201100307 六、發明說明： 【發明所屬之技術領域】 本發明係有關一種自動決定搬送路徑之無人搬送裝置 及其搬送路徑決定方法。 【先前技術】 在使用複數個搬送機器的無人搬送裝置中，以決定搬 送路徑的手段而言，已有各種提案（例如日本的專利文獻1 至7)。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an unmanned transport apparatus that automatically determines a transport path and a method of determining a transport path. [Prior Art] In the unmanned transport apparatus using a plurality of transport apparatuses, various proposals have been made for the means for determining the transport path (for example, Japanese Patent Literatures 1 to 7).
專利文獻1 :日本特開平6-83445號公報之「使用自 動行駛移動體之無人搬送系統中的行駛路徑選定方法」 專利文獻2 :日本特開2003-337626號公報之「路徑 決定裝置及方法」 專利文獻3 :日本特開2003-345439號公報之「自動 搬送系統及搬送車的路徑探索方法」[Patent Document 1] Japanese Laid-Open Patent Publication No. Hei. No. Hei. No. Hei. No. 6-83445. Patent Document 3: "Automatic Transfer System and Path Search Method for Transport Vehicle", JP-A-2003-345439
專利文獻4 :日本專利第3485755號公告之「無人搬 送車控制裝置及無人搬送車控制方法」 專利文獻5 :日本專利第3539838號公告之「無人搬 送車控制裝置及無人搬送車控制方法」 專利文獻6 :日本專利第3755268號公告之「無人搬 送車控制裝置及無人搬送車控制方法」 專利文獻7 :日本專利第3728864號公告之「無人搬 送車控制裝置及方法」 【發明内容】 (發明所欲解決之課題） 4 321340 201100307 上述以往的無人搬送裝置中的搬送路徑決定方法係能 • 大致區分為下述三種類型。 (1) 手動定義對應於搬送對象的搬送起始端與搬送目的端 之路徑。 此方法需要手動設定對應於搬送起始端與搬送目的端 之所有的組合。此外，以此方式所決定的路徑為固定路徑， 對於彈性地對應故障或阻塞等路徑狀態而言有其限度。 (2) 將搬送路徑視為網絡（graph)，決定權重變成最小的路 〇 徑 此方法為了再計算權重，需要監視搬送機器的詳細位 置和通過時間，且需要用以獲得各種機器的資訊之感測 器、通訊路徑、以及計算裝置。 (3) 準備複數個對應於搬送對象的搬送起始端與搬送目的 端之路徑，從該路徑中選擇最適當的路徑。 此方法為了選擇能在迴避阻塞等之最短時間内移動的 jp 路徑，係以現在路徑的混雜和未來路徑的混雜預測為依據 來決定。因此，需要精密度良好的預測，當混雜預測失敗 時，因為會以依據錯誤的預測而得之路徑來進行搬送，故 不會通過適當的路徑。 (解決課題的手段） 本發明乃為解決上述以往的問題點而研創者。亦即， 本發明的目的在提供一種無人搬送裝置及其搬送路徑決定 方法，係能自動決定搬送路徑，無須監視搬送機器的詳細 位置和通過時間，且無需用以監視的感測器、通訊路徑、 321340 201100307 以及計算裝置，並無需精密度高的預測，即使混雜預測失 敗亦能選擇適當的路徑。 本發明提供一種無人搬送裝置，係個別地控制複數個 搬送機器，並具備有：複數個機器控制裝置，係控制各搬 送裝置；搬送控制裝置，係探索路徑並對各機器控制裝置 發送指示；以及記憶裝置，係記憶搬送機器狀態、搬送路 徑網絡、以及預定路徑；前述記憶裝置係記憶將搬送機器 的轉乘地點、分歧地點、合流地點、搬送起始端、以及搬 送目的端作為「點」'將能從各點直接移動至鄰接地點者作 為「枝」、以及將各枝的負載作為「權重」之網絡；前述搬 送控制裝置係依各物品將構成從搬送起始端至搬送目的端 之搬送路徑之各枝的權重總和變成最小之路徑決定為「預 定路徑」，對所決定的預定路徑的各枝的現在權重加算預定 權重，對各物品沿著所決定的預定路徑個別地控制搬送機 器搬送物品，且從已搬送過的預定路徑的各枝的現在權重 減算掉前述預定權重。 此外，本發明提供一種無人搬送裝置的搬送路徑決定 方法，係個別地控制複數個搬送機器，藉由記憶裝置記憶 將搬送機器的轉乘地點、分歧地點、合流地點、搬送起始 端、以及搬送目的端作為「點」、將能從各點直接移動至鄰 接地點者作為「枝」、以及將各枝的負載作為「權重」之網 絡，藉由搬送控制裝置，依各物品將構成從搬送起始端至 搬送目的端之搬送路徑之各枝的權重總和變成最小之路徑 決定為「預定路徑」，對所決定的預定路徑的各枝的現在權 6 321340 201100307 重加算預定權重，對各物品沿著所決定的預定路徑個別地 • 控制搬送機器以搬送物品，且從已搬送過的預定路徑的各 枝的現在權重減算掉前述預定權重。 依據本發明的較佳實施形態，前述所決定的預定路徑 中，關於超過預定臨限值之路徑的各枝，係以預定貢獻率 ,-λ. 降低所加算之權重。 此外，各物品每次從點移至點時，係從現在的「預定 路徑」各枝的現在權重減算掉前述預定權重，接著再次決 ^ 定構成從現在的點至搬送目的端之搬送路徑之各枝的權重 總和變成最小之路徑以作為「預定路徑」，且對所決定的預 定路徑的各枝的現在權重加算前述預定權重。 接著，對某物品曾經通過的路徑的現在權重加算預定 罰值（penalty)，再次決定該物品的「預定路徑」。 (發明之效果） 依據上述本發明的裝置及方法，係記憶將搬送機器的 0 轉乘地點、分歧地點、合流地點、搬送起始端、以及搬送 目的端作為「點」、將能從各點直接移動至鄰接地點者作為 「枝」、以及將各枝的負載作為「權重」之網絡，依各物品 將構成從搬送起始端至搬送目的端之搬送路徑之各枝的權 重總和變成最小之路徑決定為「預定路徑」，對所決定的預 定路徑的各枝的現在權重加算預定權重，因此將搬送路徑 視為網絡，並將由路徑探索部所暫時決定的路徑作為將來 的阻塞狀態並應用於權重，而能預測變成阻塞的路徑並決 定迴避該阻塞路徑之路徑。 7 321340 201100307 口此mi地對應故障和阻塞等路徑的狀態，且無 需用以獲得各種機n的權重計算所t㈣訊之感測器、通 訊路徑、以及計算裝置等。 μ此外’各物品每次從點移動至點時，從現在的權重減 :掉現在的「預定路徑」各枝的權重，接著再次決定構成 攸，在的點至搬送目的端之搬送路徑之各枝的權重總和變 成最】之路径以作為「預定路徑」，並對現在的權重加算所 決定的預定路㈣各枝的㈣，因此能在各物品每次移動 至搬送路㈣分離、合流、齡料,_再讀討搬送路 徑並更新路徑。 因此’恆常地以最新的路徑資訊選擇路徑，即使未來 的混，預敎敗而選擇料#的路徑，亦能在搬送途中返 回適田的路I。因此’無須高精密度的預測。 【實施方式】 M J' 此外，各 /，、?、咐圔矹明本發明的較佳實施例。此外 圖中共通的部分係附上相同的符號並省略重複的說明。 第—1圖係本發明的無人搬送裝置的整體構成圖。 數個^第=中’本發明的無人搬送裝置10係具備有：複 示12，係控制複數個搬送裝置U，·機器相 π ’係對各機11控㈣置12發送指示，·搬送指 不接收部15’係從上位裝置14 部…係探索路經；以及記憶裝置17a、17b、i7^^d索 ==億搬送機器狀態、搬送路徑網絡、物品資訊、以 、疋么。各搬送裝置η亦可為一次搬送一個物品之裝 321340 201100307 - 置。 . 上述的機器指示發送部13、搬送指示接收部15、以及 路徑探索部16係構成搬送控制裝置18整體。該搬送控制 裝置18係例如單一或複數個電腦。 此外’上述記憶裝置17a、nb、17c、17(1亦可為單一 個記憶裝置17。 在第1圖的構成中，從上位裝置14對搬送指示接收部 ❹15賦予物扣的搬送指示（移動目的地及移動優先度等）。 路牷探索部16係將搬送機器的狀態17a(故障資訊等） 反映至搬送路彳 1：網絡17b。 路役捸索部16係評價搬送路徑網絡17b上的權重並產 生搬送路經。 機器指示發送部13係依據所產生的搬送路徑對各控 制裝置12進行搬送指示。 第2圖係顯示搬送路徑網絡的例子之圖。 ❹ 搬送路徑係由天井行驶台車⑽v:Qverhead 如cle)、無人搬送車⑽：㈣Guided硫似、以 動起重機（crane)等各種搬送機器的組合而構成 該等搬送機器的組合來表賴送純網絡仙下簡= 網絡」）。 Ί 1係由「點」2以及連轉接的點與點之「枝3 二：合而構成。枝係能具有方向性。在本例中，點2以矩 形表不’枝3則以箭頭表示。 為了以網絡表現搬送機器（在第2圖令為A至D)所構 321340 9 201100307 . 、 各機器的轉乘地點、分歧地點、合流地點、 srtr地點、以及成為搬送目的端之地點皆定 ί技」機：將能從_ 2技」I歸可如㈣㈣形，也有僅可單向移動 ⑽形’將移動方向定義為「枝的方向」。 到遠ίΓ在第2圖的例子中，考量「從出發點α搬送至 達點^的情形。此時，雖然有「㈣」與「㈠+ 針」兩條搬达路彳i，但須指示哪條路徑為「較佳路徑」之指 發明中，於各枝3定義權重4，將構成搬送路經 %枝的權重4的總合變成最小之路徑定義為「最佳路 權重4雖為任意的數值’但例如將該數值設為「路徑 $通過時間」時’各枝的權重總和係變成「從出發點移動 到達點所需的時間」。此時，所謂「最佳路 以最短時間通過的路徑」。 又成月匕 、^述搬送路徑網絡的記憶裝置i7b係記憶將搬送機器 ，轉乘地點 '分歧地點、合流地點、搬送起始端、以及搬 送目「的端作為「點」、將能從各點直接移動至鄰接地點者作 枝」以及將各枝的負載作為「權重」之網絡。 ^「第4圖係變更第3圖的權重之圖。在第4圖中，各數 予5」、「10」、及「3〇」為權重4。 321340 10 201100307 在求出任意的兩點間的「最料 的權重量下會怪常地得到相此7在固定 特精阻塞，亦無法獲得能迴避的路==徑的 應阻基，需動態性地變更權重4。 