Research On Technologies Of Networked Navigation Based On Dynamics Sensing

Networked navigation is one of hot and difficult issues in wireless network application research. With the wide deployment of wireless sensor networks(WSNs) and wireless local area networks(WLANs), and the rapid popularization of smart mobile devices equipped with various types of sensing unit, the navigation service has gradually been extended to the "last mile" of multifarious dynamic scenarios including disaster relief, asset management, battlefield aggregation and business guidance.Currently, the researchers in the field of indoor positioning and tracking technology have proposed a large number of innovative works. Nevertheless, the excessive dependence on different domain knowledge(e.g. the location of beacon node, the signal strength-distance model, the radio map and the floor plan, etc.), which seriously degrades the navigation system’s performance in real-time, reliability and scalability. To focus on four main factors on the performance of network navigation methods(link dynamic state, regional emergency dynamic, population dynamic behavior and roadmap to produce dynamic), this dissertation in-depth studies the network navigation method by dynamics sensing in WSNs and WLANs. According to the requirement of dynamics sensing, it studies on the following four issues: the network data dissemination reliability, the generation path reachability, mobile population identification and the scalability of navigation road maps. The details are as follows:Firstly, in the light of the dynamics of link quality, this paper proposes a lightweight data dissemination protocol based on the cooperative acknowledgement, which breaks through the constraints of link independence assumption. Taking advantage of a real sensor network experiment, it is observed that there is a significant link correlation between multiple adjacent links to receive the same packet. Based on this observation, a finite state machine model is designed, which contains the state of suspension, waiting, sending and receiving. With the joint estimation of the received acknowledgement probability of the correlation link, which forms by one transmission node and multiple neighbor nodes, the proposed method reduces the redundancy retransmission and the coverage delay of the entire network, and increases the reliability of data dissemination.Secondly, in the light of the environment dynamics, most of the existing work focus on the instantaneous impact of the emergency area on the navigation path, and the passive response pattern is used to recalculate the path, however, the pattern is lack of active estimation of the dynamics of emergency regions which is easy to guide navigation users back and forth in a local(named oscillation movements). This work exploits Consine Theorem and the inherent topological constraints of the network to address this issue. Then, we propose the basic idea for proactive relieving the oscillation movements, which is to translate the dynamic sensing of spacial-temporal variation into the local network topology sensing. Moreover, we design a model based on virtual triangle construction to measure the path reachability, and theoretically prove that the model can effectively alleviate oscillation movements. Finally, our work implements a path reachability measurement-based networked navigation protocol in WSNs. This protocol enables to sense the emergency region’s dynamic characters in real-time pattern, e.g., generation, diffusion, movement and disappearance.Thirdly, in the light of the dynamics of crowded mobility, the network navigation environments are extended to more widespread WLANs from WSNs. we considers the identification problem of the dynamic individual movement behavior and the nearest neighbor crowd in the generative proceed of each navigation path and the entire navigation roadmap. On the dynamic behavior of the mobile crowd, this paper proposes a classification model based on accelerometer samples of smartphone, which is used to recognize mobile individuals. On the basis of statistical analysis and multi-level classification, the navigation system distinguishes the various behaviors. The identification problem of crowd mobility is transformed into the neighbor discovery problem. Finally, by the principal component analysis of behavior similarity, the similarity of background and the spacial-temporal similarity, the movement characteristics of the crowd are recognized.Finally, in the light of the dynamics of navigation roadmap, this paper presents an inverse trajectory mapping based navigation method without the input of roadmap. Firstly, the method first defines the hinges on the navigation path utilizing Dead Reckoning. These hinges are used to divide the users’ trajectories into a discrete segment set. Leveraging the probability graph model into segment bridging, the logical roadmap is generated. Taking advantage of the reasoning and elimination on the factor graph, the logical map is mapping into the physical space. Finally, the navigation path and the floor plan are simultaneously generated from the target to any user in reverse order.