200526976 玖、發明說明： 【發明所屬之技術領域】 本發明係關於一種定位方法，尤指一種使用浥亂樣本策 略之室内定位方法及系統。 5 【先前技術】 隨著無線行動設備的日漸普及，越來越多使用者透過行 動裝置完成生活上的各種活動，而各種無線上網環境也與之 遽增。而由於無線環境的成熟，各種無線環境的加值應用便 10 因應而生，其中一項就是以提供位置資訊的位置感知服務， 透過此種位置感知服務，使用者可以獲得在其附近的各種資 訊與服務，例如，使用者可以查詢「最近的餐廳」、「最近 的大眾交通工具」，而服務提供者就可以根據使用者的位置 提供最適合的資訊。200526976 发明 Description of the invention: [Technical field to which the invention belongs] The present invention relates to a positioning method, and particularly to an indoor positioning method and system using a random sample strategy. 5 [Previous Technology] With the increasing popularity of wireless mobile devices, more and more users complete various activities in their lives through mobile devices, and various wireless Internet environments have also increased. Due to the maturity of the wireless environment, various value-added applications of the wireless environment have emerged in response to the 10, one of which is a location awareness service that provides location information. With this location awareness service, users can obtain a variety of information in their vicinity. And services, for example, users can query "the nearest restaurant", "the nearest public transportation", and the service provider can provide the most suitable information based on the user's location.
15 在戶外，最常利用的位置感知裝置為全球定位系統GPS (Global Position System)，透過GPS可以提供方便且正確 的使用者位置資訊，而依硬體設備的不同而有數公尺到數十 公尺的正確度。然而，另一方面，在室内的環境下則尚無一 套廣為大眾所使用之定位系統，其原因之一為誤差度太大， 20 因為室内相對於室外是一個比較小範圍的空間，且室内的格 局也相對比較稠密，所以定位系統的準確度就直接考驗整個 系統的可行性，目前室内定位方法主要可分為兩個步驟，步 驟一主要以 ToA(Time of Arrival)、TdoA(Time Difference of Arrival)、AoA(Angle of Arrival)及 RSSI(Received Signal 6 20052697615 In the outdoors, the most commonly used position sensing device is the Global Positioning System GPS (Global Position System). GPS can provide convenient and accurate user location information. Depending on the hardware device, there are several meters to tens of meters. Ruler accuracy. However, on the other hand, in the indoor environment, there is no set of positioning systems that are widely used by the public. One of the reasons is that the error is too large. 20 Because indoor is a relatively small space compared to outdoor, and The indoor pattern is relatively dense, so the accuracy of the positioning system directly tests the feasibility of the entire system. At present, the indoor positioning method can be divided into two steps. The first step is mainly based on ToA (Time of Arrival) and TdoA (Time Difference). of Arrival), AoA (Angle of Arrival), and RSSI (Received Signal 6 200526976
Strength Indicator)等技術來獲得基本的位置資訊，例如距離 或是角度等’然這些技術都存在一定程度的誤差；再利用這 些位置資訊，在步驟二中配合三角定位演算法或者樣本比對 演异法就可以達到定位的目的，而這個過程中，因為各種因 5素所造成步驟一所得到資訊的誤差會直接反映到步驟二的 定位結果。另一造成無法在室内的環境下提供為大眾所使用 之定位系統的原因在於額外硬體的成本，因為如果使用者進 入個特疋的至内環境，要先配戴一個額外的硬體設備，則 相對的就增加使用者的不便，另一方面也增加系統的成本。 10因此，室内定位一直都無法有一個很有效的解決方法。 【發明内容】 本發明之主要目的係在提供一種使用混亂樣本策略之 至内定位方法及系統，以透過無線訊號強度的特性達到在室 15内環境能夠準確的定位出使用者的位置。 — 為達成上述目的，本發明提出一種使用混亂樣本策略之 室内定位方法，此室内空間之不同處設置有多個無線 存取點’每一行動裝置偵測無線存取點之訊號強 度’此方法包括一訊號累積步驟、一混亂處理步驟、 2〇 疋位步驟及一位置選擇步驟，訊號累積步驟用以 累積多組訊號強度向量，其中一組訊號強度向量代 表仃動裂置在一時間點偵測得到之所有無線存取 的几號強度，混亂處理步驟用以一 一將所有訊號 強度向量之每一訊號強度與其餘訊號強度向量之 7 200526976 其他訊號強度任意組合而產生多組混亂後訊號強 度向量；定位步驟依據前述之多組混亂後訊號強度 向量而分別產生多組位置資訊；位置選擇步驟由前 述多組位置資訊選出一組最佳位置資訊作為最後 5 預測的位置。 本發明亦提出一種使用混亂樣本策略之室内定位系 統，其包括多個無線存取點、一訊號累積模組、一混 亂處理模組、一定位模組、及一位置選擇模組，無 線存取點係設置於一室内空間之不同處，一行動裝 10 置偵測每一無線存取點之訊號強度；訊號累積模組 用以累積多組訊號強度向量，其中一組訊號強度向 量代表行動裝置在一時間點偵測得到之所有無線 存取點的訊號強度；混亂處理模組用以——將所有 訊號強度向量之每一訊號強度與其餘訊號強度向 15 量之其他訊號強度任意組合而產生多組混亂後訊 號強度向量；定位模組依據前述之多組混亂後訊號 強度向量而分別產生多組位置資訊；位置選擇模組 由前述多組位置資訊選出一組最佳位置資訊作為 最後預測的位置。 