外文翻譯--在機(jī)器人足球賽中多智能體系統(tǒng)的發(fā)展

1、<p>　　中文5200字，英文2900單詞</p><p>　　2013屆本科畢業(yè)設(shè)計(jì)(論文)外文</p><p><b>　　文獻(xiàn)翻譯</b></p><p>　　學(xué) 院 電氣與自動(dòng)化工程學(xué)院 </p><p>　　專(zhuān) 業(yè)： Y自動(dòng)化

2、 </p><p>　　姓 名： </p><p>　　學(xué) 號(hào)： </p><p>　　外文出處：Robotics and Automation，1997. </p><p>　　Proceedings.，1997 I

3、nternational Conference Engineering on. IEEE，1997，1：626-631 </p><p>　　附 件：1.外文資料翻譯譯文；2.外文原文</p><p>　　附件1：外文資料翻譯論文</p><p>　　在機(jī)器人足球賽中多智能體系統(tǒng)的發(fā)展</p><p><

4、b>　　一、介紹</b></p><p>　　隨著現(xiàn)代工業(yè)的發(fā)展，應(yīng)用機(jī)器人的需要在增加。尤其是移動(dòng)機(jī)器人在工業(yè)和研究領(lǐng)域中逐漸地?cái)U(kuò)大了其應(yīng)用。對(duì)移動(dòng)機(jī)器人的研究主要集中在單一移動(dòng)機(jī)器人。但是，隨著機(jī)器人執(zhí)行復(fù)雜任務(wù)的增加，多智能體系統(tǒng)的發(fā)展是必須的。</p><p>　　許多學(xué)者研究多智能體系統(tǒng)。一般來(lái)說(shuō)，多智能體系統(tǒng)定義為超過(guò)2個(gè)的機(jī)器人，通過(guò)協(xié)作完成給定任務(wù)的系統(tǒng)。

5、整個(gè)系統(tǒng)相對(duì)于單一機(jī)器人系統(tǒng)有許多不同特點(diǎn)。首先，機(jī)器人所在的環(huán)境是動(dòng)態(tài)的。在多智能體系統(tǒng)中，機(jī)器人本身構(gòu)成了動(dòng)態(tài)環(huán)境，因?yàn)槊恳粋€(gè)機(jī)器人應(yīng)該識(shí)別出其余的機(jī)器人為移動(dòng)障礙物。早些的對(duì)于移動(dòng)機(jī)器人的研究都是假定環(huán)境是靜態(tài)的，甚至是單一移動(dòng)機(jī)器人的研究。第二，由于是通過(guò)協(xié)作完成既定任務(wù)的，很有必要對(duì)于機(jī)器人的角色做整個(gè)系統(tǒng)規(guī)劃。多智能體系統(tǒng)的一個(gè)明顯特征是協(xié)作，舉例來(lái)說(shuō)，2個(gè)機(jī)器人搬東西。為了相互協(xié)作，機(jī)器人位置的改變必須要預(yù)測(cè)。有一些方法可

6、以知道機(jī)器人的位置。舉個(gè)例子，機(jī)器人相互通訊自己的位置，或者管理員探測(cè)到機(jī)器人的位置并傳遞給每個(gè)機(jī)器人。這是一個(gè)與通訊有關(guān)的問(wèn)題。第三，系統(tǒng)的每一部分，例如機(jī)器人、管理員、傳感器、通訊設(shè)備必須正常，因?yàn)樗麄儗?huì)影響整個(gè)系統(tǒng)的運(yùn)行，并且必須采用合適的結(jié)構(gòu)。</p><p>　　機(jī)器人足球賽對(duì)于研究多智能體系統(tǒng)是一個(gè)讓人感興趣的領(lǐng)域。足球機(jī)器人必須一起工作，這就是一種協(xié)作。而且，對(duì)于本方的機(jī)器人或者是對(duì)手機(jī)器人，他們

7、都是在可以預(yù)測(cè)的和不可預(yù)測(cè)的動(dòng)態(tài)環(huán)境中進(jìn)行比賽。主要的目標(biāo)是在對(duì)方機(jī)器人參與的情況下。盡可能多的把足球踢到對(duì)方的球門(mén)中，當(dāng)然對(duì)方的機(jī)器人也有相同的目的。因此，基于態(tài)勢(shì)，本方機(jī)器人決定將要采取的行動(dòng)進(jìn)攻或防守，如何協(xié)作等等。在這點(diǎn)上，系統(tǒng)需要實(shí)時(shí)感知，快速?zèng)Q策和采取行動(dòng)。這些與系統(tǒng)結(jié)構(gòu)和算法有關(guān)。如下所述，機(jī)器人足球賽包含有許多多智能體系統(tǒng)的特征，很適合（人工智能）的應(yīng)用。</p><p>　　機(jī)器人足球賽的一個(gè)優(yōu)

8、點(diǎn)是對(duì)于不同系統(tǒng)的直接比較。許多足球機(jī)器人系統(tǒng)參與了一些競(jìng)賽。我們參加了在朝鮮大田舉行的MIROSOT96（機(jī)器人微型組足球賽）。機(jī)器人微型組足球賽有許多規(guī)則。規(guī)則詳細(xì)描述了足球賽的規(guī)范。場(chǎng)地是正方形的，長(zhǎng)130厘米，寬90厘米。足球是橙色的高爾夫球。機(jī)器人的尺寸限制在7.5*7.5*7.5厘米內(nèi)。每個(gè)隊(duì)包含三個(gè)機(jī)器人。我們的目的就是制作有三個(gè)機(jī)器人的足球機(jī)器人系統(tǒng)。</p><p>　　本文中，我們闡述了在建立

9、整個(gè)系統(tǒng)需要考慮的一些因素。首先，因?yàn)橄到y(tǒng)結(jié)構(gòu)很重要，我們?cè)诜治龆嘀悄荏w系統(tǒng)的基礎(chǔ)上采用了在線的集中式系統(tǒng)。第二，整個(gè)系統(tǒng)可以分成三部分機(jī)器人，通訊和視覺(jué)設(shè)備。我們描述了每一部分的說(shuō)明和采用他們的理由。這將會(huì)對(duì)以后的更進(jìn)一步提高有所幫助。</p><p>　　文章的以下內(nèi)容是這樣安排的。第二部分給出了一些系統(tǒng)結(jié)構(gòu)的研究和我們選擇的足球機(jī)器人系統(tǒng)結(jié)構(gòu)。第三部分給出了執(zhí)行硬件的具體說(shuō)明，尤其是微型機(jī)器人。第四部分介紹

10、了機(jī)器人足球賽的協(xié)作和路徑規(guī)劃算法。第五部分是結(jié)論和將來(lái)的研究工作。</p><p>　　二、多智能體系統(tǒng)的分類(lèi)和足球機(jī)器人系統(tǒng)的選擇</p><p>　　這個(gè)主題的研究是與Arai的工作相關(guān)的。我們基于兩個(gè)標(biāo)準(zhǔn)對(duì)多智能系統(tǒng)進(jìn)行分類(lèi)。一個(gè)是從誰(shuí)決策和命令考慮分為集中式分散式，另一個(gè)從什么時(shí)候系統(tǒng)制定計(jì)劃考慮分為在線離線。</p><p>　　在集中式系統(tǒng)中，管理員把

11、所有的有用數(shù)據(jù)合成，計(jì)劃所有機(jī)器人的行為并作出命令。由于管理員同時(shí)考慮了所有的機(jī)器人，系統(tǒng)將會(huì)優(yōu)化所有機(jī)器人的行動(dòng)。但是隨著機(jī)器人數(shù)目的增加，管理員需要更多的計(jì)算。如果管理員發(fā)生錯(cuò)誤，機(jī)器人將沒(méi)能力進(jìn)行更正。</p><p>　　在分布式系統(tǒng)中，每個(gè)機(jī)器人從自身的傳感器和別的機(jī)器人獲得信息，進(jìn)行規(guī)劃。在系統(tǒng)中，當(dāng)機(jī)器人數(shù)目增加時(shí)，沒(méi)有計(jì)算負(fù)載的大幅度增加。即使一個(gè)機(jī)器人出故障，別的機(jī)器人還可以工作的很好。但是系統(tǒng)

12、不能保證所有機(jī)器人行動(dòng)的優(yōu)化。</p><p>　　離線系統(tǒng)意味著所有的計(jì)劃在機(jī)器人工作之前就已經(jīng)制定好。因?yàn)闆](méi)有時(shí)間和計(jì)算力的限制，系統(tǒng)整體是優(yōu)化的。但是由于系統(tǒng)是假定的靜態(tài)環(huán)境，在環(huán)境發(fā)生微小變化時(shí)，系統(tǒng)是非魯棒的。在實(shí)際環(huán)境中，系統(tǒng)會(huì)隨著一些變化發(fā)生故障。</p><p>　　在線系統(tǒng)意味著實(shí)時(shí)規(guī)劃。它對(duì)于動(dòng)態(tài)環(huán)境是魯棒的。但它需要強(qiáng)大的計(jì)算能力和有效的算法。</p>

13、<p>　　在大量的研究中，以上的兩種分類(lèi)是相互聯(lián)系的。我們?cè)诒碇羞M(jìn)行了總結(jié)。</p><p><b>　　表1多智能體的分類(lèi)</b></p><p>　　在MIROSOT96中，對(duì)機(jī)器人的尺寸進(jìn)行了限制。因此，使一個(gè)機(jī)器人具有強(qiáng)大計(jì)算力是困難的。同時(shí)，決定一個(gè)機(jī)器人必須的裝備是重要的?；旧希粋€(gè)機(jī)器人必須配備激勵(lì)和通訊模塊。在考慮了機(jī)器人空間大小的情況下

14、，可以選擇和應(yīng)用額外的設(shè)備。足球賽需要機(jī)器人位置的全局信息。因此，我們采用集中式系統(tǒng)。從路徑規(guī)劃時(shí)間來(lái)看，我們采用了在線系統(tǒng)。足球賽有快速變化的特點(diǎn)，需要實(shí)時(shí)感知，要求迅速的行為和決策。采用在線系統(tǒng)是合理的。</p><p>　　在集中式在線系統(tǒng)中，管理員獲得整個(gè)環(huán)境和機(jī)器人的所有有用信息。同時(shí)，管理員應(yīng)該實(shí)時(shí)規(guī)劃所有機(jī)器人的路徑。這要求快速的計(jì)算能力。為了減小管理員的負(fù)載，我們選擇了模塊系統(tǒng)，它把主要的規(guī)劃和執(zhí)

15、行進(jìn)行了分離。這在某些方面與Shakey[15]和Firby[16]的工作相似。在我們的足球機(jī)器人系統(tǒng)中，在某個(gè)策略中，管理員制定整個(gè)機(jī)器人的規(guī)劃。接著，管理員把下一步理想的位置信息傳遞給每個(gè)機(jī)器人。每個(gè)機(jī)器人獲得理想的位置信息并執(zhí)行控制算法來(lái)進(jìn)行位置和速度控制。同時(shí)，每個(gè)機(jī)器人把自身獲得的數(shù)據(jù)和理想位置數(shù)據(jù)融合成它自身控制環(huán)的理想輸入。這樣的話，一個(gè)機(jī)器人必須擁有某種邏輯：具有“大腦”功能。因此，我們的機(jī)器人配置有微處理器。在這種結(jié)構(gòu)

16、中，系統(tǒng)分離出計(jì)算負(fù)載。圖1表示我們足球機(jī)器人系統(tