Page 296 - Autonomous Mobile Robots
P. 296

Adaptive Control of Mobile Robots                          283

                                                                        ˜
                                                                ˜ ˜ ˜
                                     ˙
                              We have V ≤ 0. Accordingly, ˜u, ˜ θ, Z 2 , e I , L, R, K a , K N are all bounded in the
                              sense of Lyapunov.
                                 From Equation (7.55), using the Corollary of Baralart’s theory [31], ˜u → 0
                                          → 0 (1 ≤ j ≤ m − 1) as t →∞, and e I → 0as t →∞.
                              as t →∞, z j,n j
                                 Next, let us prove the asymptotic stability of Z.
                                 The first equation of the controlled system is
                                                                                       (7.56)
                                                  ˙ z 1 =−k u1 z 1 + h(Z 2 , t) +˜u 1

                              From Assumption 7.3, we know that h(Z 2 , t) is uniformly bounded. In addition,
                              with ˜u 1 converging to zero, (7.56) is a stable linear system subjected to the
                              bounded additive perturbation h(Z 2 , t) + e u1 . Therefore, z 1 (t) is also bounded
                              uniformly.
                                 Because z 1 and h(Z 2 , t) are bounded, it is clear that u d1 is bounded from
                              (7.30). Together with ˜u converging to zero, u 1 is bounded. Since u 1 and Z 2
                              are bounded, ˜u goes to zero, u dj and u j (2 ≤ j ≤ m) are bounded. Under
                              the condition that Z 2 , u 1 , and u j (2 ≤ j ≤ m) are bounded, ˙z j,n j  and ˙z j,i
                              (1 ≤ j ≤ m − 1, 2 ≤ i ≤ n j − 1), from (7.25), are bounded.
                                 In the following, let us show that u d1 Z 2 tends to zero. For 1 ≤ j ≤ m − 1,
                                                                   2
                                                                   d1
                              since u d1 is bounded and z j,n j  tends to zero, u z j,n j  tends to zero. Taking the
                                             2
                                             d1
                              time derivative of u z j,n j ,wehave
                                    d
                                       2        2
                                                d1
                                       d1
                                     (u z j,n j  ) = u (−k u j +1 z j,n j  − ρ j,n j −2 u d1 z j,n j −1
                                    dt
                                                                    d    2
                                                                     (u d1 )
                                               +˜u 1 L h1 z j,n j  +˜u 2 ) + z j,n j
                                                                   dt
                                                       3                          2
                                                       d1                         d1
                                             =−ρ j,n j −2 u z j,n j −1 + (2˙u d1 u d1 z j,n j  − k u j +1 u z j,n j
                                                  2            2
                                                  d1           d1
                                               + u ˜u 1 L h1 z j,n j  + u ˜u 2 )       (7.57)
                              Since
                                             d  3        3          2
                                              u z j,n j −1 = u ˙z j,n j −1 + 3u ˙u d1 z j,n j −1
                                                         d1
                                                                    d1
                                               d1
                                            dt
                              is bounded, the first term in (7.57) is uniformly continuous. Together
                                                                                         and
                              with the fact that all other terms in (7.57) tend to zero (since u d1 z j,n j
                                                                                       2
                              ˜ u tend to zero), from the extended version of Barbalat’s Lemma, (d/dt)(u z j,n j )
                                                                                       d1
                                                   3
                              tends to zero. Therefore, u z j,n j −1 also tends to zero. So u d1 z j,n j −1 also tends
                                                   d1
                              to zero.
                              © 2006 by Taylor & Francis Group, LLC



                                FRANKL: “dk6033_c007” — 2006/3/31 — 16:43 — page 283 — #17
   291   292   293   294   295   296   297   298   299   300   301