一個不限制平面平行的LED-PD室內三維定位系統最佳化考量硬體指向與朗博次方參數

Abstract

隨著科技與物連網的發展，各領域對於量測資訊的需求大量增加，其中相對位置的資訊實為重要。然而，現今室內定位仍仰賴多個參考點進行定位，缺乏僅以「可攜式單位」達到兩物體之間三維定位的方法。因此，我們由探討不同的室內定位方法開始，根據以上需求將研究重點聚焦在光波段中發光二極體(Light Emitting Diode，簡稱LED)與光電二極體(Photodiode，簡稱PD)的定位方法，此方法有達到兩單位之間三維定位的潛力，但現今文獻中對使用情境與系統設計的限制仍許多，大多需限制接收與發射平面平行且僅能達到二維定位，除此之外也會對硬體進行限制。因此，本研究針對LED與PD的定位方法，建立一個可以不限制接收與發射平面平行的三維定位演算法，也不限制硬體的朗博次方(Lambertian Order)、硬體數量以及各硬體的擺放指向，使系統具有根據不同情境進行改變與設計的能力，改善此領域中對系統設計以及使用情境較多的限制。建立演算法後，本研究由該演算法建立一模擬環境與系統成效的量化方式，於模擬環境中，我們可以評估各系統設計下的定位效能，並透過改變不同的系統設計、使用情境與應用範圍(Region of Interest)、以及誤差參數，來觀察以及探討各參數對系統成效的影響。除此之外，本研究針對不同的使用情境，將系統設計作為變數進行最佳化，可將該最佳系統設計作為實際硬體系統搭建的參考。總結來說，本研究提供一僅利用兩單位進行三維定位的定位演算法，並建立一模擬環境，讓使用者得以在不浪費硬體搭建的成本下對系統設計進行分析與評估，也可以針對特定的使用情境進行系統設計的最佳化，在硬體系統搭建前達到有效的評估。

The development of the Internet of Things(IoT) increases the demand of sensors and data, especially positioning information. Current indoor positioning systems rely on multiple reference points in obtaining position information. An efficient approach using portable devices to obtain their 3-dimensional relative position is yet to be developed. In this research, we start from investigating existing positioning techniques. In order to have an efficient way to measure 3D relative position using portable devices, we focus on light emitting diode(LED) and photodiodes(PD). Most current research activities in LED and PD positioning can only be used in limited scenarios with certain given hardware configurations. Specifically, they could only calulcate 2D position data while restricting transmitting and receiving planes to be parralel. We develope a positioning algorithm to abtain 3D position without parallel planes assumption. In addition, the Lambertian order and hardware orientation are flexible with respect to different scenarios. A simulation environment is set up to quantify the performance of the position system in different scenarios. By adjusting system design variables and simulated noise respectively, the influence of each variable can be discussed. We then propose an optimization method to find the optimized system design in specific scenarios. Overall, the proposed algorithm can provide flexible 3D position system design without parallel planes assumption while considering Lambertian order, hardware amount and placing orientation as variables. The simulation environment provides users an approach to evaluate positioning system performance without the cost of setting up the actual hardware system. And the optimized result provides users a guide to construct the hardware system whenever he/she has the need to measure 3D position of multiple objects.