Technologies fusing WiFi-based indoor positioning and pedestrian dead reckoning for indoor target localization have already been proposed for a variety of occasions. Among them, the methods of multisource fusion with particle filter perform well in a random noise environment. However, the positioning accuracy can be significantly affected by the personnel density of the interest area. In this paper, we propose a novel indoor collaborative positioning system which performs dynamic user pairing and adaptive particle-pair filtering with the help of the Chirp acoustic signal distance measurement. The proposed scheme mainly includes two parts in each time slot: first, the users are clustered into cells for dynamic user-pairing, and then the adaptive particle-pair filter iterates the comprehensive weight to obtain the final positioning result. The experiment results show that the proposed scheme can effectively solve the problem of the error increase caused by the intensive personnel, and remarkably improve the indoor positioning accuracy in the crowded public.