ad 11). a. Aalysis o resose time: The Fig. (9) shows the seed ste resose curve of PID cotrol.

8 Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume Fig. (9). Seed PID Cotrol Ste Resose Curve. Fig. (10). Force Neural Network PID Cotrol Ste Resose Curve. Fig. (11). Seed Neural Network PID Cotrol Ste Resose Curve. c. Aalysis o ascedig time t r. The Fig. (8) shows the ste resose curve of PID cotrol. From the coordiate of the oit a ad c. the ascedig time is 0.4s. The Fig. (10) shows the ste resose curve of eural etwork PID cotrol. From the coordiate of the oit a ad c, the ascedig time is 0.7s. d. Aalysis o delay t d. The Fig. (8) shows the ste resose curve of PID cotrol. From the coordiate of oit b, the delay is 0.6s. The Fig. (10) shows the ste resose curve of eural etwork PID cotrol. From the coordiate of oit b, the delay is 0.6s. ) Comarative aalysis o seed ste resose curve i the Fig. (9 ad 11). a. Aalysis o resose time: The Fig. (9) shows the seed ste resose curve of PID cotrol. From the coordiate of the oit f, the resose time is.5s. The Fig. (10) shows the seed ste resose curve of eural etwork PID cotrol. From the coordiate of the oit f, the resose time is 1.s. b. Aalysis o stable state error: the stable state error of the resose curve is 5mm/s i the Fig. (9). The stable error of the resose curve is oly 10mm/s i the Fig. (11). The above exerimetal data idicate that the ANN- PID cotrol system ad PID cotrol system have same resose seed o the iitial hase. Although the ascedig time of ANN-PID cotrol system is loger slightly tha it

50 The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 Xu Jiaqiag Fig. (1). Force PID Cotrol Sie Resose Curve. Fig. (13). Seed PID Cotrol Sie Resose Curve. Fig. (14).

The ste resose of the ANN-PID cotrol has o overshoot, so it shortes the ste resose time of the ANN- PID cotrol system, which is caused by itelliget adjustmet of the eural etwork cotrol algorithm.

cotrol system. 6.. Comarative Test of Sie Iut Sigals 6..1. Exerimetal Coditios The iut sigal is Sie sigal. The force F is the sie sigal with 100N amlitude ad 4s cycle.

9 50 The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 Xu Jiaqiag Fig. (1). Force PID Cotrol Sie Resose Curve. Fig. (13). Seed PID Cotrol Sie Resose Curve. Fig. (14). Force PID Cotrol Cie Resose Curve. of PID cotrol system, the resose time of ANN-PID cotrol system is 0.9s less tha it of the PID cotrol system. The ste resose of the ANN-PID cotrol has o overshoot, so it shortes the ste resose time of the ANN- PID cotrol system, which is caused by itelliget adjustmet of the eural etwork cotrol algorithm. The ste resose of the ANN-PID cotrol system features first quick ad later slow, o overshoot, short resose time, ad small stable state error, so it imroves robustess, accuracy ad quickess of the cotrol system. 6.. Comarative Test of Sie Iut Sigals Exerimetal Coditios The iut sigal is Sie sigal. The force F is the sie sigal with 100N amlitude ad 4s cycle. The seed V is the sie sigal with 00mm/s amlitude ad s cycle Exerimetal Results 1) The sie resose curve of PID cotrol is show as the Fig. (1 ad 13). ) The sie resose curve of eural etwork PID cotroller is show as the Fig. (1 ad 14) Exerimet Aalysis 1) Comarative aalysis o force sie resose curve i the Fig. (1 ad 14). To comare the curve i the Fig. (1 ad 14), the force outut resose of ANN-PID cotrol system ot oly tracks chage of the iut sigals, but also has smoother curve tha it of the PID cotrol system. ) Comarative aalysis o seed sie resose curve i the Fig. (13 ad 15).

Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 51 Fig. (15). Seed PID Cotrol Sie Resose Curve. To comare the curve i the Fig.

