Page 74 - Human Inspired Dexterity in Robotic Manipulation
P. 74

70    Human Inspired Dexterity in Robotic Manipulation


                                       1  ð T  2
                                   J¼      u ðtÞ dt,                  (5.38)
                                       2  0
          under the dynamic constraint (5.36) and the boundary conditions
                             T                          T
          xð0Þ¼ 0,0,0,0,0,0,0½  Š , and xðTÞ¼ L,0,L,0,L,0,0Š . Proceeding along
                                           ½
          the same line of reasoning as in the case of the minimum hand-jerk model,
          one establishes the trajectory of the mass center in a form structurally similar
          to Eq. (5.23). It can be shown then that the hand trajectory, established from
          Eq. (5.31), will have the following representation:

                                                                 !
                        5       2
                       X    k  X
             x h ðtÞ¼ L   α k t +  β sinðω k t + φ Þ + γ tsinðω k t + ψ Þ ,  (5.39)
                                                   k
                                                               k
                                               k
                                    k
                       k¼0      k¼1
          where the constant coefficients α k , β k , γ k , φ k , and ψ k are functions of ω 1 ,ω 2 ,
          and T.
          5.4 HAND MASS IDENTIFICATION
          To use the minimum hand-force-change model, one needs to know the
          mass of the hand. To estimate the parameter m h reliably, it is not enough
          to know the mass of the hand predicted by anthropometric measures, that
          can be done as, for instance, in [20]. This parameter actually accumulates
          three factors: the mass of the hand itself, the configuration-depended inertia
          of the human arm in multijoint movements reduced to the hand, and the
          end-point inertia of the haptic device.
             Typically the human hand is modeled as a mass-damper-spring system,
          and this model holds for small movements and for a short period of time.
          Based on this model, the hand impedance can be evaluated by perturbing,
          by a robotic manipulandum, the hand during maintenance of a given pos-
          ture, and measuring the hand displacement. This perturbation technique is
          reported in [21–23]. However, its use for a lightweight haptic device with-
          out additional hardware, such as the one proposed in [24], is problematic.
             A simple method, based on the accommodation to forced vibrations and
          applicable to the estimation of the hand mass only, can be constructed as fol-
          lows. Assume that a human subject can follow the motion of a virtual object
          of mass m o , connected to the hand by a spring of stiffness k and a damper of
          viscosity b, without developing his or her own driving force (see Fig. 5.2).
          Assume also that a periodic force of amplitude F and frequency Ω is applied
          to the hand, and the total haptic force is defined as
                        f haptic ¼ bð_x o   _x h Þ + kðx o  x h Þ + F cosΩt:  (5.40)
   69   70   71   72   73   74   75   76   77   78   79