Kalman Filter and Cross-Multiply Algorithm With Adaptive DC Offset Removal
Li Zhang, Chang–Hong Fu, Zhongxu Zhuang, Xuan Yang, Genming Ding, Hong Hong, Xiaohua Zhu
Abstract
Recently, microwave radar technologies are taking more important roles in human vital signs monitoring. The modified differentiate and cross multiply is the state-of-the-art demodulation algorithm to extract the phase shifts of the microwave signal modulated by vital signs. However, this method suffers from the interference of the direct current (dc) offset. In order to solve this problem, a Kalman filter and cross-multiply (KFCM) approach is proposed. With a Kalman filter, the dc bias of the I/Q channel could be estimated and removed. Moreover, a real-time circle fitting algorithm is raised to compensate for the loss of phase information. Simulation results show that the proposed KFCM algorithm effectively eliminates the coupling caused by the dc offset and improves the performance of demodulation in the presence of noise. Finally, two practical scenarios, including real-time human vital sign detection and long-term breathing monitoring during sleep, are demonstrated to show the potential applications of this algorithm.