為了對 c 第3圖中，考量「從出發點Α搬送至到達點 Ο Ο 的權重間產生阻塞時，如第4圖所示增大「A〜C」門 :，重’即能獲得「“一」作為「最佳路徑。= 從係阻塞的运迴路徑。 」該路 為I以此方式動態性地變更權重4，以 二;徑阻塞之感測器，以及計測機= :間=並將移動時間實績值作為新的一 ^ 二：'然而，此方法需要設置用以求出阻塞、位置、 感測器以及對尋求路徑的裝置發送所檢測的資訊 而不取r路在月中’係以下述方法動態性地變更權重， 不取侍路徑阻塞和移動時間的實績值。 ⑴藉由記憶裝置17預先記憶將搬送機器的轉乘地點、分 ，地點、合流地點、搬送起始端、以及搬送目的端作 為點」、絲從各點錢移動轉接地點者作為 」、以及將各枝的負載作為「權重」之網絡。 () = ’藉由搬送控㈣置18’依各物品將構成從搬送 純端至搬送目的端之搬送路徑之各枝的權重總和變 成瑕小之辆㈣「預定路徑」，對所決定的預定路徑 321340 11 201100307 ^枝^現在權重加算預定權重。預定權重係例如相 δ於一合物品的權重。 徑的==路徑時作為預定路徑’對構成該路 數的增加函數約。技3的權重係設為加入該枝的預約 八之權重（預定權重）—台機11移動於路徑的該部 ^j 計值）Χ預約數。在此，所謂機器係指搬送裝置 *定ΐ該技3’所謂預約數係指 徑的構成之該枝3、的合計路徑時使用於該等預約路 新的搬送路徑係以上述權重求出。 ⑶⑽，#由搬送控制裝χ 18，依各物品沿著所決定的 預疋路#個別地控制搬送機器以搬送物品，且從已搬 =的預定路徑的各枝的現在權重減算前述預定權重。 =圖係顯示本發明的方法的第一實施形態之圖。如 ’ Μ第5圖⑷至（C)的順序變更各枝的權重。 圖⑴係最初处出發點Α搬送至到達點C」的情 元。亦即，第5圖传顧 定致〜 圖⑷知顯不決疋用以搬送最初的物品之預 路從刖的狀態。此時，從「A_盥 條搬送路徑中選摞.枯的掘壬ώ '、 B D—C」兩 m「 重總和變成最小之「A—c」（以 下％為「技ACj )。 2〇。依據此選擇，預約枝AC，將枝AC的權重從10變更至 之 第5圖⑻係第二次「從出發點a搬送至到達點c 321340 12 201100307 小的ΓΑ 此時，亦選擇各枝的權重總和變成最 小的「“C」（枝AC)。 ¾^取 30 依據此選擇，預約技AC，將枝AC的權重從2〇變更至 心第=圖（?係第三次「從出發點Α搬送至到達點C」之 ^路二1’弟5圖（c)係顯示決定用以搬送第三個物品之 〇 =Γ的狀態。此時，從「“c」與「“一」 、R搬送路徑中，選擇各枝的權重總和變成最小的「Α ϋ—c」° DC 此選擇，預約枝AB、仙、DC，分別將枝ΑΒ、仙、 DC的推重變更成10、20、20。 以減管各枝的權重4係在實際通過預約的枝時依序予 雜上述本發明的方法，能獲得下述效果。 〇 1)-般而言，由於路徑預約較少的路徑之權重較小，因 此新的搬送路徑係優先選擇路徑預約較少的路徑。由 於預約較少的路徑可視為阻塞較少的路彳f，因此可避 開阻塞。 ， (2)由於賴數亦表示未來路徑的利㈣度，因此亦成為 混雜的預測。因此，能獲得可考慮未來的混雜並予以 迴避之路徑。 在此所明「未來」係指已對複數個搬送裂置進行路 徑預約時，直至執行至少一個該預約為止之時間。 321340 13 201100307 第6圖係本發明的方法 垂 預測之圖。 第一只鈿形悲，並顯示未來 在第6圖中，A至〇為上述的點2 存在各枝3。 如第6圖所不，亦 約f 的搬送路徑’接著，預 枝CD，枝CD的權重係增加。 料中’重複預約 在此情形中，當接菩直、卡「了， 、求 Κ—Ε」的搬送路徑睥 :彳:圖的虛線所示’能獲得—雜部分二: =圖係顯示未來預測的其他參考例之圖。 在第7圖中，人至〇為上述的點2 存在各枝3。 ” 如第7圖所不，亦 百先，預約「Αϋϋ~>ί—ΐκ ρ 送路徑兩次。在此情形中 ―」的搬 重係增加。 设預路從，該路徑的權 在此情形中，當接著尋求「 第7圖的纽戶特， N/」的搬送路徑時，如 的路經。行—仔迴避未來的混雜部分「H—G」 實際的使用尚是遙遠 」的搬送路徑，則會 然而，「i^G」的路禋雖被預約 的未來。若在崎形巾料求「N— k:成迴避該路徑的路經。 如同此例 徑—;為遙來論該路 321340 14 201100307 ， 帛8圖係本發_方法㈣三實施形態， • 預測之圖。 不木 在如第7圖之情形中，亦可定義臨限值，且不預約超 過該臨限值的未來路徑。亦印，在所決定的預定路徑中， :於=預定臨限值之路徑的各技，係以預定貢獻率降低 具的推重。貢獻率係為1以下的正數，亦可為〇。 臨限值係例如該模型的路徑步階數（亦即直 ❹ =止所通過的枝的數目）、移動距離（亦即直至移動^ ^止所移動的距離）、以及移動時間（亦即直至達至^ =所需的時間)。此外，亦可不是完全地廢棄，而是階段 性地減少對於權重的貢獻率。 白奴 為了 = =^;:法中’在開始搬送時， 徑，係預先準徑中選擇適當的路 預先4備⑽條職候補，從複數條路徑 〇 ^擇，且簡絲賴送路徑，並將路徑反= 重，藉由it行阻塞迴避。 減映至核 ^由在使用發出搬送指示時的路徑狀態進 =:令，由於亦僅考慮到該時間點 、： ===。此外’雖亦從搬送機器的移夂： 行預：需要阻塞預測等來選擇路經，但為了進 準確，從而無二賴异。此外，當未來愈遙遠時預測愈不 的路簡搬示時 >、，亦針對搬送機 321340 15 201100307 :的轉乘地點/分離魅㈣送目的端，在絲得的複數 條路徑的處所中以最新的路徑狀態再次求出路徑。 亦即’依據本發明’在各物品每次從點移動至點時， =從現在的「預定路徑」的各枝的現在權重減算掉前述預 二權重’接著再次決定構成從現在的點至搬送目的端之搬 =路徑之各枝的權重總和變«小的路彳i以作為「預定路 ，」’亚對所蚊的預定路徑的各枝的現在權重加算前述預 夂權重。 、 藉由如此頻繁地進行路徑的再檢討，不僅能求出因應 迎時改變的路徑狀況之搬送路徑，即使在未來的預測失敗 而無法選擇適當的路徑時，亦會在再檢討時修正至適當 路徑。 -第9圖#、本㈣的方法的第四實施形態，且顯示搬送 路輕的再檢討之圖。 、如第9圖（A)所示’當以發出搬送指示時的路徑條件尋 求路l時，雖由於無阻塞而欲經由f_h之間，但由於兩個 物品在未來同時使用F-H間的區域，因此產生阻塞。 κ因此’_如第9圖（B)所示，藉由以屬於分離地點之£再 人h 〇寸路l，能求出不阻塞的迴避路徑。 第1〇圖係顯示進行搬送路徑的再檢討之其他參考 之圖。 設想有複數個從某地點移動至目的地之候補，該等候 補之路徑狀況相同’不管使用哪不會改變。在此， 在所選擇的路彳&上的點亦有複數個移動至目的地之候補， 321340 16 201100307 ，且該等候補之路徑狀況相同，不論使用哪一個都不會改變 • 時，有時會求出持續環繞相同地方之路徑。 例如，如第10圖（A)所示，在點D中，由於路徑（DFEA) 與路徑（DFECA)的路徑狀況相同，因此路徑（DF£CA)被作為 路徑求出。接著，當物品移動至點C時，如第10圖（B)所 示，在點C中，由於路徑（CBA)與路徑（CBDFEA)的路徑狀況 相同，因此路徑（CBDFEA)被作為路徑求出。接著，當物品 移動至點D時，返回最初的狀態，而持續環繞相同的地方。 第11圖係本發明的方法的第五實施形態，且顯示搬送 路徑的再檢討之圖。 為了解決第10圖的問題，在本發明中，對某物品曾經 通過的路徑的現在權重加算預定罰值，再次決定該物品的 「預定路徑」。所謂罰值係設為例如相當於一台物品的權 〇 亦即，如第11圖所示，對某物品曾經通過的路徑賦予 q 罰值6，而優先選擇無罰值的路徑。例如，該罰值係藉由 增加該路徑的權重而實現。 依據上述本發明的裝置及方法，係記憶將搬送機器11 的轉乘地點、分歧地點、合流地點、搬送起始端、以及搬 送目的端作為「點」2、將能從各點2直接移動至鄰接地點 者作為「枝」3、以及將各枝3的負載作為「權重」4之網 絡1，並依各物品決定將構成從搬送起始端至搬送目的端 之搬送路徑之各枝的權重總和變成最小之路徑以作為「預 定路徑」，對所決定的預定路徑的各枝的權重加算現在權 17 321340 201100307 重，因此將搬送路徑視為網絡丨，並將路徑探索部所曾經 決定的路徑作為未來的阻塞狀態而應用於權重，藉此能預 測變成阻塞的路徑，並決定迴避該阻塞之路徑。 因此，彈性地對應故障和阻塞等路徑的狀態，且無須 用以獲侍各種機器的權重計算所需的資訊之感測器、通訊 路徑、以及計算裝置等。 此外，各物品每次從點2移動至點2時，從現在的權 重減算掉現在的「預定純」各枝的權重，接著再次決定 構成從現在的點至搬送目的端之搬送路徑之各枝的權重碑 t變成最小之魅以料「預定路徑」，㈣現在的權重加 开所決疋的預定路控的各枝的權重，因此能在各物品每次 私動至搬Id路㈣分離、合流、轉乘等地點時再次檢討搬 送路徑並更新路徑。 因此 、 恆^地以最新的路徑資訊選擇路徑，即使未來 的此雜預測失敗而選擇不適#的路徑，亦能在搬送途中运 回適當的路徑。因此，無須高精密度的預測。 路a曰此外’本發明並未限定於上述實施形態，在未逸離才 發明的精神下當然能施予各種變更。 【圖式簡單說明】 =係本發明的無人搬送裝置的整體構成圖。 弟2圖係顯示搬送路經網絡的例子之圖。 第3圖係對第2圖賦予權重之圖。 第4圖係變更第3圖的權重之圖。 弟5圖（A)至⑹係顯示本發明的方法的第—實施形態 321340 18 201100307 '之圖。 第6圖係本發明的方法 預測之圖。 的第二實施形態 且顯示未來 且顯示未來 第7圖係顯示未來_叫轉考例之圖 第8圖係本發_方㈣第三實施形態， 第9圖（A)及（B)係本發明的古 知乃的方法的第四實施形態，且 顯示搬送路徑的再檢討之圖。 第10圖（A)及（B)係顯示進行搬送路徑的再檢討之其 他參考例之圖。 八 第11圖係本發明的方法的第五實施形態，且顯示搬送 路徑的再檢討之圖。 