201034410 . 六、發明說明： 【發明所屬之技術領域】 本發明係有關於一種無線感測網路系統的錯誤感測節點 偵測方法及感測方法，且特別係有關於可減少感測錯誤率之無 線感測網路系統的錯誤感測節點偵測方法及感測方法。 【先前技術】 由於微型製造技術、通訊技術及電池技術的進步，目前已 發展出微型感測器’用以感應、無線通訊及處理資訊。微型感 測器可感應及偵測環境的目標物及改變，並可處理收集到的數 據，再將處理過後的資料以無線傳輸的方式送到資訊融合中心 或基地台（Base Station)。而無線感測網路（Wireless Sensor201034410. VI. Description of the Invention: [Technical Field] The present invention relates to a method and a method for detecting an error sensing node of a wireless sensing network system, and particularly relates to reducing a sensing error rate The error sensing node detecting method and sensing method of the wireless sensing network system. [Prior Art] Due to advances in microfabrication technology, communication technology, and battery technology, miniature sensors have been developed to sense, wirelessly communicate and process information. The micro sensor senses and detects the target and changes of the environment, and can process the collected data, and then send the processed data to the information fusion center or base station by wireless transmission. Wireless Sensor Network (Wireless Sensor)
Networks)係由一或多個無線資料收集器以及多個感測器 (Sensor)所構成的網路系統’其中無線感測網路中的通訊方式係 採用無線通訊方式’因此，感測器或是無線資料收集器可方便 地設置於任意位置，並可節省佈線費用。 然而’在無線感測網路中，往往不可避免地存在錯誤的感 測節點’感測器可能因其能源用盡或硬體損壞而送出不正確的 訊息到資訊融合中心’因而造成系統之估計正確度的下降，並 導致整體的估測準確度效能降低。 【發明内容】 因此本發明之一方面係在於提供一種無線感測網路系統 的錯誤感測節點偵測方法及感測方法，藉以提高無線感測網路 系統的感測準破率。 根據本發明之實施例，本發明之無線感測網路系統的錯誤 3 201034410 - 感測節點偵測方法係用以偵測至少一錯誤感測節點，其中無線 感測網路系統包含複數個感測節點和資訊融合中心，感測節點 係用以感測偵測目標’並分別傳送測量值至資訊融合中心，此 方法包含：接收感測節點的測量值；假設感測節點的其中至少 一者為至少一假設錯誤感測節點；對假設錯誤感測節點以外之 其他感測節點的觀察值進行同質性檢測，以檢測假設錯誤感測 節點以外之其他感測節點的測量值是否實質相同；當檢測出假 設錯誤感測節點以外之其他感測節點的測量值為實質相同 0 時，判斷假設錯誤感測節點為錯誤感測節點；以及當檢測出假 設錯誤感測節點以外之其他感測節點的觀察值為實質不相同 時，判斷假設錯誤感測節點不是錯誤感測節點，並重新假設新 的至少一假設錯誤感測節點，以重新進行同質性檢測。 又，根據本發明之實施例，本發明之無線感測網路系統的 感測方法係用以估測偵測目標，其中無線感測網路系統包含複 數俩感測節點和資訊融合中心’此方法包含：利用感測節點來 感測偵測目標，並分別傳送測量值至資訊融合令心；利用資訊 〇 融合中心來根據感測節點的測量值，而進行錯誤感測節點偵 測，以偵測至少一錯誤感測節點；以及排除錯誤感測節點的測 量值，並根據錯誤感測節點以外之感測節點所傳送的測量值來 進行估測。 又，根據本發明之實施例，本發明之無線感測網路系統包 含資訊融合中心及複數個感測節點。感測節點係用以分別感測 偵測目標，而接收到觀察值，並分別傳送測量值至資訊融合中 心，其中感測節點的其中至少一者為錯誤感測節點。資訊融合 中心係用以接收感測節點的測量值來進行估測，其中資訊融合 中心包括摘錯單元，用以根據感測節點的測量值來偵測出錯誤 201034410 感測節點’資訊融合中心係招擔纽μ a ““旦 係根據錯誤感測節•點以外之感測節點 所傳送的測量值來估測偵測目標。 =此，本發明之無線感測網路系統的錯誤感測節點備測方 ί及感測方法可預域測出異常的感測節點，藉以排除異常的 感測節點所發出錯誤的感測資訊，因而可提高無線感測網路系 統的感測正確率。 【實施方式】 ❹ $讓本發明之上述和其他目的、特徵、優點與實施例能更 明顯易懂’本說明書將特舉出一系列實施例來加以說明。但值 得注意的係，此些實施例只係用以說明本發明之實施方式而 非用以限定本發明。 請參照第1圖和第2圖，第1圖緣示依照本發明之-實施 例之無線感測網路系統的系統示意圖，第2圖繪示依照本發明 之一實施例之資訊融合中心的結構示意圖。本實施例之無線感 測網路系統100例如為分散式網路偵測系統，用以感測至少一 〇偵測目標（未繪示），而可應用於例如軍事或環境監控。無線感 別網路系統1〇〇包含資訊融合中心（Fusi〇n center)丨2〇和複數個 感測節點110。感測節點110可任意地設置於偵測目標的周圍， 以進行感測，並對所接收到的至少一觀察值來進行初步之目標 事件的判斷，例如是否發生火災事件。感測節點110可再將所 判斷的測量值傳送至資訊融合中心120來做最後的事件判斷。 虽無線感測網路系統i 〇〇之感測節點丨1〇的其中至少一者因能 源用盡、硬體損壞或意外故障而發生異常時，此異常的感測節 •點（即錯誤感測節點1U)會傳送與真實觀測狀況不符合的資料 到資訊融合中心120，而這些不正確的資料會影響到資訊融合 201034410 • 中心120的參數估計，導致無線感測網路系統loo的感測正確 率大幅地下降。此時，本實施例之無線感測網路系統1〇〇及其 錯誤感測節點偵測方法與感測方法可偵測出異常的感測節點 110,並在感測過程中排除錯誤之感測節點110所送出的資訊， 以提升無線感測網路系統100的感測準確率。其中，資訊融合 中心120包含接收單元121、偵錯單元122及判斷單元123。 接收單元121係用以接收感測節點11〇所發出測量值，偵錯單 元122係用以根據感測節點11 〇的測量值來偵測此些感測節點 參 110中是否有錯誤感測節點U1(異常的感測節點11〇)，以及在 此些感測節點1 1 〇中何者為錯誤感測節點1 1 1。在偵錯單元122 偵測出錯誤感測節點111後，判斷單元123可根據錯誤感測節 點111以外之感測節點110所傳送的測量值來進行估測。 咕參照第3圖和第4圖，第3圖繪示依照本發明之一實施 例之無線感測網路系統之感測方法的方法流程圖，第4圖繪示 依照本發明之一實施例之無線感測網路系統的系統架構圖。當 進行本實施例之無線感測網路系統1〇〇的感測方法時，首先， Q利用感測節點110來感測谓測目標，並分別接收到觀察值（步驟 210)。在本實施例中，無線感測網路系統1〇〇係使用貝氏估計 法來估測感測節點110的觀測參數(觀察值)〜θ。假設無線感 測網路系統設有8他，".，知}共Ν個感測節點110來觀測 參數0，而感測節點110會在每一個回報時間點傳送其測量值 4資訊融σ中〜120»再者，Θ的分佈函數假設已知為崩。感 測節點m觀測到的值為Α其中η為感測節點11〇的編號， ^為觀測時間的索引。每-感測節點㈣會根據觀測到的⑽ 出一個決定屺（測量值）並傳送丨眘 η . 得廷到貝訊融合中心120，而W可以 被對應於一向量訊號g & «，而此向量可用以代表 201034410 * {q1，…，qM )共M個不同的值。因此，資訊融合中心120可根據 此些感測節點110所發出的4來使用貝氏估計’而估測所觀察 的0值。 