disturbance term (Choi  &  Kwak,  2003;  Eom,  Suh,  & Chung, 1997; Hu & Muller, 1996; Komada, Miyakami, Ishida,  &  Hori, 1996; Park & Lee, 2007; Zhongyi, Fuchun, & Jing, 2008)。 Efficient suppression of these disturbances decouples the dynamics of the

 joints and allows simple controllers to be designed for each DOF。

0967-0661/$ - see front matter & 2012 Elsevier Ltd。 All rights reserved。 http://dx。doi。org/10。1016/j。conengprac。2012。10。008

Fig。 1。  Block diagram of a typical disturbance observer in a robotic  application。

Table 1 Disturbance observer applications in robotics。

Disturbance rejection Independent joint control Zhongyi et al。 (2008), Park and Lee (2007), Choi and Kwak (2003), Eom et al。  (1997),

Komada et al。 (1996), and Hu and Muller  (1996)

Friction estimation and compensation Bona and Indri (2005), Sawut et al。 (2001), and Mohammadi, Marquez et al。   (2011) Time-delayed and transparent teleoperation control  Mohammadi, Tavakoli et al。 (2011) and Natori et al。 (2006, 2007,   2010)

Force/torque estimation  Sensorless force control Shimada et al。 (2010), Katsura et al。 (2003), Eom et al。 (1998), and Lee et al。   (1993) Shadow robot system Katsura et al。 (2010)

Fault detection Sneider and Frank (1996), Chan (1995), and Ohishi and Ohde (1994)

 and compensation (Bona & Indri, 2005; Sawut, Umeda, Park, Hanamoto, & Tsuji, 2001)。 In Mohammadi, Marquez, and Tavakoli (2011), the authors proposed a nonlinear disturbance observer- based  control  law  that  guaranteed  asymptotic  trajectory  and    dis-

 turbance  tracking  in the  presence  of slow-varying disturbances。

One  important  aspect  of  disturbance  observer-based  friction

 compensation schemes is in that they are not based on any particular friction models (Bona & Indri, 2005)。 Disturbance observers   have   recently   been   used   in   time-delayed  bilateral

 teleoperation in order to improve the transparency of telerobotic systems (Natori, Tsuji, & Ohnishi, 2006); (Natori, Kubo, & Ohnishi, 2007); (Natori, Tsuji, Ohnishi, Hace, & Jezernik, 2010)。 In time- delayed teleoperation, the delayed position/force signals are received from  the  communication  channel  in  the  master and the slave sides。 The time-delayed position/force signals are then added to the output of the disturbance observer in order  to provide the master and the slave robots with estimation of the undelayed versions of the position/force signals and thus improve the teleoperation system transparency (Natori et al。, 2010)。 In Mohammadi, Tavakoli, and Marquez (2011), the authors imple- mented a pair of nonlinear disturbance observers in a 4-channel bilateral teleoperation architecture to achieve full transparency in the absence of communication time delays in both free and constrained motions。 In that work, however, the availability  of joint acceleration measurements, which is necessary for achieving full transparency, simplified design of disturbance    observers。

Besides   disturbance   rejection,   disturbance   observers   have

 found applications in other robotic contexts。 In many robotic applications,   the   robot   end-effector   comes   in   contact   with a