State-of-the-Art Elderly Service Robot: Environmental Perception, Compliance Control, Intention Recognition, and Research Challenges

摘要

This article surveys the literature on environmental perception, compliance control, and intention recognition for elderly service robots. Population aging is an inevitable trend in current society, leading to an urgent need for service robots. A high-performance collaborative robot should be able to construct an environment of harmonious and effective human–robot symbiosis. This survey starts with an overview of the environment perception system (e.g., environment 3D reconstruction and object location) for elderly service robots, as environment perception is the basis of human–robot interaction. Furthermore, we report various learning methods deployed in compliance control, which opens the way for the generalization of robots. In this case, we structure our review into three primary domains delimited by learning processes. The article also presents several aspects focusing on processing multimodal signals (e.g., vision, speech, and electromyograms) related to intention recognition. Finally, we briefly discuss several challenges that robots need to overcome to reach interactive performance comparable to humans.