【主要元件符號說明】 1 網絡 3 枝 〇 10 無人搬送裝置 12 機器控制裝置 14 上位裝置 16 路徑探索部 18 搬送控制裝置 2 點 4 權重 11 搬送裝置 13 機器指示發送部 15 搬送指示接收部 17a 至 17d 記憶裝置 39 321340Patent Document 4: "Unmanned Transport Vehicle Control Device and Unmanned Transport Vehicle Control Method", Japanese Patent No. 3485755, Patent Document 5: "Unmanned Transport Vehicle Control Device and Unmanned Transport Vehicle Control Method", Japanese Patent No. 3539838 6: "Unmanned transport vehicle control device and unmanned transport vehicle control method" of Japanese Patent No. 3755268. Patent Document 7: "Unmanned Transport Vehicle Control Device and Method" of Japanese Patent No. 3728864 [Invention] Problem to be solved) 4 321340 201100307 The above-described conventional unmanned transport device determination method can be roughly classified into the following three types. (1) The path corresponding to the transport start end and the transport destination end of the transport target is manually defined. This method requires manual setting of all combinations corresponding to the transfer start end and the transfer destination end. In addition, the path determined in this way is a fixed path, and there is a limit to elastically corresponding to a path state such as a fault or a blockage. (2) The transfer path is regarded as a graph, and the weight is determined to be the smallest path. In order to recalculate the weight, it is necessary to monitor the detailed position and transit time of the transport machine, and it is necessary to obtain information on various machines. Tester, communication path, and computing device. (3) A plurality of paths corresponding to the transfer start end and the transfer destination end of the transfer destination are prepared, and the most appropriate path is selected from the paths. In order to select the jp path that can be moved in the shortest time to avoid blocking, etc., this method is determined based on the hybrid prediction of the current path and the hybrid prediction of the future path. Therefore, a prediction with good precision is required, and when the miscellaneous prediction fails, since the route is obtained by the path obtained based on the erroneous prediction, the appropriate path is not passed. (Means for Solving the Problem) The present invention has been made in order to solve the above-mentioned conventional problems. That is, an object of the present invention is to provide an unmanned transport apparatus and a transport path determining method thereof, which can automatically determine a transport path without monitoring the detailed position and transit time of the transport apparatus, and without requiring a sensor or communication path for monitoring. , 321340 201100307 and computing devices, without the need for high-precision predictions, even if the hybrid prediction fails, the appropriate path can be selected. The present invention provides an unmanned transport device that individually controls a plurality of transport devices, and includes a plurality of device control devices that control each of the transport devices, and a transport control device that searches for a route and transmits an instruction to each device control device; The memory device is a memory transfer device state, a transport path network, and a predetermined route; the memory device memorizes the transfer point, the divergence point, the merge point, the transfer start end, and the transfer destination end of the transport device as "points" It is possible to move directly from each point to a neighboring place as a "branch" and a network in which the load of each branch is "weight". The transport control device configures a transport path from the transfer start end to the transfer destination end depending on each item. The route whose total weight of each branch becomes the smallest is determined as the "predetermined route", and the current weight of each branch of the determined predetermined route is added with a predetermined weight, and the articles are individually controlled to transport the articles along the determined predetermined route. And subtract the current weight of each branch of the predetermined path that has been transferred Said predetermined weight. Further, the present invention provides a method for determining a transport path of an unmanned transport device, which controls a plurality of transport devices individually, and memorizes a transfer point, a divergence point, a merge point, a transfer start end, and a transfer destination of the transport device by the memory device. The terminal is a "point", and a network that can move directly