20 【實施方式】 為能讓貴審查委員能更瞭解本發明之技術内 容，特舉一較佳具體實施例說明如下。 8 200526976 —圖1顯不應用本發明之使用混亂樣本策略之室 =疋位方法的環境。#中，在一室内空間之不同處 ^ 夕個無線存取點"（Access p〇int，ap)，而 當f帶有行動裝置13 (M〇bUe N〇de)之使用者進入 5此，内二間時，無線存取點1 1與行動裝置1 3進行無 線a息父換，例如，無線存取點1 1會持續發出訊標 (beaC〇n)訊號並由行動裝置13接收以得其訊號強 又收集所彳于之訊號將透過無線網路送回一伺服端 1 2並儲存為訊號特徵資料庫（此為訓練階段），而伺 10服、1 2可依照先前行動裝置1 3在不同位置對每_ 無線存取點1 1收集之訊號強度，比對目前行動裝置 1 3於某位置對每一無線存取點丨丨收集之訊號強 度’估測該行動裝置1 3之位置，因而定位使用者所 在之地點。 15 圖2顯示前述伺服端1 2進行定位之運作圖，其首 先由 成號累積模組21(Signal Accumulation Module)累積行動裝置13在某固定位置與時間點收 集每一無線存取點1 1之訊號強度，如圖3所示，S S x y 代表一行動裝置1 3在時間點tx(x= 1〜h，h為正整數） 20 所偵測到無線存取點y(y=l〜η，η為正整數）之訊號 強度，而行動裝置1 3在時間點X偵測得到之所有無 線存取點1 1的訊號強度則以一訊號強度向量SVX表 示。圖中顯示伺服端1 2共累積了· h組訊號強度向量。 9 200526976 再請參照圖2所示，此h組訊號強度向量進一步 由訊號混亂模組2 2 (S i g n a 1 S c r a m b 1 e Μ 〇 d u 1 e)進行 混亂處理（Scramble)，以--將所有訊號強度向量 SVX之每一訊號強度與其餘訊號強度向量之其他訊 5 號強度任意組合而產生hn組訊號強度向量SSVX (x=l〜hn)，舉例而言，2組訊號強度向量SVi^SS!1， SS!2)及SVpUS/，SS22)經混亂處理後將產生22 = 4 組訊號強度向量SSVeCSS〆，SS〗2)、SSVfiSS/， SS22)、SSVhssA SS22)、SSVdSS^，SSi2)。 10 而定位模組23(Location Estimation Module)則 依據前述之hn組訊號強度向量SSVX而分別產生hn 組位置資訊Lx(x=l〜hn)，其中，對每一組訊號強度 向量SSVX’可採用比對此訊號強度向量SSVx之各個 訊號強度SSxy和訊號特徵資料庫而判定其位置資 15 訊。 位置選擇模組 24(Location Selector Module)則 依據移動預測2 5、歷史記錄2 6與地圖知識2 7，而由 前述hn組位置資訊lx選出一組作為最後預測的位 置’其中’移動預測25用以由使用者之前一位置及 20移動方向所預測之可能位置，而賦予每一位置資訊 一個位置權重值wp，歷史記錄26用以由長時間觀察 使用者在此室内空間之移動記錄比對目前位置所 預測之可能位置，而賦予每一位置資訊一個歷史權 重值wh ’地圖知識27係描述此室内空間的地圖，例 10 200526976 如，可行走路徑及阻擋物位置等，以避免定位出不 合理的結果，藉此，位置選擇模組可地圖知識27而 去除不可能之位置資訊，並依據移動預測2 5、歷史 記錄2 6、或移動預測2 5與歷史記錄2 6之組合而求出 5 具有最大權重值之位置資訊，以作為最後預測的位 置。 前述移動預測2 5係基於當使用者移動的時候， 大部分時間都會有一個固定的模式，例如傾向等速 前進，而不會時快時慢或者短時間内作方向上的大 10 幅度改變。所以，以此特性可以增加使用者對於混 亂後的訊號樣本傾向選擇某個特定的方向和距 離。舉例而言，可以藉由短時間内使用者的位置 l(t-l)以及l(t-2)來判斷出移動向量，並且預測下一 點的位置為l(t) + [ l(t-l) - 1〇2)]，就可以簡單的達 15 到預測的目的。除此之外，並可以利用一個對整·個 地圖的加權函數來描述每個位置的可能性，詳細言 之，假設使用者目前位置為（x0，y0 )，距離函示 為d ( X，y) = V(x - ^〇)2 + (y~ y〇)2，歷史平均移動距離 (dCa,y)-d)2 — _i_p 2p^ 為，則任一點的機率p ( x，y ) = VM〆 ， 20 利用高斯（Gaussian)分佈來加權每個位置出現的機 率，而高斯函式中的變異數variance )可以依使 用者的歷史速度變化程度來做調整，圖4即顯示此 範例之移動預測的示意圖。 11 200526976 前述歷史記錄26係具有 使用者的移動模式所歸納 二長時間的觀察所有 新的使用者在移動時，可 &見樣板。以當一個 以及歷史軌跡，判定使艮據使用者目前的位置 板最相似，從而預測下—#歷史記錄中的哪個樣 資訊中，增加選擇預測：：並在混亂出的位置 评頂別位置中的機率。 由以上之說明可知 I々 度的特性達到在室内二“係透過無線訊號強 者的 长兄旎夠準確的定位出使用 ίο 15 样，、’利用混雜訊號樣本的技術來增加定 位樣本空間，另輔以銘&汉t 1以移動預測、歷史經驗及環境描 述來提升整體定位♦铖沾 糸統的準確度，而能夠將環境的 干擾效應減到最低’俾有效改進既有定位系統之穩 定性與正確性。 〜 上述貫施例僅係為了方便說明而舉例而已，本 發明所主張之權利範圍自應以申請專利範圍所述 為準，而非僅限於上述實施例。 【圖式簡單說明】 圖1係應用本發明之使用混亂樣本策略之室内定位 2〇 方法的環境示意圖。 圖2係本發明之使用混亂樣本策略之室内定位系統 之祠服端進行定位之運作圖。 12 200526976 圖3係本發明之使用混亂樣本策略之室内定位系統 之伺服端的訊號累積模組所累積行動裝置在無線 存取點之訊號強度示意圖。 圖4係依據本發明之移動預測的示意圖。 5 【圖號說明】 (1 1) 無線存取點 (12) 伺服端 (13) 行動裝置 10 (21)訊號累積模組 (22) 訊號混亂模組 (23) 定位模組 (24) 位置選擇模組 (25) 移動預測 15 (26)歷史記錄 (27) 地圖知識 13Strength Indicator) and other technologies to obtain basic position information, such as distance or angle. However, these technologies have a certain degree of error; and then use these position information to cooperate with the triangular positioning algorithm or sample comparison in step 2. Method can achieve the purpose of positioning, and in this process, because of various factors caused by 5 factors, the error of the information obtained in step 1 will be directly reflected in the positioning result of step 2. Another reason that the positioning system for public use cannot be provided in an indoor environment is the cost of additional hardware, because if the user enters a special internal environment, he must first wear an additional hardware device, Relatively, the inconvenience of the user is increased, and the cost of the system is also increased. 