Exploration on emotions continues from past to present. Nowadays, with the rapid advancement of intelligent technology, computer-aided emotion recognition using physiological signals has attracted increasing academic interests. Notwithstanding so, the research progress of this issue seemed to meet a bottleneck in the last decade without any substantial breakthrough. It naturally motivates us to thoroughly review and deeply rethink the current status and the future prospect of this issue. This work has provided a comprehensive overview and systematic taxonomy for physiological-signal-based emotion recognition. To do so, this work has made an unprecedented odyssey from signal properties through solution strategies to intrinsic principles. On the way, this paper has spared no effort to hunt the treasures about methodological essence and philosophical insight, draw the panorama of state-of-the-arts, and outlook the development of technologies. We expect all these discoveries and contributions can be beneficial for inspiring and guiding research innovations in the future.

对情感的探索从过去一直延续到现在。如今，随着智能技术的飞速发展，利用生理信号进行计算机辅助情绪识别已经引起了越来越多的学术兴趣。尽管如此，在过去的十年中，这个问题的研究进展似乎遇到了瓶颈，而没有任何实质性的突破。这自然地促使我们彻底审查并深刻地重新考虑该问题的现状和未来前景。这项工作为基于生理信号的情绪识别提供了全面的概述和系统的分类法。为此，从信号特性到解决方案策略再到内在原理，这项工作取得了空前的成功。在途中，本文不遗余力地寻找关于方法论本质和哲学见解的宝藏，绘制最新技术的全景图，并展望技术的发展。我们希望所有这些发现和贡献都将有助于将来激发和指导研究创新。