Page 337 - Handbook of Biomechatronics
P. 337

330                                                Borna Ghannadi et al.


          structure. In counterbalance-based control, the device applies active/passive
          counterbalance to the patient limb for gravity compensation (Sanchez
          et al., 2006; Sukal et al., 2006; Stienen et al., 2007; Montagner et al.,
          2007; Jackson et al., 2007; Mihelj et al., 2007). Lastly, if the robotic system
          tracks the performance of the patient using an error-based strategy and adapts
          some features for assistance, this is performance-based adaptive control (Kahn
          et al., 2004; Krebs et al., 2003; Riener et al., 2005).
             As in Fig. 2, there are different methods to implement resistive
          (challenge-based) control. In resistance induced therapy, the robot resists
          patient’s movements (Morris et al., 2004; Patten et al., 2006).In error ampli-
          fication (feedback distortion) therapy, the robot amplifies kinematic (Patton
          et al., 2006a,b), visual (Wei et al., 2005; Brewer et al., 2006; Patton et al.,
          2006b), or tactile errors (Liu et al., 2017). Finally, sometimes constraint-
          induced therapy is used in resistive robotic control (Johnson et al., 2003;
          Shaw et al., 2005).
             Corrective control is a kind of time-independent assistive control, in
          which the assistance is done when there are large tracking, coordination,
          or skill errors. This can be achieved by tunneling, in which an impedance-
          based control is applied at the boundaries of a wider trajectory (Guidali
          et al., 2011; Klamroth-Marganska et al., 2014; Mao et al., 2015). Coordina-
          tion (synergy-based) control prevents large coordination errors between
          joints during a rehabilitation task (Guidali et al., 2009; Brokaw et al.,
          2011; Crocher et al., 2012). Finally, haptic provoke is used for providing
          real-world experience based on gaming control schemes (Burdea, 2003;
          Patton et al., 2004; Broeren et al., 2006; Yeh et al., 2013).
             It was mentioned in Section 1 that optimal care is of great importance
          for rehabilitation robotics. This optimal care can be achieved only if the
          robot has an understanding of the coupled human-robot rehabilitation
          system. Thus, one major stream of recent studies is dedicated to the
          improvement of triggered passive control methods, which will be dis-
          cussed in Section 7: Recent developments and research opportunities.
          Patient preparation is the downside for the direct use of biosignals (trig-
          gered passive control); however, partially assistive controllers use internal
          bio-inspired models of the patients to make decisions. Consequently,
          another major stream of recent research is focused on partially assistive
          control methods since these devices can assist the patients using some
          helpful bio-inspired information. Later in Section 7, recent develop-
          ments and research opportunities, some of these developments, will be
          discussed.
   332   333   334   335   336   337   338   339   340   341   342