如第3圖所示，接著，利用資訊融合中心12〇的偵錯單元 122來根據感測節點11〇的測量值，而進行一錯誤感測節點偵 測（步驟220)，以偵測出錯誤感測節點u丄。 以下，進一步說明本實施例的錯誤感測節點偵測方法。請 參照第5圖，其繪示依照本發明之一實施例之無線感測網路系 0 統之錯誤感測節點偵測方法的方法流程圖。首先，假設所有的 感測節點110皆為正常的感測節點（步驟S221)。接著，對所有 感測節點110的觀察值進行同質性檢測，以檢測所有感測節點 110的測量值是否實質相同（步驟S222)<>當檢測所有感測節點 110的測量值為實質相同時，判斷所有的感測節點i 1〇皆為正 常的感測節點（步驟S223)，亦即這些感測節點i i〇中並無任何 錯誤感測節點111。當檢測出這些感測節點丨1〇的觀察值為實 質不相同時，判斷並非所有的感測節點11〇皆為正常的感測節 ❹點（亦即這些感測節點110的其中至少一者為錯誤感測節點 U1)，並假設這些感測節點110的其中至少一者為假設錯誤感 測節點（步驟S224)。接著，對假設錯誤感測節點以外之其他感 測節點110的觀察值進行同質性檢測，以檢測此假設錯誤感測 節點以外之其他感測節點11〇的測量值是否實質相同（步驟 S225)。當檢測出此假設錯誤感測節點以外之其他感測節點u 〇 的測量值為實質相同時，判斷此假設錯誤感測節點為錯誤感測 節點m(步驟S226)。當檢測出假設錯誤感測節點以外之其他 感測節點110的觀察值為實質不相同時，判斷此假設錯誤感測 筇點不是錯誤感測節點（步驟S227)，並重新假設新的至少一假 7 201034410 設錯誤感測節點，以重新 同質性檢麻把 新進仃同質性檢測。在本實施例中，此 檢利係假設無線感測網路系統⑽中大多 =為正常運作，則在無線感測網路系統⑽ ^數刹 =:具有相同性質，亦即大多數正常之感測節=的 測量值應為實質相同。 ,在本實施例中，首先，假設Ft為錯誤感測節點iu的集合， 述a式（1)之一元假设檢定問題可用來測試與求得此錯誤 感測節點的集合。 .⑴Networks) is a network system consisting of one or more wireless data collectors and a plurality of sensors. The communication method in the wireless sensing network is wireless communication. Therefore, the sensor or The wireless data collector can be conveniently placed anywhere and saves on wiring costs. However, in wireless sensing networks, it is inevitable that there are erroneous sensing nodes. The sensor may send incorrect information to the information fusion center due to energy exhaustion or hardware damage. The decrease in accuracy leads to a reduction in the overall accuracy of the estimation accuracy. SUMMARY OF THE INVENTION Therefore, an aspect of the present invention provides a method and a method for detecting an error sensing node of a wireless sensing network system, thereby improving a sensing probability of a wireless sensing network system. According to an embodiment of the present invention, the wireless sensing network system of the present invention has error 3 201034410 - a sensing node detecting method is used for detecting at least one error sensing node, wherein the wireless sensing network system includes a plurality of senses The measuring node and the information fusion center, the sensing node is configured to sense the detecting target and transmit the measured value to the information fusion center respectively, the method includes: receiving the measured value of the sensing node; and assuming at least one of the sensing nodes For at least one hypothetical error sensing node; performing homogeneity detection on observation values of other sensing nodes other than the hypothetical error sensing node to detect whether the measured values of other sensing nodes other than the hypothetical error sensing node are substantially the same; When it is detected that the measured values of other sensing nodes other than the hypothetical error sensing node are substantially the same 0, the hypothetical error sensing node is determined to be an error sensing node; and when other sensing nodes other than the hypothetical error sensing node are detected When the observed values are substantially different, it is judged that the error sensing node is not the error sensing node, and the new at least one false is assumed again. Set the error sensing node to re-test homogeneity. Moreover, in accordance with an embodiment of the present invention, the sensing method of the wireless sensing network system of the present invention is for estimating a detection target, wherein the wireless sensing network system includes a plurality of sensing nodes and an information fusion center. The method comprises: sensing a detection target by using a sensing node, and transmitting the