Page 9 - Read Online
P. 9

Ma et al. Complex Eng Syst 2023;3:10  I http://dx.doi.org/10.20517/ces.2023.14    Page 5 of 14


               and a new fixed-time sliding mode surface are proposed for the second-order subsystem (7). On this basis, a
               fixed-time controller is constructed to make the error variables,       and      , converge into a small region around
                                                                          ¤
               the origin. Then, a fixed-time controller is developed for the third-order subsystem (8), which guarantees that
               the system state variables,       ,       , and   , are all uniformly ultimately bounded, and the tracking errors,       and
                    , can converge into a small region around the origin in a fixed time.

               3.1. Tracking control laws design for the second-order subsystem
               3.1.1. Fixed-time disturbance observer
               Firstly, for the attitude error subsystem (7), define an auxiliary variable as

                                                                                                        (9)
                                                          1 =    −    1
               where    1 satisfies
                                                 1
                                                                         1
                                            ¤    1 =     1 +    11 erf(   1 ) +    12 |   1 | erf(   1 )  (10)
                                                   
               The parameters    11 and    12 are positive constants with    11 >    1   /  . Let variable exponential coefficient    1 =
                   0     2
                   1  with    0 and    0 satisfying 0 <    0 < 1 and 1 +    0 <    0.
               1+   0    2
                   1
               Theorem 1 For the second-order subsystem (7), if the disturbance observer is constructed as


                                              ˆ                       1                                (11)
                                                1 =       11 erf(   1 ) +    12 |   1 | erf(   1 )
               then it can estimate    1 accurately in a fixed time. That is to say, the observation error    1 =    1 −    1 can converge
                                                                                     ˜
                                                                                              ˆ
               into a small region within a fixed time.
               Proof of Theorem 1 Select a Lyapunov function as    2 =    , differentiating it, one has
                                                               2
                                                               1
                                   ¤
                                      2 = 2   1 ( ¤   − ¤   1 )

                                           1           1
                                                                                1
                                     = 2   1  (   1 +    1 ) −     1 +    11 erf(   1 ) +    12 |   1 | erf(   1 )
                                                         
                                                                        1
                                                                1
                                     = 2   1 −    11 erf(   1 ) −    12 |   1 | erf(   1 ) +     1
                                                                          

                                                       1
                                                                         1
                                     = −2    11    1 erf(   1 ) −     1    1 +    12    1 |   1 | erf(   1 )
                                                                                                       (12)

                                                        1
                                                                          1
                                     ≤ −2    11    1 tanh(   1 ) −     1    1 +    12    1 |   1 | erf(   1 )
                                                          

                                                            1  
                                                                           1
                                     ≤ −2    11 |   1 | −    11    1 −  |   1 | +    12    1 |   1 | erf(   1 )
                                                           
                                                     1
                                     ≤ −2   12 |   1 ||   1 | erf(|   1 |) + 2   11    1
                                                  1 +1
                                     = −2   12 |   1 |  erf(|   1 |) + 2   11    1
               where    1 is a positive constant.
               Case 1 When    2 > 1 and |   1 | > 1, one has erf(|   1 |) > erf(1) and     0    2 1  ≥     0  > 1. Then (12) can be
                                                                          1+   0    1 2  1+   0
               rewritten as
                                                                           0  +1
                                              ¤
                                                2 ≤ −2(   12 erf(1) −    11    1 )|   1 |  1+   0
                                                                                                       (13)
                                                                        0 +   0 +1
                                                ≤ −2(   12 erf(1) −    11    1 )    2(1+   0 )
                                                                     2
                                               0 +   0 +1
               As    12 erf(1) −    11    1 > 0 and ¯   1 =  > 1, then all the solutions of {   2 > 1} will reach the set {   2 ≤ 1}
                                             2(1+   0 )
                                            1
               within a fixed time       1 ≤         .
                                    2(   12 erf(1)−   11    1 )( ¯   1 −1)
   4   5   6   7   8   9   10   11   12   13   14