Int ern at i onal  Journ al of Ele ctrical  an d  Co mput er  En gin eeri ng   (IJ E C E)   Vo l.   9 , No .   5 Octo ber   201 9 , pp.  4099 ~ 41 13   IS S N: 20 88 - 8708 DOI: 10 .11 591/ ijece . v9 i 5 . pp4099 - 41 13     4099       Journ al h om e page http: // ia es core .c om/ journa ls /i ndex. ph p/IJECE   Dynami c anal ysis and  QF T - bas ed robust  contr ol d esig of   coaxial m icro - he licopt er       Aissa   Me ksi,  Ah med  H ami da Bo udin ar, B enatma n K oua dri   El e ct ri ca l   Eng in ee ring   Depa r tment,   Fa cul t y   of El ec tr ic a l Engi ne er ing,     Univer sit y   of   Sc ie nc es  and   Tech nolog y   of  Or an,  A lg eri a       Art ic le  In f o     ABSTR A CT    Art ic le  history:   Re cei ved Ma r   9 , 201 9   Re vised  A pr1 9 , 201 9   Accepte Apr 25 , 2 01 9       Thi pape pre s ent the   d y namic  beha v ior  of  coa xi al   m ic ro - hel i copt e r ,   under   Quanti t ati ve  Feedback  Th eor y   (QF T)  con t rol .   Th fli gh dy nami cs  of   aut onom ous  ai vehi cles  (AA Vs with  rota ting  rings  is  no n - li ne ar  and   complex.   The n ,   it   bec om es  nec e ss ar y   to  cha r ac t e riz th ese   non - linear i ti es  for   ea ch  f li ght   conf i gura ti on ,   in  ord er  to  provid th ese   aut onom ous  ai veh icle s   (AA Vs )   with  aut onom ous  fli ght  a nd  navi gati on  ca pabi l it i es.   The n,   the   nonli ne ar  m ode is  l ineari z ed   aro und  the   o per ating  poin using  som assum pti ons.  Fi nal l y ,   robust  QF cont rol  law   over   the   coaxia m ic ro - hel i copt er  is  app li ed  to  m eet  some  spec ifica ti ons.  QF (qua nti tativ fee db ack   th eor y is  a   control  l aw  design   m et hod  that  use fre quency   dom ai concept s   to  m ee p erf orm anc spe ci f ic a ti o ns  while   m ana gi ng  unce rt ai nt y .   Thi m et hod   is  base on  th fee db ac con trol   when  th pla nt  is  unce r tain  or  when   unce rt ai distur banc es  are   aff e ct ing  th pl ant.   The   QF design  appr oac invol ves  conv en ti onal  fre qu ency  response  loop  s hapi ng  b y   m anipulat ing   th e   gai v ari ab le   wi th  the   pole and  ze ros  of  the   no m ina tra nsfer  func ti on .   Th e   design  proc ess i s   accom pli shed  b y   using MATL A envi ronm ent software .   Ke yw or d s :   Coaxial  m ic ro - helic op te   MATLAB   t oolbo c k   Non - li near   Qu a ntit at ive  f e edb ac k t he ory   Copyright   ©   201 9 Instit ute of   Ad v ance Engi ne eri ng  and  Sc ie n ce   Al l   rights re serv ed .   Corres pond in Aut h or :   Aissa   Me ksi   Ele ct rical  En gi neer i ng D e par t m ent, F acul ty   of Elec tric al  E ng i neer i ng,    Un i ver sit y o f S ci ences a nd Te chnolo gy  of Or an ,   B.P 15 05, El  Mnao uer,  Or a n, A l ger ia .   Em a il ai ssa. m eksi@ un i v - us t o.dz       1.   INTROD U CTION   Un m ann e aer ia veh ic le ( U AV s a re  ai rc r aft  capa ble  of  fly ing   a nd  pe rfor m ing   m issio with o ut   hu m an  prese nc on  bo a rd.  I ni ti al l dev el ope as  pa rt  of  m i li ta ry  act ivit ie s ther is  now  gr e at   pote nti al   for  ci vil  act ivit ie (survei ll ance,  carto gr a ph y. ..).   Ma ny  unres ol ved   co ns trai nts  rem ai fo t he   us of   ci vil  dron e s   in  the  pu blic  sp ace.  Am on t h ha r points  to  res olv a re  e m bed de deci sion - m aking   a utono m y,  al l - weathe r   per ce ptio cap abili ty safety   and   dep e ndabil it y.  These  featur es  will in  the  fu tu re,  be  s ubj ect   to  certi f ic at ion   process  b ei ng  dev el op e d, but  current  dro nes suffer  fro m  a lack  of   r obus t ne ss and a uton om y.   In   the  li te ratu r e,  m od el ing   an de velo pm ent  of   new   form ulati on of  coa xi al   m ic ro - helic op te rs  ha ve   been   t he  sub j e ct   of   seve ral  r esearch  w orks   su ch  as:   Mo ham m ad  Har un   work [1 ] a nd  Che [ 2]  on  cl assic  helic op te rs,   Alvaro  [ 3]  for  th m inidr one  w it keeled  pro pelle r,   as  w el as  the  wo r ks   of  Sc hafro t a nd   Christi an [4]  on c oax ia l m ic ro - helic opte r (w it fixe a nd  va riable ste ps ).   Am on these  m ilit ary  dr on e s     m ic ro   coaxi al   helic op te presente in  this  arti cl was  dev el oped  by  par t ner I SL,  CR AN   an H EUDI ASYC  Syst e m s.Th m or detai le m od el of  so m con fi gurati on of   the   coax ia m ic ro - helic op te   ca be  fou nd  in  [5 , 6].T he  c oax ia l   m ic ro - helic opte with  swas hpla te   co ns ist of   t w counter - r otati ng  r ot or e nsuri ng  the  li ft  a nd  pro pu lsi on  o t he  ai rc raf a nd   swa shpla te   m ou nted  at   the   lowe ro t or   t co ntr ol   the  cra ft.  T he   two  c ounte r - ro ta ti ng  r otors   sh a re  the  s am axis  of   ro ta t ion ,   w hich  m akes  it   po s sible  to  c om pen sat m utu al ly   fo the  rea ct ion   to rque  produce by  eac of  the  r oto r s,   and   co n se qu e ntly   the   Evaluation Warning : The document was created with Spire.PDF for Python.