10 Therefore, indoor positioning has not been able to have a very effective solution. [Summary of the Invention] The main object of the present invention is to provide an internal positioning method and system using a chaotic sample strategy, so as to reach the position of the user in the environment in the room 15 accurately through the characteristics of wireless signal strength. — In order to achieve the above-mentioned object, the present invention proposes an indoor positioning method using a chaotic sample strategy. Multiple wireless access points are set at different places in the indoor space, and each mobile device detects the signal strength of the wireless access point. It includes a signal accumulation step, a chaos processing step, a 20-bit step, and a position selection step. The signal accumulation step is used to accumulate multiple sets of signal strength vectors, where a set of signal strength vectors represents automatic split detection at a point in time. The measured signal strengths of all wireless accesses. The chaos processing step is to combine each signal strength of all the signal strength vectors with the rest of the signal strength vectors. 200526976 Other signal strengths are arbitrarily combined to generate multiple sets of chaotic signal strengths. Vector; the positioning step generates multiple sets of position information respectively according to the aforementioned multiple sets of chaotic signal strength vectors; the position selection step selects a set of best position information from the aforementioned multiple sets of position information as the last 5 predicted positions. The invention also proposes an indoor positioning system using a chaotic sample strategy, which includes a plurality of wireless access points, a signal accumulation module, a chaos processing module, a positioning module, and a position selection module for wireless access. The points are set at different places in an indoor space. A mobile device detects the signal strength of each wireless access point. The signal accumulation module is used to accumulate multiple sets of signal strength vectors, one of which represents the mobile device. The signal strengths of all wireless access points detected at a point in time; the chaos processing module is used to-arbitrarily combine each signal strength of all signal strength vectors with the rest of the signal strength to 15 other signal strengths and generate Multiple sets of signal strength vectors after chaos; the positioning module generates multiple sets of location information according to the aforementioned multiple sets of signal strength vectors after chaos; the location selection module selects a set of best location information from the aforementioned multiple sets of location information as the final prediction position. [Embodiment] In order to allow your review committee to better understand the technical content of the present invention, a preferred specific embodiment is described below. 