10 Research o Peumatic Servo Cotrol for Double-Cylider The Oe Electrical & Electroic Egieerig Joural, 014, Volume 8 51 Fig. (15). Seed PID Cotrol Sie Resose Curve. To comare the curve i the Fig. (13 ad 15), the seed outut resose of ANN-PID cotrol system ot oly tracks chage of the iut sigals, but also has the wave form which aroaches to the iut sigals better tha it of the PID cotrol system. 3) The curve obtaied i the exerimet has higher oises because the system uses the simulated shift sesor ad the motio seed sigals of the cylider are comuted via the differetial method. Whe the cylider oerates at a low seed, the frictio is bigger ad is ot uiform ad crawlig is easy to occur. The above aalysis results that the eural etwork cotrol algorithm ca effectively weake ifluece of the system s o-liear factors ad coulig amog variats. The resose of the ANN-PID cotrol system has o sigificat laggig comared to it of the PID cotrol system. CONCLUSION For the strog o-liearity ad low rigidity of the eumatic force servo system, this aer rooses the eural etwork cotrol strategy ad desigs the double closedloo PID hybrid cotrollers ad double-variat eural etwork PID cotroller for comarative exerimetal research. The exerimetal research idicates that the object model eed ot be kow ad the system eed ot be idetified i system cotrol based o a eural etwork, so this method is suitable for cotrol over the strog o-liear ad ucertai ukow system without accurate model. Better effect is achieved i the research exerimet for the eumatic force servo cotrol system. The eumatic force servo cotrol system ca imrove the system robustess ad cotrol recisio i case of double-cylider collaborative low-seed motio loadig. CONFLICT OF INTEREST The authors cofirm that this article cotet has o coflicts of iterest. ACKNOWLEDGEMENTS This work is suorted by Fudig Project of Cometece Develomet Program for Beijig VET Teachers i 1th Five-Year Pla ad Professioal Iovatio Team Costructio Pla of Beijig Vocatioal College. The roject ame is Car Detectio ad Maiteace Professioal Iovatio Team Costructio of Beijig Vocatioal College (No.JJH01357), which is led by Xu Jiaqiag. ABOUT AUTHOR Xu Jiaqiag (1965-), male, be bor i Beijig of Chia, master of egieerig, majorig i Mechaical maufacturig ad automatio secialty of Xi a Jiaotog Uiversity, a associate rofessor of Beijig Labor ad Social Security Vocatioal College ad Beijig Exerimetal Techical School, ad be iterested i research o fluid cotrol ad comuter cotrol. REFERENCES [1] Zhao Hog, Research o Cotrol Strategy of Peumatic Noliear System ad Its Alicatio, Ph. D. thesis, Xi a Jiaotog Uiversity, Xi a, Shaxi, Chia, 003. [] Zhog Hogsheg, Peumatic Trasmissio ad Cotrol, Beijig: Chia Machie Press, 199,. 1~116. [3] He Yue, Moder Cotrol Theory Foudatio, Beijig: Chia Machie Press, 1988,. 1~15. [4] Hachi G.D, Ozguer U ad Drakuov S, Noliear Cotrol of a Rodless Peumatic Servoactuator, Cotrol Alicatios, 199,. 516~51. [5] Li Shiyog, Fuzzy cotrol, Neural Network ad Itelliget Cotrol Theory, Chia: Harbi Istitute of Techology ress, 1998,. 8~19. [6] Yag Gogju, Research o Peumatic Itelliget Multi-oit Positio Cotrol System, Beijig Istitute of Techology, Beijig, Chia, [7] Ga Chegyu, Research o Cotrol Method ad Imlemetatio of Peumatic Pressure Loadig System, Postgraduate dissertatio, Xi a Jiaotog Uiversity. Xi a, Shaxi, Chia, [8] Xu Hog ad Tao Guoliag, Theoretical Aalysis o Damig Pressure-variable Proortioal Servo System ad Exerimetal Research o Its Lows-seed Features, Chia Mechaical Egieerig, vol. 13, , No.15, 00. Received: October 16, 014 Revised: December 3, 014 Acceted: December 31, 014 Xu Jiaqiag; Licesee Betham Oe. This is a oe access article licesed uder the terms of the Creative Commos Attributio No-Commercial Licese (htt://creativecommos.org/liceses/byc/3.0/) which ermits urestricted, o-commercial use, distributio ad reroductio i ay medium, rovided the work is roerly cited.

An Advanced GPS Carrier Tracking Loop Based on Neural Networks Algorithm

An Advanced GPS Carrier Tracking Loop Based on Neural Networks Algorithm Iteratioal Joural of Egieerig ad Alied Scieces (IJEAS ISSN: 394-366, Volume-3, Issue-9, Setember 06 A Advaced GPS Carrier Tracig Loo Based o Neural Networs Algorithm Jichu She, Shuai Che, Chaghui Jiag,