Duowen Liu

TL;DR: A general probabilistic model is presented to shed light on a fundamental issue: how good the RSS fingerprinting based indoor localization can achieve and the interaction among accuracy, reliability, and the number of measurements in the localization process.

TL;DR: In this article, the authors proposed an AP choosing method based on WLAN indoor positioning, comprising steps of measuring signal intensity of all access points in the position by a user, obtaining a complex vector for describing AP signal characteristics according to the AP signal by the system, eliminating AP candidates which are invalid for positioning according to characteristics of the complex characteristics, calculating an optimal signal measurement strategy based on a simulation annealing algorithm according to user-given measurement times, performing measurement on the signal of the access point according to calculated strategy, uploading the measurement result to the server according, calculating the

TL;DR: The design principles of the localization system are revisited with the opportunity provided by the theory of best fingerprints reporting strategy, the localization reliability bounds are presented, and algorithms to find the best strategy in practice are developed.

TL;DR: In this paper, the authors proposed an indoor positioning RSS fingerprint database recovery method based on a propagation model, comprising the following steps: a user gives the sampling amount of positioning; the size of positioning grids in a space (namely, the positioning accuracy) is determined based on channel correlation; the user samples signal strength (RSS values) in specified positions; and a complete positioning database is recovered by only using part of sampling amount acquired by the user according to linear transformation.

TL;DR: In this article , the authors present Runespoor, an edge-cloud video analytics system to manage the tail accuracy and enable emerging robotics applications, which trains and deploys a super-resolution model tailored for the tail-accuracy of analytics tasks on the server to significantly improves the performance on hard-todetect classes and sophisticated frames.