                IS S N :   2088 - 8708   In t J  Elec  &  C om En g,   V ol.  9 , N o.   5 Oct ober  20 19  :   4099   -   4 1 1 3   4100   ste ering   of   the   la ce.  This  syst e m   is  al so   com pact  anti - t orq ue  so luti on,   wh ic sa ves  t he  au xili ary  syst e m   pr ese nt  on  c onve ntio nal  helic op te a nd  the r efore  reduce  t he   le ng t of  the   fu sel a ge.   The   m ass  of  the  dr on e   is   li fted  by  t he  t ot al   thru st  pro duced   by   the  t w ro t or s ro t ary  wi ng  dr one  can   be   co ns i der e as   rigi body  with si x degree s of free do m  f or whic the  aer od y nam ic  f or c es g e ner at e d b y i ts act uator s   are a dd e d   [7] .   The  pro blem   of   this  coa xial  m ic ro - helic op t er  is  non - li nea r,   com plex,   qu asi   sta ti on naire   and   s ubj ect   to  wi nd  disturbance a ff ect in it sta ble  fli gh t   an ea sy  la nd i ng  a nd  the   dyna m ic   equ at ion of  m otion   for   a   coax ia m ic ro - helic op te incl ud i ng  str uctured  a nd   unstr uc ture unce rtai nties.  The   ap plica ti on   of   t he   QFT   te chn iq ue  f or   t raj ect or trac ki ng   i the   pres ence  of  uncert ai nties  of  m as va riat ion   pa r a m et ers  is  stu died.I this  pa per,  t he   QFT  Horowit I. m et hod  as  an   ap pro pr i at rob us c on t ro m et ho is  app li ed   to   c oax ia l   m ic ro - helic op t er, w hich has  m any adv a nta ge s over ot he r  re le van t t ec hn i ques  [8 ] .   The  Q FT  H oro witz   I. desi gn  proce dure  is  no on ly   base on   the  am plit ud of  the  tra ns f er  f un ct io in  the  fr e quen cy   do m ai n,   but  al so   ta kes  into  acco un the   ph ase  in f or m at ion   [ 8].Th qu a ntit at ive  fe edb a c theo ry  QFT,  c al le (La  t or ie   qu a ntit at ive  de  la   ret ro act i on  TQR in   Fre nch  H oro witz   I. M,   was   de ve lop e by  Isaac  Horowit z.  It  is  frequ e ncy  te ch ni qu us in the   Nich ols  diag r a m   to  reali ze  robust  co ntr ol  [9 ] .   The  desi red   te m po ral  responses  are  translat ed  into  tolera nc es  in  the  fr e quency  dom ai n.   The  desig pr ocess  is  highly   trans pa ren t;   al lowi ng  desig ne to  see  wh at   tr adeoff  is  nee ded   t ac hiev desire le vel  of   perform ance.   In   this  m anu s cript,  we  a ddr ess  the  three  m ai pr oble m of   the  a utom at ic   app li ed  to  the  coa xia l   helic op te m icr o - ty pe  dro ne,  since  we  a re  interest ed   in  de te rm ining   a   dy nam ic   m od el   of  the   syst em a nd   aerona utica m od el , as wel l as cal culat ing  a s ta bili zi ng  input . W e are m or sp eci fical ly  intereste in ac hi evin the ob j ect ives t hu s  define d:   -   Mod el in the  transla ti on   an ro ta ti on  dy na m ic of   the  c oa xial  m ic ro - he li cop te dur in it autonom ou fligh phase [10] .   -   Esta blish  the  li near   m od el   of  the  nonli near   m od el   by  app l yi ng   the  Tay l or  ap pro xim ation   m et ho ar ou nd  the ope rati ng point.   -   Determ ine  st abili zi ng   co rr e ct or   f or   c oa xial  m ic ro - helic op te in  the  case of  q ua si - sta ti onary  flig ht,  f r om   su f f ic ie ntly  sim ple synthesis m od el [11] .      T he  rem ai nd er  of   t his  pa pe is  organ iz e as  fo ll ow.  I Sect ion2,  the  dynam ic   m od el   of   coa xial  m ic ro - helic op t er  is  f or m ulated   in  the  Ca rtes ia sp ace I S ect ion 3,  Q FT  con t ro ll er  is  de velo ped   a nd  a pp li ed   to  the  direct  dy nam ic  m od el   of   dro ne  in  Ca rtesi an  sp ace.  