measurement value to the information fusion command heart respectively; using the information fusion center to perform error sensing node detection according to the measured value of the sensing node, to detect Measure at least one error sensing node; and exclude the measured value of the error sensing node, and perform estimation based on the measured value transmitted by the sensing node other than the error sensing node. Moreover, in accordance with an embodiment of the present invention, the wireless sensing network system of the present invention includes an information fusion center and a plurality of sensing nodes. The sensing node is configured to respectively sense the detection target, and receive the observation value, and respectively transmit the measurement value to the information fusion center, wherein at least one of the sensing nodes is an error sensing node. The information fusion center is configured to receive the measured value of the sensing node for estimating, wherein the information fusion center includes an error clearing unit for detecting an error according to the measured value of the sensing node. 201034410 sensing node 'information fusion center system Recruitment New μ a ““Destimate the detection target based on the measured values transmitted by the sensing nodes other than the error sensing section. If the error sensing node of the wireless sensing network system of the present invention and the sensing method can detect the abnormal sensing node in advance, the sensing information sent by the abnormal sensing node is excluded. Therefore, the sensing accuracy rate of the wireless sensing network system can be improved. The above and other objects, features, advantages and embodiments of the present invention will become more apparent and understood. However, it is to be understood that the embodiments are merely illustrative of the embodiments of the invention and are not intended to limit the invention. Referring to FIG. 1 and FIG. 2, FIG. 1 is a schematic diagram of a system of a wireless sensing network system according to an embodiment of the present invention, and FIG. 2 is a diagram showing an information fusion center according to an embodiment of the present invention. Schematic. The wireless sensing network system 100 of the present embodiment is, for example, a distributed network detecting system for sensing at least one detection target (not shown), and can be applied to, for example, military or environmental monitoring. The wireless sensing network system 1 includes a information fusion center (Fusi〇n center) 丨 2 〇 and a plurality of sensing nodes 110. The sensing node 110 can be arbitrarily disposed around the detection target for sensing, and performs preliminary target event determination on the received at least one observation value, such as whether a fire event occurs. The sensing node 110 can then transmit the determined measurement values to the information fusion center 120 for final event determination. If at least one of the sensing nodes 无线1〇 of the wireless sensing network system i 发生 is abnormal due to exhaustion of energy, hardware damage or unexpected failure, the abnormal sensing node points (ie, the sense of error) The test node 1U) transmits data that does not match the actual observation status to the information fusion center 120, and these incorrect data will affect the information fusion 201034410 • Parameter estimation of the center 120, resulting in sensing of the wireless sensing network system loo The accuracy rate has dropped significantly. At this time, the wireless sensing network system 1 