Desktop version

Home arrow Computer Science arrow Robotic Assistive Technologies: Principles and Practice


Adaptation, Training, and Learning

Adaptation, training, and learning need to occur on the side of both the robot and the human for rehabilitation robots to be effectively used. For example, a robotic system can adapt and learn the behaviors it needs to interact with a person to improve engagement in cognitive training activities (Chan and Nejat 2012). Learning and adaptation have also been used to improve cognitive outcomes in these types of activities by adapting the level of difficulty of the activity to a user’s performance (Tapus, Tapu§, and Mataric 2009). Similarly, users often need to learn how to use rehabilitation robots as the majority of them, at least initially, will be novice users (Huttenrauch and Eklundh 2002). In some cases, humans will even adapt to their rehabilitation robots despite an unnatural user interface design because they are able to exploit consistent control signals (Jiang et al. 2014).


Rehabilitation robots are collocated with their users, allowing for proximate interactions (i.e., robot and human interacting in the same location). Due to these proximate interactions and the intended population of interest, physical features of rehabilitation robots often play an important role in the overall acceptance of such systems. For example, healthy adult users prefer a human-like robot for socially assistive tasks, such as exercise instruction (Goetz, Kiesler, and Powers 2003). Alternatively, children with autism are more accepting of interacting with a robot that has no facial features over a human-like robot (Robins, Dautenhahn, and Dubowski 2006). In the case of wearable physical rehabilitation robots such as robotic prosthetics, it is important for such systems to conform to human-like characteristics to be accepted by the intended users (Dellon and Matsuoka 2007).

Length of Exposure

The intended user’s length of exposure to a rehabilitation robot has been shown to have varying effects on the HRI. For example, the novelty effect (Kanda and Ishiguro 2013) can be observed: users are excited to interact with the robot at first; however, over time they lose interest, which affect the moods of the users (Wada et al. 2002) and ultimately lead to the user no longer interacting with the robot (Kanda et al. 2004). This, of course, is disconcerting as the user will no longer receive the benefits associated with using these robots. Therefore, it is necessary to consider the interaction time length in the robot design and to investigate the efficacy of rehabilitation robots over long-term studies to ensure that rehabilitation intervention outcomes are not a result of the novelty of the technology.

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >

Related topics