from each point to an adjacent location as a "branch" and a load of each branch as a "weight" is configured by the transport control device. The route to which the sum of the weights of the branches of the transfer destination is the smallest is determined as the "predetermined route", and the current weight of each branch of the determined predetermined route is added to the predetermined weight, and the predetermined weight is added to each item. The predetermined route to be determined individually • controls the transport device to transport the article, and subtracts the predetermined weight from the current weight of each branch of the predetermined route that has been transported. According to a preferred embodiment of the present invention, in the predetermined path determined above, for each branch of the path exceeding the predetermined threshold, the added weight is decreased by a predetermined contribution rate, -λ. In addition, each time the item is moved from the point to the point, the predetermined weight is subtracted from the current weight of each branch of the current "predetermined path", and then the transfer path from the current point to the destination of the transfer is determined again. The sum of the weights of the branches becomes the smallest path as the "predetermined path", and the aforementioned predetermined weight is added to the current weight of each branch of the determined predetermined path. Next, a predetermined penalty is added to the current weight of the path through which an item has passed, and the "predetermined path" of the item is determined again. (Effect of the Invention) According to the apparatus and method of the present invention described above, the 0-transfer point, the divergence point, the confluence point, the transport start end, and the transport destination end of the transport apparatus are stored as "points", and the points can be directly The network that moves to the adjacent location as the "branch" and the load of each branch as the "weight" determines the path that minimizes the sum of the weights of the branches from the transfer start end to the transfer destination end. In the "predetermined path", the current weight of each branch of the determined predetermined path is added with a predetermined weight. Therefore, the transport path is regarded as a network, and the path temporarily determined by the route search unit is used as a future blocked state and applied to the weight. It is possible to predict the path that becomes blocked and decide to avoid the path of the blocked path. 7 321340 201100307 This is the state of the path such as fault and blockage, and it is not necessary to obtain the weight of various machines n to calculate the sensor, communication path, and computing device of t(4). In addition, each time each article moves from the point to the point, it subtracts from the current weight: the weight of each branch of the current "predetermined path", and then determines the constituents, and the transfer path from the point of the transfer to the destination of the transfer destination. The sum of the weights of the branches becomes the most-preferred path as the "predetermined path", and the current path (4) of each branch determined by the current weight is added, so that each item can be moved to the transport path (4), separated, merged, and aged. Material, _ read the transfer path and update the path. Therefore, the path is always selected with the latest route information. Even if the future is mixed, the route of the material ## can be returned to the road I in the way. Therefore, there is no need for high-precision predictions. [Embodiment] M J' In addition, each of /, ??, and the preferred embodiment of the present invention. In addition, the same portions in the drawings are denoted by the same reference numerals and the repeated description is omitted. Fig. 1 is a view showing the overall configuration of the unmanned conveying apparatus of the present invention. In the unmanned transport apparatus 10 of the present invention, the unmanned transport apparatus 10 includes a repeating display 12 for controlling a plurality of transporting devices U, and a machine phase π' for each of the machines 11 to control (four) to set 12 transmission instructions, and The non-receiving unit 15' searches for a path from the host device 14; and the memory