Proactive interaction management in domestic service robots to enhance human-robot conversational interaction

Abstract

Robotics technology has long been renowned for its ability to handle com- plex, repetitive, and sometimes risky tasks that are impossible for humans.

In industries, robots are preferred for precise operations in various manu- facturing stages of a product. On the other hand, domestic service robot is

the other side of robotic technology that brings the smart robotic system into our homes. Service robots are specially designed to interact and assist

humans where they provide a range of services, including caregiving, com- panionship, and entertainment. In addition, service robots can elevate living

standards by assisting with essential daily activities, especially for people with disabilities.

Nowadays, the incorporation of generative pretrained models in robotic sys- tems to provide a fluid conversation between humans and robots is a highly

researched topic. Though pretrained generative language models can be incorporated into conversational systems, they often fail to offer healthy conversations with users. Most pretrained models are trained on billions of data points from blogs, posts, and forums, where language quality is not

a concern. However, language quality and clarity matter greatly when de-signing robotic systems to interact with humans, especially the elderly and children