Se ct ion pr ese nts   si m ulati on   resu lt of   the  pr opose con t ro ll er  an d discussi on  of   r esults. Fi nally , s om e con cl us i on s  are  prese nted  in  t he  cl os in sect io n.       2.   MA T HEM AT ICA L  MODE L OF  COA X I AL  MICRO - HEL ICO PTE R   Kno wing  that   coa xial  m ic ro - helic opte is   co ns ide red  a ri gid   body   with  fi xed  m ass  " m ",     the  ge ner ic   m od el   with  six  degrees  of  fr ee do m   (6 - DOF)  ref e rs  to  it three  translat ion s   and   it th ree  sp at ia ro ta ti ons.  It  th eref or descr i be the  dynam i cs  and   t he  ki nem at ic of   ro ta ti on   an tra ns l at ion   in  the  refe ren ce   syst e m   li nk ed  to  the   "G"   bo dy The re  a re  t wo  ways   of  e xpressi ng  the   e qu at io ns  of  m otion :   ei ther   f r om   the   la ws  of  Ne wton  an E uler  th a we  will   us or   by  refo rm ul at ing   the  New t on ia m echan ic with  the  La gr a nge   and  Ham il ton   equ at io ns.  I t his  sect io a re  def i ned  the  refe ren ce   points axes  syst em us ed  a nd  the   ki nem at ic  and  ki netic   pa ram et ers  app e arin in   Fig ur 1 ,   as   well   a al the  relat ion s   c om po sin th 6 - DOF  m od el .   It is ass um ed  that t he  i ner ti al   eff ect of the  r otors a re  neg li gib le  c om par ed  to  th os of th e m a in bod y.           Figure  1. La nd m ark us e to   descr i be  the  m ov em ent of the   coa xial m ic ro - helic op te r   Evaluation Warning : The document was created with Spire.PDF for Python.
In t J  Elec  &  C om En g   IS S N: 20 88 - 8708       Dyna mic an alysis a nd QFT - base r obus t c ontrol  desig n of    ( AissaMeksi )   4101   2.1.   Ref ere nce  an d  axes s ystems     The  m ov em ent of a  rigid b od y i s m a inly  d escribe d by tw r efere nces:   -   The  ine rtia ref ere nce  {J}:   boun to  the  E arth,   hav i ng   a   ref ere nce  po i nt  "O"  an de fine by  the  ba s e     {x e ; y e ; z e },  where { x e } poi nts nor t h,  {y e po i nts east, a nd {z h } points  up.    -   The  na vig at io m ark   (g ea or  m ob il e)  { B }:   relat ed  t t he  ce nter  of  gravit "G"   of  t he  str uctu re  of   the  dro ne  an de fi ned   by  the  ba se  {x C y G z G wh e re  {x G is  the  longit ud inal  axis  point ing   to wa rd t he  veh ic le  f r ont,  { y G def in es  t he   la te ral  axis   an po i nts  t the  v eh ic le  r ig ht  a nd  {z G } d efi ne the   ve rtic al   axis   of the  ve hicle  an d p oin ts  to  t he  top.   r otati on  m a trix  ide ntifie s   the  ori entat io of   a ny  c oor din at syst em   in  the  t hr ee  dim ension s     of   s pace.  It  is  al so   cal le transiti on   m at rix  or   at ti tud m atr ix,  no te "R η".  Con se quentl y,  the  passa ge  m at rix     of  the  ref e r ence  a pparatu s   " B"   to   the   re fe ren ce   ine rtia " J"   is   pa ram et erized  by   the  Eul er  a ng le s   desc r ibing   the  t hr ee   r otati on  a ng le s     ( r ol l),   θ   ( pitch)   a nd    (yaw ). T he   product  of  t hese   three   m at rices  m akes  it   possible   to exp ress  t he passa ge fr om  t he refe ren ce  " B"   to t he  r efe re nce " J" ,  that is  to say [ 12] :     = [       +    +          ]   (1)     w he re    is  the  ro ll   an gle,  θ  t he  pitch  a ngle   and     the  ya w   ang le res pect ive ly .Th kin e m at ic   equ at ion  of   or ie ntati on  c onnects  t he   tim der i vative   of   the   an gles   of   r ol ,   pitc θ  a nd  ya   to  the  instanta neous  s pee of rota ti on   by:     = [ 1    0   0     ]   (2)     2.2.   