“…Our formalism and analysis build on machine learning, control theory, and human-robot interaction to provide insight into shared control. We suggest possible challenges, as well as opportunities that could arise from the tight interaction between the robot and the user: adaptation to the context and (Rosenberg, 1993) no (Marayong et al, 2003) (Debus et al, 2000) (You and Hauser, 2011) (Kim et al, 2011) (Marayong et al, 2002) (Leeper et al, 2012) predefined behaviors (Demiris and Hayes, 2002) (Fagg et al, 2004) (Crandall and Goodrich, 2002) no (Aigner and McCarragher, 1997) (Gerdes and Rossetter, 2001) (You and Hauser, 2011) (Marayong et al, 2002) predefined behaviors (Yu et al, 2005) no (Kofman et al, 2005) (Shen et al, 2004) (Smith et al, 2008) predefined behaviors MPC/minimum-jerk (Anderson et al, 2010) (Loizou and Kumar, 2007) (Weber et al, 2009) predefined behaviors (Aarno et al, 2005) (Yu et al, 2005) (Vasquez et al, 2005) fixed goals (2D) no (Ziebart et al, 2009) fully flexible (2D) no the user, predicting and expressing intent, and capitalizing on the user's reactions. These challenges and opportunities are not only applicable to shared control, but conceivably to humanrobot collaboration in general.…”