〇〇 and its error sensing node detecting method and sensing method of the embodiment can detect the abnormal sensing node 110 and eliminate the sense of error in the sensing process. The information sent by the node 110 is measured to improve the sensing accuracy of the wireless sensing network system 100. The information fusion center 120 includes a receiving unit 121, a debugging unit 122, and a determining unit 123. The receiving unit 121 is configured to receive the measured value sent by the sensing node 11 ,, and the detecting unit 122 is configured to detect, according to the measured value of the sensing node 11 是否, whether the sensing node 401 has an error sensing node. U1 (abnormal sensing node 11〇), and which of the sensing nodes 1 1 为 are error sensing nodes 1 1 1 . After the error detecting unit 122 detects the error sensing node 111, the determining unit 123 can perform estimation based on the measured values transmitted by the sensing node 110 other than the error sensing node 111. 3 and FIG. 4, FIG. 3 is a flowchart of a method for sensing a wireless sensing network system according to an embodiment of the present invention, and FIG. 4 is a diagram showing an embodiment of the present invention. A system architecture diagram of a wireless sensing network system. When the sensing method of the wireless sensing network system 1 of the present embodiment is performed, first, the sensing node 110 senses the target by the sensing node 110, and respectively receives the observed value (step 210). In the present embodiment, the wireless sensing network system 1 estimates the observation parameters (observations) θ of the sensing node 110 using the Bayesian estimation method. Assume that the wireless sensing network system is provided with 8, "., a total of one sensing node 110 to observe parameter 0, and the sensing node 110 will transmit its measured value 4 information sigma at each reporting time point. In the ~120» again, the distribution function of Θ is known to be collapsed. The value observed by the sensing node m is Α where η is the number of the sensing node 11〇, and ^ is the index of the observation time. Each sensor node (4) will make a decision 测量 (measurement value) according to the observed (10) and transmit 丨 η. 得廷到贝讯融合中心120, and W can be corresponding to a vector signal g & This vector can be used to represent a total of M different values for 201034410 * {q1,...,qM ). Therefore, the information fusion center 120 can estimate the observed zero value using the Bayesian estimate' based on the 4s sent by the sensing nodes 110. As shown in FIG. 3, next, the error detecting unit 122 of the information fusion center 12 is used to perform an error sensing node detection according to the measured value of the sensing node 11 (step 220) to detect an error. Sensing node u丄. Hereinafter, the error sensing node detecting method of the present embodiment will be further described. Referring to FIG. 5, a flow chart of a method for detecting an error sensing node of a wireless sensing network system according to an embodiment of the present invention is shown. First, assume that all of the sensing nodes 110 are normal sensing nodes (step S221). Next, homogeneity detection is performed on the observation values of all the sensing nodes 110 to detect whether the measured values of all the sensing nodes 110 are substantially the same (step S222) <> when the measured values of all the sensing nodes 110 are detected to be substantially the same When it is determined that all the sensing nodes i 1〇 are normal sensing nodes (step S223), that is, there are no