Autonomy in socially assistive robotics: a systematic review

Area of Science:

• Robotics
• Human-Robot Interaction
• Artificial Intelligence

Background:

• Socially Assistive Robots (SARs) are increasingly deployed across education, healthcare, and service sectors.
• SARs must interact socially with diverse populations, from children to the elderly, posing technical and social challenges.
• A shift from Wizard-of-Oz (WoZ) methods to more autonomous SARs is evident in recent research.

Purpose of the Study:

• To systematically review the last 5 years of SAR research, focusing on the level of autonomy.
• To identify automated sensing and actuation capabilities, and potential research gaps.
• To explore the transparency of SAR autonomy and capabilities to users.

Main Methods:

• Systematic literature review following the PRISMA method.
• Analysis of SARs' sensing and actuation automation levels.
• Examination of transparency in communicating SAR autonomy and capabilities.

Main Results:

• Focus on fully and semi-autonomous SARs, analyzing their automated capabilities.
• Identification of potential gaps in SAR autonomy and automation.
• Discussion on the benefits and drawbacks of transparency in SAR communication.

Conclusions:

• The review provides a holistic view of SAR characteristics, roles, and application domains.
• Highlights the need for further research into SAR autonomy and transparent user communication.
• Informs future development of SARs for diverse societal roles.