Balance  of  fo r ces a nd  m ome nt appl ie d   Eq uations   of   f or ces  a nd   m ome nts  are  req ui red   to  com plete   the  dynam ic  m od el ing   of   ae ro m echan ic al   syst e m s.  More ov e r,   t he  m od e li ng   of  the se  f or ces  a nd  m ome nts  (that  is  to  say F res  a nd  M rem ai ns   m ajor   pro blem du to  the  com plexity   of   fl uid   dynam ic and   inte racti on s   bet we en  the   stu died  veh ic le   (r igi body   equ i pp e with   m ov ing  act ua tors)  a nd  the   su r rou nd i ng  fl uid .   T he  m ai f or ces   act in on  a   r otorcra ft  are   pr ese nted   [ 12] .     2.2.1.   Eff ort   set ting s   The  e ffor t   pa r a m et ers  are  si m pl the  thre f or ces   an t he  th ree  m ome nts  that  i nter ven e   in   the   dynam ic of   translat ion   a nd   ro ta ti on  of   the   veh ic le The  dev el op m ent  of   these  effo rts  will   con sti tute  the   aerodynam ic  m od el  p rese nte in  the  f ollow i ng secti ons  [7 ,   12 ] . We  note  th en:      = [ ] ,   the t hr ee  co m pone nts  of  t he fo rce  vecto e xpress ed  in  " B ",   M = [ ] ,   the t hr ee  co m pone nts  of  t he m o m ent v ect or  expres sed  in  " B ".   a nd     = [ ]   (3)           Figure  2. T he f or ces  and m ome nts  on the syst e m      Evaluation Warning : The document was created with Spire.PDF for Python.
                IS S N :   2088 - 8708   In t J  Elec  &  C om En g,   V ol.  9 , N o.   5 Oct ober  20 19  :   4099   -   4 1 1 3   4102   2.2.1. 1.   The f orces   The fo rces  w hich  a re acti ng  on the  syst em  are:   -   T he wei ght o the  co a xial m ic ro - helic opte r   The  ve hicle   is  su bject e to  the  gr avita ti onal   fiel as  al l   bodies  in  the  vicinit of   the  Ear th,  an the   acce le rati on   of   gravit is  de note " g".  Th weig ht  com ponen " P"  is  the   force  to   w hich   the  cra ft  is  s ubj ect ed,  and   whose  di r ect ion   is  norm al   to  the  su rf a ce  of   the  Eart h.  The  val ue  of  "g"  is  exp ress ed  an know in  the  iner ti al  r efe re nc e fr am e " B" , t hat is to  say [ 12] :     = g z e   (4)     -   T he pushi ng for ce     The  m od el   of  thr us of  ro ta r wing   can  be  expresse as  a   functi on  of  th sp ee of  r ota ti on   of  t he   ro t or  squa red  Ω p 2   at   ce rtai ae r od y nam ic   coeffic ie nt  " k" .   T he   li ft  gen e rated   by   ro ta ti ng  r otor  t her e fore  has  t he  f ollo wing e xpressi on :     = Ω 2   (5)     -   T he dra g forc e   The  aer od y na m ic   dr ag  force   of   the  ve hicle   F a   caused   by  the  fr ic ti on  of   t he  ai on   the  f us el age  ca be  expresse as:   [ 13 ]     F a = 1 Ω 2   (6)     Ω   : M oto r  rotat io s pee d.     2.2.2.   Forces  gener ate d by  th e  coa xia l mi cr o - h el ic op ter   Thrust  is  the  m ai fo rce  pro du ce by  the  coax ia m ic ro - he li cop te r,   al lo wing  the  dr one  to  m ov i three - dim ension al   sp ace.  T he  i m m ob il up pe ro t or   pro duce on ly   ver ti cal   thru st  force  F 1 wh ic is  directl pro portion al   to   the  s quare   of  the  s peed  of  r ot at ion   of  it bl ades  Ω 1   an the  a erodynam ic   coef fici ent  of  th r us t    suc h t hat:     1 = [ 0 0 Ω 1 2 ]   (7)     The  ex pressi on  of  the  f orce 2   gen e rated  by  the  lowe r otor  de pends  on   t he  tw an gles   δ cx and   δ cy of   t he   swas hp la te   wh i ch fo rm  the inputs  of   t he  syst e m . Th is force i s g i ven b y:     2 = Ω 2 2 [           ]   (8)           Figure  3.Sum m ary of  t he  el e m ents r el at ed  t the  m od el ing o the  co a xial  m ic ro - helic op t er   Evaluation Warning : The document was created with Spire.PDF for Python.
