Page 165 - Adaptive Identification and Control of Uncertain Systems with Nonsmooth Dynamics
P. 165

APPC of Strict-Feedback Systems With Non-linear Dead-Zone  161

                                                                    2 2
                                            ˆ
                                     k 1z 2 1  θ 1 |z 1 |  T        ˆ ε z    z 1e ˙μ
                                                                    1 1
                                 ˙
                                 ξ 1 =    +        (Z 1 )  1 (Z 1 ) +      −          (10.14)
                                                  1
                                       r     2η 2                ˆ ε 1 |z 1 |+ σ 11  μ
                                               1

                                            ˙ ˆ     |z 1 |  T             ˆ
                                            θ 1 = r  1  2    (Z 1 )  1 (Z 1 ) − σ 12 θ 1  (10.15)
                                                         1
                                                     2η
                                                       1
                                                   ˙
                                                   ˆ ε 1 = r  a1 |z 1 | − σ 13 ˆε 1   (10.16)
                            where z 1 is the transformed error defined in (10.9), r can be calculated
                            based on e(t), μ(t),and   1 > 0,   a1 > 0, k 1 > 0,η 1 > 0and σ 11 ,σ 12 ,σ 13 > 0
                            are design parameters. It should be noted that the fact ˆε 1 (t) ≥ 0,t ≥ 0 holds
                            for any initial conditions ˆε 1 (0) ≥ 0 based on (10.16). Thus the term ˆε 1 |z 1 |+
                            σ 11 in (10.13) is always positive (i.e., ˆε 1 |z 1 |+ σ 11 > 0), and there is no
                            singularity problem in the proposed control design.
                               Consider the following Lyapunov-Krasovskii function
                                           m 1
                                  1  2  c 11     e  à 1m 1     t  − (t−ς) 2    1  2    1   2
                                                                                 ˜
                             V 1 = z +                    e     k (x 1 (ς))dς +  θ +      ˜ ε 1
                                    1
                                                                 1j
                                                                                  1
                                                 τ
                                  2     2     1 −¯ 1  t−τ 1j (t)              2  1    2  a1
                                           j=1
                                                                                      (10.17)
                                                                         , ¯ i are positive scalars
                                                                           τ
                            where c 11 > 0,  > 0 are positive constants, and τ im i
                                                                                   ˜
                                                                                           ˆ
                            defined in Assumption 10.1,and ˜ε i = ε − ε i ,i = 1,···n−1and θ i = θ −θ i ,
                                                                                        ∗
                                                             ∗
                                                             i                         i
                                                                                     ∗
                            i = 1,···n are parameter errors between the bounded constants ε = ε iN +
                                                                                     i
                             2
                            η /2, θ = W ∗T W , and their estimations ˆε i, θ i.
                                  ∗
                                            ∗
                                                                    ˆ
                             i    i         i
                                        i

                               Consider h ij (t, ¯x i ) ≤ k ij (¯x i ) with k ij (¯x i ) ≥ 0,thetimederivativeof V 1


                            along (10.13)–(10.16) can be given as
                                     ⎛                                                   ⎞
                                                             m 1

                              ˙ V 1 ≤z 1r f 1 (x 1 ) + g 1 (x 1 )(z 2 + α 1 ) +          ⎠
                                                                h 1j (x 1 (t − τ 1j (t))) − e ˙μ/μ −¨y d
                                     ⎝
                                                             j=1
                                        m 1      à 1m
                                     c 11     e   2         2
                                  +              k (x 1 (t)) − k (x 1 (t − τ 1j (t))) −  V d1
                                                            1j
                                                  1j
                                               τ
                                     2      1 −¯ 1
                                        j=1
                                     1   ˙    1
                                                  ˙
                                       ˜ ˜
                                  +    θ 1 θ 1 +  ˜ ε 1 ˜ε 1
                                       1       a1
                                  m 1r 2 M  2  c 11    m 1  2
                                ≤      z +          h (x 1 (t − τ 1j (t)))
                                                     1j
                                        1
                                   2c 11    2    j=1

                                  + z 1r f 1 (x 1 ) + g 1 (x 1 )(z 2 + α 1 ) − e ˙μ/μ −¨y d

                                     c 11    m 1  e  à 1m  2   2
                                  +                 k (x 1 (t)) − k (x 1 (t − τ 1j (t))) −  V d1
                                                     1j
                                                               1j
                                     2    j=1 1 −¯τ 1

                                        |z 1 |  T
                                  − θ 1r   2    (Z 1 )  1 (Z 1 ) − σ 12 θ 1 −Èε 1r |z 1 | − σ 13 ˆε 1
                                                              ˆ
                                     ˜
                                             1
                                        2η
                                           1
   160   161   162   163   164   165   166   167   168   169   170