8 200526976 — Figure 1 shows the environment where the room-to-bit method using the confusion sample strategy of the present invention is not applied. # 中 ， different in an indoor space ^ evening wireless access point " (Access point, ap), and when a user with a mobile device 13 (M〇bUe No.de) enters 5 In the second room, the wireless access point 11 and the mobile device 13 exchange wirelessly. For example, the wireless access point 11 will continue to send a beacon signal and it is received by the mobile device 13 to The strong signal and the collected signal will be sent back to a server 12 through the wireless network and stored as a signal feature database (this is the training phase), while the server 10 and 12 can follow the previous mobile device 1 3 The strength of the signal collected by each wireless access point 1 1 at different locations is compared with the current mobile device 1 3 The strength of the signal collected by each wireless access point at a certain location 丨 丨 estimates the strength of the mobile device 1 3 Location, thereby locating where the user is. 15 Figure 2 shows the positioning operation of the aforementioned server 12. First, the number accumulation module 21 (Signal Accumulation Module) accumulates the mobile device 13 to collect each wireless access point 1 1 at a fixed location and time point. Signal strength, as shown in Figure 3, SS xy represents a mobile device 1 3 at time tx (x = 1 ~ h, h is a positive integer) 20 wireless access point y (y = l ~ η, η is a positive integer), and the signal strengths of all the wireless access points 11 detected by the mobile device 13 at time X are represented by a signal strength vector SVX. The figure shows that the servo end 12 has accumulated a total of h signal strength vectors. 9 200526976 Please refer to FIG. 2 again. This h-group signal strength vector is further scrambled by the signal chaos module 2 2 (Signa 1 S cramb 1 e Μ 〇 DU 1 e) to- Each signal strength of the signal strength vector SVX is arbitrarily combined with the other signal strengths of the other signal strength vectors to generate a hn group of signal strength vectors SSVX (x = 1-hn). For example, two sets of signal strength vectors SVi ^ SS ! 1, SS! 2) and SVpUS /, SS22) will generate 22 = 4 sets of signal strength vectors SSVeCSS〆, SS〗 2), SSVfiSS /, SS22), SSVhssA SS22), SSVdSS ^, SSi2) after chaotic processing. 10 And the Location Module 23 (Location Estimation Module) generates the hn group of position information Lx (x = l ~ hn) according to the aforementioned hn group of signal intensity vectors SSVX. Among them, for each group of signal intensity vectors SSVX ', Compare this signal strength vector SSVx with each signal strength SSxy and the signal feature database to determine its location data. The Location Selector Module 24 is based on the mobile prediction 25, the historical record 26, and the map knowledge 27, and the aforementioned hn group of position information lx selects a group as the last predicted position. Based on the user's previous position and the possible position predicted by the 20 moving direction, each position information is given a position weight value wp, and the history record 26 is used to compare the current movement record of the user in this indoor space by comparing the current Location predicted possible location, and each location information is given a historical weight value wh 'Map Knowledge 27 is a map describing this indoor space, eg 10 200526976 For example, walkable paths and obstacle locations, etc. to avoid unreasonable positioning As a result, the location selection module can map knowledge 27 to remove impossible location information, and obtain 5 based on the combination of mobile prediction 25, historical record 2 or mobile prediction 2 5 and historical record 2 6 Location information with the highest weight value as the last predicted location. The aforementioned movement