In t J  Elec  &  C om En g   IS S N: 20 88 - 8708       Dyna mic an alysis a nd QFT - base r obus t c ontrol  desig n of    ( AissaMeksi )   4103   The  t otal  th rus t,  = [ 0 0 ] i def i ned  by  the   s um   of  the   in div id ual   thr us ts   of  eac of  the  ro t or s   [ 1 ]    and   [ 2 ]    de scribe d by  ( 5)  and ( 6). T he  e xpressi on  of  t he t otal t hrust ca n be  w ritt en:       = [ Ω 1 2 +     Ω 2 2 ]   (9)     Finall y, the t otal force F  g e ne rated  by the  tw o ro t or s  can  b e  w ritt en:      = [ ] = [     Ω 2 2   Ω 2 2 Ω 1 2 +     Ω 2 2 ]   (10)     2.2.3.   Moment act i ng   on the  c oaxi al micr o - heli copter   The  m o m ent  M p   gen erate by   the  total   fo rc app li ed  to  t he  coa xial  m icr o - helic opte a it center  of   gr a vity   is  th eref or t he  sum   of   the  m o m e nts  res ulti ng   from   the  fo rce   of  the  upper   r oto F 1 that  of  th lowe r   ro t or   F 2   a nd th e w ei ght P .     M =    1 1 +    2 + 2 +      (11)     As  co nclusi on, t he  s um  o t he  m o m ents o f  the  forces  F is  giv e n by:     M = [   Ω 2 2     0 ]   (12)     Howe ver,  m om ent  aro un t he  a xis  due  t eac of  the  r otors  is  di rectl propor ti onal   to  the  s qu a re  of  their  velocit y wit h t he  ae rodynam i c coe ff ic ie nts  1 > 0 , 2 > 0 ,   su c as:     = [ 0 0 1 Ω 1 2 2 Ω 2 2 ]   (13)     Finall y, the t otal  m o m ent  M p   is  giv e n by:     M = [ ] = [   Ω 2 2     Ω 2 2 1 Ω 1 2 2 Ω 2 2 ]   (14)     Accor ding to  (2. 10)  a nd (2. 14 ),  t he  e xpr essio n of t he  t otal f or ce  F  ca n be  wr it te as:     = ( 0 , 0 , ) + wh e re = 1  ( )   (15)     2.2.4.   Dev el op me nt   of t he  m at he mat ic al m od el  a cc ordin g to  Newton - Eul er   Af te prese nting   t he  diff e re nt  eq uatio ns   we  ca no de velo the  m od el   m at he m a tical   us in th e   New t on - E uler fo rm ulati on . T he  c orres ponding eq uatio ns  a r e wri tt en  in t h e  foll ow i ng form  [ 7 , 12] :     ̇ = ̇ = + +  ̇ = Ω Ω ̇ = Ω × Ω + +    (16)     w he re:     :   Is  the  posit io n vecto r of  t he c oax i al  m ic ro - helic op te r = [ , , ] ,   m:   the total  m a ss of the  c oax i al   m ic ro - he li co pter,   J:   sym m et ric m at rix  inerti of d im ension   ( 3x3),  giv e n by:     = [ 0 0 0 0 0 0 ] ,   (17)     Ω :   The  ang ular v el ocity  ex press ed  in  the  fixe d i ner ti al  r efe re nc e .     Evaluation Warning : The document was created with Spire.PDF for Python.
                IS S N :   2088 - 8708   In t J  Elec  &  C om En g,   V ol.  9 , N o.   5 Oct ober  20 19  :   4099   -   4 1 1 3   4104   2.2.5.   Dyna mi c  simu lation  of   th e  s ys te m in  M A TL AB - Simul i nk  en viro nme nt   MATLAB   Si m ul ink   s of t wa re  is  us ed   to   si m ulate   the  pro pu lsi on  s yst e m In   Fi gure  4,  asi is  m entioned,  s ect ion   of  bo dy  m od el ing cal l ed  rigid  bo dy,  is  show nin  Si m ul ink   e nv i ronm ent.  In   this  syst e m ,   inputs,  f or ces  and   th total   dr ag  f orce  are  e xer te on  the  center  of  m ass   of   the  ve hicle Total   dr a f orces,   befor e   enteri ng  the  body  dy nam ic   cal culation syst em wh ic are   w ritt en  base on   New t on  Eule r’s  la w,    are  tra ns fe rred   from   body  c oor d inate   syst e m   to  inerti c oor din at syst e m In   this  syst e m pro pu lsi on   syst e m   and   t he  tra ns it ion   of   f or ces  a nd  m o m ents  fr om   bo dy  c oord i nate  syst e m   to  inerti co ordin at syst e m   are  add e to  the  rigi body  sub - syst e m . The  physi cal   pr ope rtie   of   the  c oa xial  m ic r o - hélic o pter  are  c ons idere   as in Ta ble 1 .           Figure  4 . N on li near sy ste m  o the co a xial m icr o - helic opte r       Table 1 .   