基于可见光辅助的边缘智能室内外定位系统OA

Global navigation satellite systems(GNSS)provide meter-level positioning accuracy in open outdoor environments but fail completely indoors due to signal attenuation.In contrast,visible light positioning(VLP)technology achieves centimeter-level accuracy indoors but is difficult to apply outdoors due to its dependence on light sources.This paper proposes an edge intelligence-based indoor-outdoor seamless positioning system,which for the first time tightly couples BeiDou Navigation Satellite System(BDS)and VLP through a graph optimization framework,solving the problem of heterogeneous and asynchronous fusion in transition areas.The core innovations include:(1)a multimodal factor graph-based seamless positioning framework that uniformly processes two types of heterogeneous observations,namely BDS pseudorange/carrier phase and VLP angle-of-arrival/received signal strength;(2)an adaptive uncertainty propagation mechanism that achieves imperceptible switching in transition areas through online estimation of optical channel status and satellite geometric dilution of precision;(3)a lightweight positioning solution architecture based on edge computing that reduces the computational load on end devices by 62%.Experimental results in typical indoor-outdoor transition scenarios demonstrate that the proposed system achieves 0.8 m RMS accuracy in open outdoor environments,0.12 m RMS accuracy in deep indoor areas,100%switching success rate within a transition zone of 2-meter range,and a positioning jump of less than 0.25 m caused by switching.Compared with traditional handover-based schemes,Compared to traditional switching schemes,the positioning accuracy is significantly better.