prediction 25 is based on the fact that when the user moves, most of the time there will be a fixed pattern, such as the tendency to move forward at a constant speed, without changing the direction by a large amount in a short time. Therefore, this feature can increase the user's tendency to choose a specific direction and distance for the scrambled signal samples. For example, the user's position l (tl) and l (t-2) can be used to determine the movement vector in a short time, and the position of the next point is predicted to be l (t) + [l (tl)-1 〇2)], you can easily achieve 15 to the purpose of prediction. In addition, a weighting function for the entire map can be used to describe the possibility of each location. In detail, suppose the user's current location is (x0, y0), and the distance function is d (X, y) = V (x-^ 〇) 2 + (y ~ y〇) 2, the historical average moving distance (dCa, y) -d) 2 — _i_p 2p ^ is, then the probability of any point p (x, y) = VM〆, 20 Gaussian distribution is used to weight the probability of occurrence at each position, and the variation in the Gaussian function can be adjusted according to the historical speed change of the user. Figure 4 shows this example. Illustration of motion prediction. 11 200526976 The aforementioned historical record 26 is summarized by the user's movement pattern. Two long observations of all new users can be & Based on the current and historical trajectories, it is determined to make the user ’s current position board the most similar, so as to predict which information in the # history record, increase the selection prediction :: and in the chaotic position evaluation Chance. From the above description, it can be known that the characteristics of the I degree can be achieved indoors. "The older brother of the strong wireless signal can accurately locate the use of 15 samples." The technique of using mixed signal samples is used to increase the positioning sample space. Using Ming & Han t 1 to improve the overall positioning with mobile prediction, historical experience and environmental description. 铖 铖 The accuracy of the system can be minimized, and the stability of the existing positioning system can be effectively improved. And correctness. ~ The above-mentioned embodiments are merely examples for the convenience of description. The scope of the rights claimed in the present invention should be based on the scope of the patent application, rather than limited to the above-mentioned embodiments. [Schematic description of the drawings] Figure 1 is a schematic diagram of the environment of the indoor positioning method 20 using the chaotic sample strategy of the present invention. Figure 2 is an operation diagram of the positioning of the temple end of the indoor positioning system using the chaotic sample strategy of the present invention. 12 200526976 Figure 3 Series The mobile device accumulated in the signal accumulation module of the server end of the indoor positioning system using the chaotic sample strategy of the present invention is in a wireless access point. Schematic diagram of signal strength. Figure 4 is a schematic diagram of mobile prediction according to the present invention. 5 [Explanation of figure numbers] (1 1) Wireless access point (12) Servo terminal (13) Mobile device 10 (21) Signal accumulation module (22 ) Signal confusion module (23) Positioning module (24) Position selection module (25) Mobile prediction 15 (26) History (27) Map knowledge 13