Pa ram et ers  m od el  of  coax ia l m ic ro - helic op te   Para m eter   Descripti o n   Valu e   Un it   m   I x   I y   I z   d           m a ss (kg )   Mo m en t of  inertia  ab o u t X  ax is   Mo m en t of  inertia  ab o u t Y  ax is   Mo m en t of  inertia  ab o u t Z  axis   Distan ce c en ter  of  grav ity   aerod y n a m i c coef f icien t of  thru st   aerod y n a m i c coef f icien t of  thru st   Aerod y n a m i c   co ef f icien t y aw   o f    u p p er  roto r   Aerod y n a m i c   co ef f icien t y aw of    lo w er  roto r   0 .29 0   1 .38 3  ×  1 0  −3   1 .38 3  ×  1 0  −3   2 .72  ×  1 0  −4   0 .06 7 6   3 .68  ×  1 0  −5   3 .77  ×  1 0  −5   1 .47  ×  1 0  −6   1 .32  ×  1 0  −6   Kg   k g m 2   k g m 2   k g m 2   m   N/rad 2 s 2   N/rad 2 s 2   N. m . rad - 2 .s - 2   N. m . rad - 2 .s - 2       2.3.   Expressi on   of  li nea ri z ed t angen m od el     The  nonlinea syst e m   descr ib ed  by  eq uatio (2.16 will   ser ve  as  a   ba sis  f or   t he  desig of  the  c ontr ol  al gorithm by  nonlinea a ppr oach e s,  but  it   is  abs olu te ly   not  s uitable   f or   the  sy nth esi s   of  li near  al gor it h m s.  Howe ver,  thes appro ac h es  are  base on  dynam ic  m od el   li near iz ed  arou nd   certai point  of  ope rati on,    in  this  case   hove rin g.  The   li near iz at ion  as su m ption of   t he  nonlinea m od el   relat ing  to  the   ho ver   are  a s   fo ll ows:   -   The  tra ns la ti on  sp ee v   of the   veh ic le  is sm al l:   = ( ) ,   -   The p os it ion   ξ   of the  ve hicle  is any,   -   The rotat io s pe ed  Ω   of the  v e hicle  is lo Ω   o (e),   -   Roll  angle  φ   a nd p it ch  angle  θ   are l ow  = o (e) ,   -   The  ya a ng le   ψ   is  regulat ed  to  ze ro.   The  li near iz e m od el   ta ng e nt   to  the  vicinit y   of   th hove i com po sed  of   the  f ollow i ng  four   SI S (S in gle In pu -   Sing le   O utput)  ch a nn el s:   -   A n al ti tud e c hain ha ving it sta te   [ ]   :     [ ̇ ̇ ] = [ 0 1 0 0 ] [ ] + [ 0 1 ]     Evaluation Warning : The document was created with Spire.PDF for Python.
In t J  Elec  &  C om En g   IS S N: 20 88 - 8708       Dyna mic an alysis a nd QFT - base r obus t c ontrol  desig n of    ( AissaMeksi )   4105   -   A  roll ing chai n hav i ng it s stat [ ]     [ ̇ ̇ ] = [ 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 ] [ ] + [       0 0 0 1  ]             -   A  p it c c hain  h a ving it s stat [ ]     [ ̇ ̇ ̇ ̇ ] = [ 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 ] [ ] [       0 0 0 1  ]             -   A  stri ng of la ce ha ving as  stat [ ]   :     [ ̇ ̇ ] = [ 0 1 0 0 ] [ ] + [ 0 1  ]         3.   QFT C ONT R OLL ER OF  COA X IAL  M I CRO - HEL IC OPTE R   Qu a ntit at ive  F eedb ac The ory   is  rob us c on t ro desig t echn i qu e   that  us es  feedbac to  achie ve   respo ns es  t ha m eet   sp eci fied  s pecifica ti on des pite  struct ur e plant  unc ertai nty  an plant     disturba nces  [ 14 15] This  te chn i qu has  been   a pp li ed  t m any  c la s ses  of   pro blem su c as  Sing le   In put  Sing le  Outp ut ( SISO ),   M ulti pl Inp ut  Sin gle   O utput  (M ISO a nd  Mult iple Inp ut  Mult iple Ou t pu t (MIM O)  f or   bo t co ntinuo us  and   disc rete  cases.  For  this   researc h,   SISO  syst em   is  a ssu m ed  fo c ontr ol  la desig an d   seq uen ti al   loop  cl os ures  are  util iz ed  as  wit the  po le   plac e m ent  con tr oller.  The  Q FT  de sign   m et ho do log is  qu it tra nspare nt,  al lo wing  t he   desi gn e t see  the  necess ary  tra de - offs  to  ac hieve  t he   cl os e d - l oop  s yst e m   sp eci ficat io ns   [16 17] Th basic  ste ps   of  the  pr oce dure  a sh ow in Fig ur 5   are  prese nted  in  the  f ollow i ng   su b - sect io ns .  