Title: Domino: Dominant Path-based Compensation for Hardware Impairments in Modern WiFi Sensing Show Chinese title

Title: 多米诺：现代WiFi感知中基于主导路径的硬件损伤补偿

Abstract: WiFi sensing faces a critical reliability challenge due to hardware-induced RF distortions, especially with modern, market-dominant WiFi cards supporting 802.11ac/ax protocols. These cards employ sensitive automatic gain control and separate RF chains, introducing complex and dynamic distortions that render existing compensation methods ineffective. In this paper, we introduce Domino, a new framework that transforms channel state information (CSI) into channel impulse response (CIR) and leverages it for precise distortion compensation. Domino is built on the key insight that hardware-induced distortions impact all signal paths uniformly, allowing the dominant static path to serve as a reliable reference for effective compensation through delay-domain processing. Real-world respiration monitoring experiments show that Domino achieves at least 2x higher mean accuracy over existing methods, maintaining robust performance with a median error below 0.24 bpm, even using a single antenna in both direct line-of-sight and obstructed scenarios. Show Chinese abstract

Abstract: WiFi感知由于硬件引起的射频失真面临关键的可靠性挑战，尤其是在支持802.11ac/ax协议的现代主流WiFi网卡中。这些网卡采用灵敏的自动增益控制和独立的射频链路，引入了复杂且动态的失真，使得现有的补偿方法无效。在本文中，我们引入了Domino，一种新的框架，它将信道状态信息（CSI）转换为信道脉冲响应（CIR），并利用它进行精确的失真补偿。Domino基于一个关键见解，即硬件引起的失真对所有信号路径产生均匀影响，使得主导的静态路径可以作为可靠的参考，通过时延域处理实现有效的补偿。现实世界中的呼吸监测实验表明，Domino相比现有方法至少提高了2倍的平均准确率，在直接视距和受阻场景下，即使仅使用单天线，也能保持稳健的性能，中位误差低于0.24次/分钟。