Addressing the challenge of Wi-Fi signal susceptibility to environmental fluctuations, cumulative errors in pedestrian dead reckoning (PDR) system and the inability of unscented Kalman filter (UKF) to convey additive nonlinear process noise covariance in the indoor UKF fusion localization algorithm of Wi-Fi signal and PDR, an innovation augmented unscented Kalman filter(IAUKF) fusion algorithm was introduced. The Sage-Husa filtering enhanced the algorithm's resistance to signal singularities by real-time adjustment of the supervisory covariance using innovation vectors. While ensuring that high-order moment information can be transmitted, the augmented form of filtering also autonomously assessed the process noise covariance. This augmentation not only streamlined computational processes but also enhanced the autonomy of covariance estimation compared to hypothetical statistical estimations, thereby improving the overall algorithm accuracy. The final experimental results demonstrate that the indoor pedestrian positioning accuracy of IAUKF improves by 30.7% compared to UKF algorithm, with an average step error of 0.523 m.