T hey are:  [18 - 20]   -   Plant m od el  (w it uncertai nty) , Tem plate s g e ner at io a nd  nom inal plant se le ct ion   P o (j ω ).   -   Sp eci fy  acce pt able  trackin m od el s,  wh i ch  the  cl os ed - lo op   res pons sat isfie s,  T L   ≤  T R   ≤  T RU   and d et e rm ine trackin g b ound s   -   Determ ine  disturba nce  re j ect i on  m od el s,  T D   base on  disturba nce  re j ect i on  sp eci ficat ion s,  a nd  determ ine   disturba nce  bounds   -   L oop - s hap i ng the c on t ro ll er   G (j ω ).   -   Synthesize   no m inal l oo p t ra nsfer  fun ct io n,  L0 (j ω )   -   Pr e - filt er s ynt he sis  F (j ω ).   -   Si m ulati on  a nd D esi gn V al ida ti on .           Figure  5.QFT   con t ro l l oop        3.1.   Genera tin pl an t  t e mpla tes   Am on t he  fir s ste in  Qu a nt it at ive  Feedb ac T h eo ry  ( QFT)  desig proc ess  are  ge ner at ed  the  plant   tem plate s.  The   pla nt  te m plate ca pture  t he   uncertai nties   in   the  plant   tra ns f er  functi on,   (  )   a nd  a re  pl otted   as  boun dar ie s   of  m agn it ud e   an ph ase   va riat ion on  t he  Nich ols  ch art  at   sp eci fic  desig fr e q ue ncies .     These tem plate s ar e t hen use d t c reate b ound on the  N ic hols c har t as  sho wn in Fi gure  5 [19 21] .   Evaluation Warning : The document was created with Spire.PDF for Python.
                IS S N :   2088 - 8708   In t J  Elec  &  C om En g,   V ol.  9 , N o.   5 Oct ober  20 19  :   4099   -   4 1 1 3   4106   3.2.   Genera tin pe rfo rm an ce  b ounds   Give the   te m plate s,  QF c onve rts  the  cl os e lo op  m a gn it ude  sp eci fi cat ion i nto   m agn it ude  a nd   ph a se  co ns trai nt  on  the  open   loop  tra ns fe f un ct io for  the   no m inal  plant ( 0 = ( ( ) 0 ( ) ) . Us ually   syst e m   perform ance is d esc ribe as  re fer e nce trac ki ng, ro bust st abi li ty  an d   re j ect ion t i nput  disturba nce[20 21] .   -   Re fer e nce T rac king   To  m eet  the tra ckin g per form a nce th e  contr ol le sho uld   sat isfy t t he  f ollo wing:      | | |  1 +  | | |   (18)     w he re  t he  uppe an lo we boun ds   a re   de fined  us in ti m e   dom ai fig ur e   of  m erit   su c as  pe ak   ove rsho ot   and sett li ng  ti m e.     = ( ω n 2 a ) ( s + a ) s 2 + 2 Ɛ ω n s + ω n 2 /   ω n = 1    / ,   a=1 Ɛ   = 0 . 6   (19)     = s + 1 s 2 + 1 . 2s + 1   (20)     = K ( + 1 ) ( + 2 ) ( + 3 ) /   1 = 0 . 5   2 = 1   , 3 = 2 , K = 1   (21)     = 1 ( s + 0 . 5 ) ( s + 1 ) ( s + 2 )   (22)     Nu m erical  v al ues ( ω n , a , Ɛ , 1 , 2 , 3 , K ) are take f ro m  the sp eci ficat ion s .   -   Ro bust Sta bili ty   Robust  sta bili ty   in  QF am ou nts  to  c hec kin sta bili ty   us ing   no m inal  plant    ( 0 ( ) = ( ) ( ) )   and the n dem on strat in sta bili ty  o the  wh ole set pla nts b assigni ng a se nsi ti vity  r at ing 1 giv en  b [ 19 ]     |  1 +  | 1 =   (23)     wh e re:   μ is t he  ci rcle M   sp eci ficat io in  m agn it ud e:     M m  = 2 0 l og10( μ )   Φ   =2  c os - 1( 0 . 5 μ ) ϵ [ 0,   180º ]   In   [ 20] O dedYan i translat ed  this  co nd it i on   to  desire phase  m arg in  ( s how in   ( 24 a nd   gain   m arg in  ( )  s ho w in   ( 25 ).      = 2  1 ( 1 2 ( 1 ) ) ,     PM 180º  -   Φ   ( deg)   (24)      = 20 log ( 1 + 1 1 ) ,        GM    1 μ    ( m agn it ud e )   (25)     -   Dist urba nce rejecti on at  plan t ou t pu t   An   uppe lim it is set to the s ensiti vity  f un ct i on  t lim it   the p eak  value o disturba n ce am plific at ion  t the p la nt out put as f ollow [ 19 - 21]     | 1 1 +  |    (26)      = = 1 1 +  2 + ω n 2 + ω n + ω 2     ω n = 1    /      a=1 Ɛ = 0 . 6   (27)     = 1 1 +  2 + 1 . 2 2 + 1 . 2 + 1   (28)     These  boun ds   are the n use d f or lo op sh a ping to  d e sig a c on t ro ll er.       Evaluation Warning : The document was created with Spire.PDF for Python.