devices 17a, 17b, i7^^d==100 million transfer machine status, transport path network, item information, and so on. Each of the transporting devices η can also be used to transport one item at a time 321340 201100307 -. The above-described device instruction transmitting unit 13, the transport instruction receiving unit 15, and the route search unit 16 constitute the entire transport control device 18. The transport control device 18 is, for example, a single or a plurality of computers. In addition, the above-mentioned memory devices 17a, nb, 17c, and 17 (1 may be a single memory device 17. In the configuration of Fig. 1, the transport instruction is given to the transport instruction receiving unit 15 from the host device 14 (moving purpose) The route search unit 16 reflects the state 17a (failure information, etc.) of the transport device to the transport path 1: the network 17b. The road support unit 16 evaluates the weight on the transport path network 17b. The machine instruction transmitting unit 13 transmits an instruction to each of the control devices 12 in accordance with the generated transport path. Fig. 2 is a view showing an example of the transport path network. 搬 The transport path is driven by the patio (10) v: Qverhead (clear), unmanned transport vehicle (10): (4) Guided sulfur-like, a combination of various transporting machines such as cranes, which constitutes a combination of these transporting machines, and is based on the network. Ί 1 is composed of "point" 2 and the point of the transition and the "twig 3: combination of the point. The branch can have directionality. In this example, point 2 is represented by a rectangle. In order to transfer the machine by network (A to D in the second picture), 321340 9 201100307 . , the transfer point of each machine, the place of divergence, the location of the confluence, the location of the srtr, and the location to be the destination of the transfer. The machine can be changed from _ 2 technology to I (4) (four), or only one-way (10) shape to define the direction of movement as "direction of the branch". In the example of the second picture, consider the case of "moving from the starting point α to the point ^. At this time, although there are two ("four)" and "(one) + needle", it is necessary to indicate which In the invention in which the strip path is the "better path", the weight 4 is defined in each branch 3, and the path in which the sum of the weights 4 constituting the transport path via the % branch is minimized is defined as "the optimal road weight 4 is arbitrary. The value 'However, for example, when the value is set to "path $ passage time", the sum of the weights of the branches becomes "the time required to move from the departure point to the arrival point". At this time, the "the best way to pass the shortest time". In addition, the memory device i7b of the transport route network is a memory transfer device, and the transfer point "difference point, merge point, transfer start end, and transfer destination" as "points" will be able to pass from each point. The system moves directly to the neighboring location and the load of each branch as the "weight". ^ "Graph 4 is a diagram for changing the weight of Figure 3. In Figure 4, each number is 5", "10", and "3" is a weight of 4. 321340 10 201100307 In the calculation of the arbitrary weight between the two points, it will be strangely obtained that the 7 is fixed in the fixed fine, and the path that can be avoided can not be obtained. Change the weight 4 in order to change the "A~C" door as shown in Fig. 4 in the figure 3, considering the "transfer from the starting point to the point of arrival". That is, you can get "1" as the "best path. = the path of the blockage from the system." This path is I in this way to dynamically change the weight 4, to 2; the sensor for the path blocking, and the measuring machine = : between = and the moving time performance value is taken as a new one: 'However, this method needs to be set to find the blocking, position, sensor and send the detected information to the device seeking the path without taking r The road in the middle of the month is dynamically changed in weight by the following method, and does not take the path value of the traffic path blocking and moving time. (1) The memory device 17 pre-memorizes the transfer point, the point, the place, the merge point, the transfer start end, and the transfer destination end as the point ", the wire transfer destination from each point", and The load of each branch acts as a network of "weights". () = 'With the transfer control (four) set 18', the total weight of each branch that constitutes the transfer path from the transfer end to the transfer destination is changed to a small (four) "predetermined route" for the determined reservation. Path 321340 11 201100307 ^The branch now weights the predetermined weight. The predetermined weight is, for example, the weight of the singular item. The == path of the path acts as a predetermined path to the increasing function constituting the number of paths. The weight of the technique 3 is set to the weight of the reservation eight (the predetermined weight) to which the branch is added - the unit 11 moves the part of the path ^j count) Χ the number of reservations. Here, the device refers to the transport device. When the total route of the branch 3 in the configuration of the number of reservations is defined, the new transport path used for the reserved route is obtained by the weight. (3) (10), # By the transport control device 18, the transport device is individually controlled to carry the articles along the determined pre-route #, and the predetermined weight is subtracted from the current weight of each branch of the predetermined route that has been moved. = Figure shows a diagram of a first embodiment of the method of the invention. The weights of the branches are changed in the order of '5' (4) to (C). Figure (1) is the sentiment of the initial departure point to the arrival point C". That is to say, the fifth picture is given to the stipulations. (4) The state of the shackles used to transport the original items is unclear. At this time, from the "A_ 搬 搬 路径 摞 枯 枯 枯 、 、 、 、 、 、 、 、 、 、 、 、 、 BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD ( ( ( ( ( ( ( According to this selection, the branch AC is reserved, and the weight of the branch AC is changed from 10 to 5 (8) for the second time. "From the departure point a to the arrival point c 321340 12 201100307 Small ΓΑ At this time, the branches are also selected. The sum of the weights becomes the smallest "C" (branch AC). 3⁄4^取30 According to this selection, the reservation skill AC changes the weight of the branch AC from 2〇 to the heart = map (? is the third time "transfer from the starting point to the arrival point C" ^ 2 2 1 brother 5 Figure (c) shows the state of 〇 = 决定 which is used to transfer the third item. At this time, from the "c" and "one" and R transport paths, the total weight of each branch is selected to be the smallest. Α ϋ—c”° DC Select this option, and make a reservation for branches AB, 仙, and DC, and change the weight of the branches, 仙, and DC to 10, 20, and 20, respectively, to reduce the weight of each branch. In the case of the above-described method of the present invention, the following effects can be obtained. 〇1) In general, since the path with fewer path reservations has a smaller weight, the new transfer path is preferred to have fewer path reservations. path of. Since the path with fewer reservations can be regarded as a less obstructed path f, the blocking can be avoided. (2) Since the Lai number also indicates the profit (four) degree of the future path, it also becomes a mixed forecast. Therefore, it is possible to obtain a path that can be considered and avoided in the future. As used herein, "future" means the time until a reservation has been made for a plurality of transport splits until at least one of the appointments is executed. 321340 13 201100307 Figure 6 is a diagram of the method of the present invention. The first one is sorrowful and shows the future. In Figure 6, A to 〇 is the above point 2 and there are 3 branches. As shown in Fig. 6, the transfer path of the f is also followed by the pre-branch CD and the weight of the branch CD is increased. In the case of 'repeated appointments, in this case, when the connection is straight, the card is ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, A diagram of other reference examples for prediction. In Fig. 7, there is a branch 3 in the point 2 where the person is at the above point. As shown in Figure 7, it is also a hundred-first, and the reservation "Αϋϋ~> ί-ΐκ ρ sends the path twice. In this case ―" the weight system increases. Pre-route, the right of the route In this case, when you follow the route of "Newton, N/" in Figure 7, the path is as follows. In the future, the "H-G", which is a distant part of the future, is still a distant route. However, the road of "i^G" is reserved for the future. If you ask for "N-k in the shape of the smear-shaped towel, it is a way to avoid the path. Like this case--" is the road to 321340 14 201100307, 帛8 is the implementation of the method _ method (four) three implementations, The map of the forecast. In the case of Figure 7, the threshold can also be defined, and the future path beyond the threshold is not reserved. Also printed, in the determined predetermined path, : == predetermined Each of the techniques of the limit value reduces the weight of the tool by a predetermined contribution rate. The contribution rate is a positive number of 1 or less, and may be 〇. The threshold value is, for example, the path step number of the model (ie, straight ❹ = The number of branches that pass through, the distance traveled (ie, the distance moved until the movement), and the time of movement (ie, the time until it reaches ^ =). In addition, it may not be completely discarded. Instead, the contribution rate to the weight is reduced in stages. The white slaves are in the ==^;: in the law, when starting the transfer, the path is selected in advance, and the appropriate route is selected in advance (4). The path is selected, and the wire is sent to the path, and the path is reversed = heavy, with the line blocking回 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 : Need to block predictions, etc. to choose the path, but in order to be accurate, it is no different. In addition, when the future is farther away, the more predictive the road is when the road is moved, &, also, for the conveyor 321340 15 201100307 : The transfer point/separation charm (4) is sent to the destination end, and the path is again obtained in the position of the plurality of paths of the silk with the latest path state. That is, according to the present invention, each time the item moves from the point to the point, = Calculate the pre-two weights from the current weight of each branch of the current "scheduled path" and then decide again that the sum of the weights of the branches from the current point to the destination of the transport destination is changed. The aforementioned weight is added to the current weight of each branch of the predetermined path of the "predetermined road," 'A pair of mosquitoes.' By re-examining the route so frequently, it is possible to find not only the route of the route that changes in response to the change in time, but also to correct the appropriate route when the future forecast fails and the appropriate route cannot be selected. path. - The fourth embodiment of the method of Fig. 9 and (4), and showing a re-examination of the light conveyance path. As shown in Fig. 9(A), when the route 1 is sought by the path condition when the transport instruction is issued, although it is not blocked, it is intended to pass between f_h, but since the two items use the area between the FHs in the future, Therefore, a blockage occurs. κ Therefore, as shown in Fig. 9(B), the avoidance path of the non-blocking can be obtained by taking the position of the separation point. The first diagram shows a different reference to the re-examination of the transport path. Imagine that there are multiple candidates for moving from a location to a destination, and the path of the waiting route is the same 'will not change regardless of the usage. Here, there are a plurality of points on the selected path & the candidate to move to the destination, 321340 16 201100307, and the path of the waiting route is the same, no matter which one is used, it will not change. It will find the path that continues around the same place. For example, as shown in Fig. 10(A), in the point D, since the path condition of the path (DFEA) and the path (DFECA) is the same, the path (DF£CA) is obtained as the path. Next, when the article moves to the point C, as shown in FIG. 10(B), in the point C, since the path condition of the path (CBA) and the path (CBDFEA) is the same, the path (CBDFEA) is obtained as the path. . Then, when the item moves to point D, it returns to the original state and continues to surround the same place. Fig. 11 is a view showing a fifth embodiment of the method of the present invention, and showing a re-examination of the transport path. In order to solve the problem of Fig. 10, in the present invention, a predetermined penalty value is added to the current weight of the path through which an article has passed, and the "predetermined path" of the article is again determined. The penalty value is set to, for example, the weight of one item, that is, as shown in Fig. 11, the path to which an item has been passed is given a penalty of 6 and the path without penalty is preferentially selected. For example, the penalty is achieved by increasing the weight of the path. According to the apparatus and method of the present invention described above, the transfer point, the divergence point, the merge point, the transfer start end, and the transfer destination end of the transport device 11 are stored as "points" 2, and can be directly moved from the respective points 2 to the adjacent The location is the "branch" 3, and the load of each branch 3 is used as the network 1 of the "weight" 4, and the sum of the weights of the branches constituting the transport path from the transfer start end to the transfer destination end is minimized according to each item. The path is used as the "predetermined path", and the weight of each branch of the determined predetermined path is added to the current weight of 17 321340 201100307. Therefore, the transport path is regarded as the network port, and the path determined by the path search unit is taken as the future. The blocking state is applied to the weight, whereby the path that becomes blocked is predicted, and the path to avoid the blocking is decided. Therefore, it is possible to elastically correspond to the state of the path such as the failure and the blockage, and it is not necessary to use the sensor, the communication path, the calculation means, and the like for obtaining the information required for the weight calculation of various machines. In addition, each time each item moves from point 2 to point 2, the weight of the current "predetermined pure" branches is subtracted from the current weight, and then the branches constituting the transfer path from the current point to the destination of the transfer are again determined. The weight of the monument t becomes the smallest charm to "scheduled path", (4) the current weight is added to the weight of each branch of the predetermined road control, so that each item can be moved to the Id Road (4) every time. When the location is merged or transferred, the transfer path is reviewed again and the route is updated. Therefore, the path is selected with the latest path information, and even if the future prediction fails, the path of the unsuitable # can be returned to the appropriate path during the transfer. Therefore, there is no need for high-precision predictions. Further, the present invention is not limited to the above-described embodiments, and various modifications can of course be made without departing from the spirit of the invention. BRIEF DESCRIPTION OF THE DRAWINGS = The overall configuration of the unmanned transport apparatus of the present invention. The Brother 2 diagram shows a diagram of an example of a transport path through a network. Figure 3 is a diagram showing the weight given to Figure 2. Fig. 4 is a diagram showing the change of the weight of Fig. 3. Fig. 5 (A) to (6) show the first embodiment of the method of the present invention 321340 18 201100307'. Figure 6 is a graph of the method of the present invention. The second embodiment shows the future and shows that the seventh picture shows the future. The figure 8 shows the third embodiment of the present invention. The fifth embodiment (A) and (B) According to a fourth embodiment of the method of the invention, a re-examination of the transport path is shown. Fig. 10 (A) and (B) are diagrams showing other reference examples for reviewing the transport path. Eighth Figure 11 is a fifth embodiment of the method of the present invention, and shows a re-examination of the transport path. [Description of main component symbols] 1 Network 3 Branch 10 Unmanned transport device 12 Machine control device 14 Host device 16 Path search unit 18 Transport control device 2 Point 4 Weight 11 Transport device 13 Machine instruction transmitting unit 15 Transport instruction receiving units 17a to 17d Memory device 39 321340