error sensing nodes 111 in the sensing nodes ii. When it is detected that the observed values of the sensing nodes 丨1〇 are substantially different, it is determined that not all the sensing nodes 11〇 are normal sensing thrift points (that is, at least one of the sensing nodes 110) The node U1) is sensed for error, and it is assumed that at least one of the sensing nodes 110 is a hypothetical error sensing node (step S224). Next, homogeneity detection is performed on the observation values of the other sensing nodes 110 other than the error sensing node to detect whether the measured values of the other sensing nodes 11A other than the assumed error sensing node are substantially the same (step S225). When it is detected that the measured values of the other sensing nodes u 以外 other than the assumed error sensing node are substantially the same, it is judged that the assumed error sensing node is the error sensing node m (step S226). When it is detected that the observation values of the other sensing nodes 110 other than the hypothetical error sensing node are substantially different, it is judged that the hypothetical error sensing defect is not the error sensing node (step S227), and the new at least one false is newly assumed. 7 201034410 Set the error sensing node to re-homogeneous detection of the new homogenization test. In this embodiment, the detection system assumes that most of the wireless sensing network system (10) is in normal operation, and the wireless sensing network system (10) has the same property, that is, most normal feelings. The measured value of the measure = should be substantially the same. In the present embodiment, first, assuming that Ft is a set of error sensing nodes iu, a one-dimensional hypothesis verification problem of equation (1) can be used to test and obtain a set of the error sensing nodes. .(1)
Ho ' Pr[^ =qi\0] = ρ.φ for all sneS\FT; : otherwise. 其中’當H〇成立時，若馬為所選擇之錯誤感測節點集合， 則下述的統計量丹趴馬)就會收斂為一個自由度為（N—|Ft丨)(M —1)的卡方分配(Chi square distribution) ° ^(5\Fr) = f1Ke5\^}|；KLZ££ .....................................( n=l M ei ?->·〇〇Ho ' Pr[^ =qi\0] = ρ.φ for all sneS\FT; : otherwise. Where 'when H〇 is established, if the horse is the selected set of error sensing nodes, then the following statistic Dan Hummer will converge to a Chi square distribution with a degree of freedom of (N-|Ft丨) (M-1) ° ^(5\Fr) = f1Ke5\^}|;KLZ££. ....................................(n=l M ei ?->·〇〇
'X{N^FTm-\y 且〜=Σ贼=沾....... Τ Σ /=1 (3)(4) 然後，利用此統計量丑(以馬）’上述公式（1)之二元假設性檢 定問題可被處理如下。 若丑〇^\巧)>/12-«,(則〜-1)(从-1)’則接受只1’ 若’ 則接受 Η0。 其中，α為檢定之顯著水準。 此錯誤感測節點偵測方法的同質性檢測可依據如下步驟 8 201034410 - 來進行： 步驟（a):設定k=0 ’並設置檢定之顯著水準α。 步驟（b”對於所有取炉進行公式（2)的檢測。如果存在 一個候選集合馬⑻滿足Η〇,則中止同質性檢測，且最後決定的 錯誤節點集合為·為滿足下列公式（5):'X{N^FTm-\y and ~=Σ Thief=Dip....... Τ Σ /=1 (3)(4) Then, use this statistic ugly (with horse)' above formula (1 The binary hypothesis verification problem can be handled as follows. If ugly ^\巧)>/12-«, (then ~-1) (from -1)' then accept only 1' if' accept Η0. Among them, α is the significant level of the test. The homogeneity detection of the error sensing node detection method can be performed according to the following step 8 201034410 - Step (a): Set k=0 ' and set the significant level α of the verification. Step (b) performs the test of formula (2) for all the furnaces. If there is a candidate set horse (8) that satisfies Η〇, the homogeneity detection is aborted, and the finally determined set of error nodes is · for satisfying the following formula (5):
Pr=^p^^S\FTk)........................................(5) 步驟（c):若在步驟(b)中不存在任何候選集合巧w滿足H〇， 則k值增加1，亦即α = 。 Φ 步驟（d):若々 = ，則接受，且所決定的為為隨機地從 Ν個感測知點110中選出個。否則回到步驟（b)並且重覆執 行步驟步驟（b)到（d) ’直到為·被決定出來。 在偵測到錯誤感測節點111後，接著，資訊融合中心12〇 可排除錯誤感測節點111的測量值，並根據錯誤感測節點J J J 以外之其他感測節點110所傳送的測量值來進行估測（步驟 230)，因而可避免所估測的資訊中具有錯誤資訊，以確保無線 感測網路系統100的估測正確率。 φ 以下舉例來說明本實施例的錯誤感測節點偵測方法與智 線感測方法。此無線感測網路系統1〇〇例如設有5個感測節點 Si、S2、S3、s4及S5，其中84和s5為錯誤感測節點。此時， 假設實際的觀測參數Θ為2，而正常的感測節點Sl、心及心應 大多會送出測量值q4到資訊融合中心120，則錯誤感測節點s4 和S5則大多會送出q4以外的測量值至資訊融合中心丨2〇。資訊 融合中心120由此些感測節點Si、S2、S3、心及85在時間點 ΤΜ到T==5所收集到的資訊如下表一： 表一 201034410 T=1 Τ=2 Τ=3 T=4 T=5 ------ ^1 殳4 ¢4 ¢4 ¢4 Si 分4 94 ~ ---- 《4 94 94 S^ ¢4 92 ?4 94 S4 q\ qi q\ 9ι qi S5 Qi <h q\ (J2 <l\ §進仃本實施例的錯誤感測節點㈣方法時，可先假設所 ⑩有的感句知，點SrSs皆為正常的感測節點，亦即感測節點n 中並無錯誤感’點（或錯誤感測節點的數量為零），接著，對 所=感測節點呂崎的觀察值進行同質性檢測，以檢測所有感 測節點SA的測量值是否實質相同。當檢測所有感測節點 s广s5的測量值為實質相同時，則判斷所有的感測節點 音為正常的感測節點。當檢測出這些感測節點的觀察值 為實質不相同時，則判斷並非所有的感測節點SA皆為正常 的感測節點，亦即這些感測節點S广85的其中至少一者為假設 參錯誤感測節點。接著，假設錯誤感測節點的數量為一個，並例 如假設Sl為錯誤感測節點，亦即^為假設錯誤感測節點。接 著對感測節點s2、s3、污4及85進行同質性檢測，以 節點S2、S3:〜及85的測量值是否實質相同。若感測節點S2、 S3 S4及~S5通過同質性檢測，則判斷S1確實為錯誤感測節點； 若感測節點S2、s3、s4及s5未通過同質性檢測，則判斷s並 非為錯誤感測節點，並重新假設其他感測節點S2、S3U 為錯誤感測節點來進行檢測。若任-感測節點Sl、s2、S3 s5 及s5中皆無法通過同質性檢測，則假設錯誤感測節點的數量為4 201034410 二個’並例如假設\和、為錯誤感測節點，以檢測任二感測 知· 1 2 S4及S5的集合是否為錯誤感測節點集人。在 此說明例中，則可依此方式偵測到心和s5為錯誤感測^點， 然不限於此，在其他實施例中，若無線感測網路系統_具有 更多的錯誤感測節點的數量（二個以上），可同理地繼續進行檢 測’直到债測到最終的錯誤感測節點集合（多個錯誤感測 110)。 在上述說明例中’右使用現有的無線感測網路系統及估測 參方法來進行感測，則資訊融合中心會根據所有感測節點SpS2、 S3、S4及S5(包含錯誤感測節點）的資訊來進行判斷，此時所估 測出的3值為0.53，因而大幅地降低估測正確率（相較於實際觀 測參數0=2) » 反之，若使用本實施例的無線感測網路系統1〇〇及感測方 法來進行感測，則資訊融合中心12G會㈣排除（或消除）錯誤 感測節點的資訊，並根據錯誤感測節點以外的感測節點^、心 及S3的資訊來進行判斷，此時所估測出的4值為187，因而可 ^ 大幅地提升估測正確率。 請參照第6圖，其繪示依照現有感測方法與本發明之一實 施例的估測誤差比較圖。當感測節點中具有二個錯誤感測節點 時，如第3圖所示，橫軸代表現有無線感測網路系統的感測方 法與本發明之無線感測網路系統1 〇〇的感測時間，縱抽之Mse 代表分別使用現有無線感測網路系統的感測方法與本發明之 無線感測網路系統100所產生的估測誤差。由第3圖可知，使 用現有感測方法的估測誤差一直無法下降，而使用本發明之感 測方法的無線感測網路系統100可隨時間的增加而大幅地降低 估測誤差’因而可大幅地改善無線感測網路系統的感測正確 11 201034410 . 念· 牛 〇 由上述本發明的實施例可知’本發明之無線感測網路系統 的錯誤感測節點偵測方法及感測方法可預先偵測系統中是否 有錯誤感測節點，並可偵測出異常的感測節點，來排除錯誤的 感測資訊’因而大幅地提高無線感測網路系統的感測正確率。 综上所述’雖然本發明已用較佳實施例揭露如上，然其並 非用以限定本發明，本發明所屬技術領域中具有通常知識者， 在不脫離本發明之精神和範圍内，當可作各種之更動與潤飾， 〇 因此本發明之保護範圍當視後附之申請專利範圍所界定者為 準。 ’ 【圖式簡單說明】 為讓本發明之上述和其他目的、特徵、優點與實施例能更 明顯易懂，所附圖式之詳細說明如下： 第1圖繪示依照本發明之一實施例之無線感測網路系統的 系統示意圖。 φ 第2圖繪示依照本發明之一實施例之資訊融合中心的結構 示意圖。 第3圖繪示依照本發明之一實施例之無線感測網路系統之 感測方法的方法流程圖。 第4圖繪示依照本發明之一實施例之無線感測網路系統的 系統架構圖。 第5圖繪示依照本發明之一實施例之無線感測網路系統之 錯誤感測節點偵測方法的方法流程圖。 第6圖繪示依照現有感測方法與本發明之一實施例的估測 誤差比較圖。 12 201034410 【主要元件符號說明】 10〇 :無線感測網路系統 Η1 :錯誤感測節點 121 :接收單元 123 :判斷單元 21〇 :接收觀察值 220 .錯誤感測節點该測 110 :感測節點 12〇:資訊融合中心 122 :偵錯單元 2Z1 :假設所有的感測節點皆為正常的感測節點 222 :檢測所有感測節點的測量值是否實質相同 223 :判斷所有的感測節點皆為正常的感測節點 224 :假設感測節點的其中至少一者為假設錯誤感測節點 225 :檢測假設錯誤感測節點以外之感測節點的測量值是 否實質相同 226 ··判斷假設錯誤感測節點為錯誤感測節點 227 :判斷假設錯誤感測節點不是錯誤感測節點 230 :根據錯誤感測節點以外之其他感測節點的測量值來 返行估測 13Pr=^p^^S\FTk).......................................... (5) Step (c): If there is no candidate set in step (b), then the value of k increases by 1, that is, α = . Φ Step (d): If 々 = , it is accepted, and the determined one is randomly selected from the one sensing points 110. Otherwise, return to step (b) and repeat steps (b) through (d) ' until it is determined. After detecting the error sensing node 111, the information fusion center 12〇 can then exclude the measurement value of the error sensing node 111 and perform the measurement according to the measurement value transmitted by the sensing node 110 other than the error sensing node JJJ. Estimate (step 230), thereby avoiding erroneous information in the estimated information to ensure an accurate rate of estimation by the wireless sensing network system 100. φ The following describes an error sensing node detecting method and a smart line sensing method in this embodiment. The wireless sensing network system 1 is provided, for example, with five sensing nodes Si, S2, S3, s4 and S5, of which 84 and s5 are error sensing nodes. At this time, assuming that the actual observation parameter Θ is 2, and the normal sensing node S1, heart and heart should mostly send the measured value q4 to the information fusion center 120, the error sensing nodes s4 and S5 are mostly sent out of q4. The measured value is transferred to the information fusion center. The information collected by the information fusion center 120 from the sensing nodes Si, S2, S3, and 85 at time point T==5 is as follows: Table 1 201034410 T=1 Τ=2 Τ=3 T =4 T=5 ------ ^1 殳4 ¢4 ¢4 ¢4 Si 4 points 4 94 ~ ---- "4 94 94 S^ ¢4 92 ?4 94 S4 q\ qi q\ 9ι qi S5 Qi <hq\ (J2 < l\ § In the error sensing node (4) method of this embodiment, it can be assumed that all of the 10 senses are known, and the point SrSs is a normal sensing node, that is, There is no error sense point in the sensing node n (or the number of error sensing nodes is zero), and then the homogeneity detection is performed on the observation value of the sensing node Lu Qi to detect the measurement of all the sensing nodes SA Whether the values are substantially the same when detecting that the measured values of all the sensing nodes s s5 are substantially the same, then all the sensing node sounds are judged to be normal sensing nodes. When the detected values of the sensing nodes are detected, the actual values are not When the same, it is determined that not all of the sensing nodes SA are normal sensing nodes, that is, at least one of the sensing nodes S wide 85 is a hypothetical parameter error sensing node. It is assumed that the number of error sensing nodes is one, and for example, S1 is assumed to be an error sensing node, that is, a hypothetical error sensing node. Then, homogeny detection is performed on the sensing nodes s2, s3, 4 and 85, Whether the measured values of the nodes S2, S3: ~ and 85 are substantially the same. If the sensing nodes S2, S3 S4 and ~S5 pass the homogeneity detection, it is determined that S1 is indeed an error sensing node; if the sensing nodes S2, s3, If s4 and s5 fail to pass the homogeneity detection, it is judged that s is not the error sensing node, and it is assumed that the other sensing nodes S2 and S3U are the error sensing nodes for detection. If any-sensing nodes Sl, s2, S3 s5 And s5 can not pass the homogeneity detection, it is assumed that the number of error sensing nodes is 4 201034410 two 'and for example, \ and , is the error sensing node to detect any two senses · 1 2 S4 and S5 Whether the set is an error sensing node set. In this example, the heart and s5 may be detected as error detection points in this manner, but is not limited thereto. In other embodiments, if the wireless sensing network Road system_ has more error sensing nodes (two Above), the detection can continue in the same direction 'until the debt measurement reaches the final set of error sensing nodes (multiple error sensing 110). In the above illustrative example, the right wireless sensing network system and estimation are used right. According to the method for sensing, the information fusion center judges according to the information of all the sensing nodes SpS2, S3, S4 and S5 (including the error sensing node), and the estimated value of 3 is 0.53. Significantly reduce the estimation accuracy rate (compared to the actual observation parameter 0=2) » Conversely, if the wireless sensing network system 1〇〇 and the sensing method of the present embodiment are used for sensing, the information fusion center 12G (4) Exclude (or eliminate) the information of the error sensing node, and judge according to the information of the sensing node ^, heart and S3 other than the error sensing node. At this time, the estimated value of 4 is 187, and thus ^ Significantly improve the accuracy of estimates. Referring to Figure 6, there is shown a comparison error estimation error according to an existing sensing method and an embodiment of the present invention. When there are two error sensing nodes in the sensing node, as shown in FIG. 3, the horizontal axis represents the sensing method of the existing wireless sensing network system and the feeling of the wireless sensing network system 1 of the present invention. The measured time, the longitudinal pumping Mse represents the estimation error generated by the sensing method of the existing wireless sensing network system and the wireless sensing network system 100 of the present invention, respectively. As can be seen from FIG. 3, the estimation error using the existing sensing method cannot be lowered, and the wireless sensing network system 100 using the sensing method of the present invention can greatly reduce the estimation error with time. The sensing of the wireless sensing network system is improved substantially. It can detect in advance whether there is an error sensing node in the system, and can detect abnormal sensing nodes to eliminate the wrong sensing information', thus greatly improving the sensing accuracy rate of the wireless sensing network system. In the above, the present invention has been disclosed in the above preferred embodiments, and it is not intended to limit the invention, and the present invention may be made without departing from the spirit and scope of the invention. The scope of protection of the present invention is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features, advantages and embodiments of the present invention will become more <RTIgt; A schematic diagram of a system of a wireless sensing network system. φ Figure 2 is a block diagram showing the structure of an information fusion center in accordance with an embodiment of the present invention. 3 is a flow chart of a method of sensing a wireless sensing network system in accordance with an embodiment of the present invention. 4 is a system architecture diagram of a wireless sensing network system in accordance with an embodiment of the present invention. FIG. 5 is a flow chart of a method for detecting an error sensing node of a wireless sensing network system according to an embodiment of the invention. Figure 6 is a diagram showing a comparison error estimation error according to an existing sensing method and an embodiment of the present invention. 12 201034410 [Description of main component symbols] 10〇: Wireless sensing network system Η1: Error sensing node 121: Receiving unit 123: Judging unit 21〇: Receiving observation value 220. Error sensing node The measurement 110: Sensing node 12〇: Information Fusion Center 122: Debug Unit 2Z1: Assume that all the sensing nodes are normal sensing nodes 222: Detect whether the measured values of all sensing nodes are substantially the same 223: Determine that all sensing nodes are normal Sensing node 224: It is assumed that at least one of the sensing nodes is a hypothetical error sensing node 225: detecting whether the measured values of the sensing nodes other than the hypothetical error sensing node are substantially the same 226. Error sensing node 227: determining that the error sensing node is not the error sensing node 230: returning the estimate based on the measured values of the sensing nodes other than the error sensing node 13