Generation of an Indoor 2D Map and Track Encryption Based on Mobile Crowdsourcing

Abstract

The widespread application of mobile crowdsourcing modes provides new ideas for generating indoor maps. By collecting and analyzing the trajectory datas of users properly, we can obtain the location information of indoor paths. Unfortunately, currently studies usually rely heavily on a satellite location, which restricts their indoor application. In this paper, a simple and practical method of generating indoor maps on Andriod platform is presented, and this method is able to correct deviation duly. User's datas collected by several bulit-in sensors are preprocessed utilizing Gaussian filter, after which we adopt feature recognition to confirming one's walking track based on multiple experiment datas. In order to integrate tracks generated by different persons, we then propose a new data structure based on a transition probability that can be updated online to store track information. In addition, we minimize possible deviations by testing the signal power launched by four Bluetooth base stations. Discrete tracks are finally integrated into a complete indoor map using a graph_based model. We then propose a novel encryption scheme exploiting chaos in a nonlinear digital filter, where secure key generation methods are discussed in detail. The secure key scheme includes: 1)channel measurement 2)a decorrelation transform 3)multibit adaptive quantization and encoding. Experiments are conducted in rectangle fields of 8m*8m, 44m*44m, respectively, and the results show our method can attain a maximum error of 5.94%.