In t J  Elec  &  C om En g   IS S N: 20 88 - 8708       Dyna mic an alysis a nd QFT - base r obus t c ontrol  desig n of    ( AissaMeksi )   4107   3.3.   Contr oller   d es ign   The  co ntr oller  is  design e by  add i ng   po le a nd   ze ro es  to  th no m inal  transf er  f unct ion   by   sat isfyi ng  al bo un ds   at   each  f reque ncy.   Durin this  sta ge,   the  desig ne co n side rs  trade - off  betw een  the  s pecifica ti on ,   con t ro ll er  com plexity  and t he  co st  of   fee dba ck  in  the  ba nd width [ 19 - 21] .     ( ) = 3 . 072 ( 12 ) ( 350 ) ( 350 )   (29)     3.4.   Pre - filter   d esi gn   pre - filt er  is  need e to  br in the  res pons e   within  the  re f eren ce  trac king  tolera nces,  and   an is   done by a dd i ng  po le s  and ze r oes   [19 - 21] .     ( ) = 1 ( 0 . 30 0 66 )   (30)     Si m ulati on  a nd  validat io n of  the  desig is  show i Fi gure  6 [22] .           Figure  6.QFT   desig n proce dure       4.   RESU LT S  AND A N ALYSIS   The  f re qu e ncy  range  f r om   0. 01  to  1000  r ad/s   was  sel ect ed  f or   t he  Q FT  c ontr oller  desi gn.   This  ra nge  cov e rs  t he  fu ll   range  of  f re quencies  f or  ty pi cal   coa xial  m ic ro - helic op t er.  T he   first  ste of  the  QFT  de sign  is  to  ge ner at the   plant  te m plates.   These  plant  t e m plate wer gen e rated  us in the  Ma tl ab  s of t war [ 23]   and   a re  sh ow in  Fi gur e 7 .     Evaluation Warning : The document was created with Spire.PDF for Python.
                IS S N :   2088 - 8708   In t J  Elec  &  C om En g,   V ol.  9 , N o.   5 Oct ober  20 19  :   4099   -   4 1 1 3   4108       Figure  7.Pl ant  tem plate       Crit erion s  n e ed ed fo the  d e si gn i ng of t he  c ontr oller are:    -   T rack i ng crit erio n   The  up per   bo und  f or  the  tr ackin crit eri on  was  ge ner at ed  by  c on si de rin 1.1 9 oversho ot  an set tl ing   tim o 4.4 s.  T he  trans fer   functi on  us e f or  th uppe boun s how in  ( 19).   Si m il arly the  lowe r   bound   was   ge ne rated  by  c on s iderin no  ove r sho ot  an set t li ng   ti m of   9. 77 s T he  tra nsf er  f unct ion  use f or  the  l ow e bound  s how i ( 22) .T hese  bounds   i ti m do m ai and   fr e quency  do m ai are  sho wn   i Figure   8. T he  trac ki ng  bounds are  s hown in t he Nic hols  plo t i Fig ure  9.             Figure  8. T rac ki ng   boun ds ; (a )  tim e d om ai n; ( b) freq ue ncy dom ai n           Figure  9 I nters ect ion   bounds  (stabili ty , trac ki ng  a nd  disturb ance  rej ect io n)   - 3 5 0 - 3 0 0 - 2 5 0 - 2 0 0 - 1 5 0 - 1 0 0 - 5 0 0 - 1 5 0 - 1 0 0 - 5 0 0 50 100 N i ch o l p l o t O p e n - l o o p   p h a se   ( d e g ) O p e n - l o o p   m a g n i t u d e   ( d B ) 0 . 0 1 0 . 0 5 0 . 1 0 . 5 1 5 10 50 100 500 1000 - 1 0 0 - 8 0 - 6 0 - 4 0 - 2 0 0 20 M a g n i t u d e   ( d B )     10 -2 10 -1 10 0 10 1 10 2 - 1 8 0 - 1 3 5 - 9 0 - 4 5 0 P h a s e   ( d e g ) B o d e   D i a g r a m F r e q u e n c y     ( r a d / s ) Tu Tl 0 2 4 6 8 10 12 14 0 0 . 2 0 . 4 0 . 6 0 . 8 1 1 . 2 1 . 4     S t e p   R e s p o n s e T i m e   ( s e c o n d s ) A m p l i t u d e Tu Tl - 3 5 0 - 3 0 0 - 2 5 0 - 2 0 0 - 1 5 0 - 1 0 0 - 5 0 0 - 2 0 - 1 0 0 10 20 30 40 50 60 70 O p e n - l o o p   m a g n i t u d e   ( d B ) O p e n - l o o p   p h a se   ( d e g ) N i ch o l p l o t Evaluation Warning : The document was created with Spire.PDF for Python.