Evolving artificial pain from fault detection through pattern data analysis

Publication Type:
Conference Proceeding
2017 IEEE International Conference on Real-Time Computing and Robotics, RCAR 2017, 2018, 2017-July pp. 694 - 699
Issue Date:
Full metadata record
Files in This Item:
Filename Description Size
Evolving artificial pain from fault detection through pattern data analysis.pdfAccepted manuscript794.99 kB
Adobe PDF
© 2017 IEEE. Fault detection is a classical area of study in robotics and extensive research works have been dedicated to investigate its broad applications. As the breath of robots applications requiring human interaction grow, it is important for robots to acquire sophisticated social skills such as empathy towards pain. However, it turns out that this is difficult to achieve without having an appropriate concept of pain that relies on robots being aware of their own body machinery aspects. This paper introduces the concept of pain, based on the ability to develop a state of awareness of robots own body and the use of the fault detection approach to generate artificial robot pain. Faults provide the stimulus and defines a classified magnitude value, which constitutes artificial pain generation, comprised of synthetic pain classes. Our experiment evaluates some of synthetic pain classes and the results show that the robot gains awareness of its internal state through its ability to predict its joint motion and generate appropriate artificial pain. The robot is also capable of alerting humans whenever a task will generate artificial pain, or whenever humans fails to acknowledge the alert, the robot can take a considerable preventive actions through joint stiffness adjustment.
Please use this identifier to cite or link to this item: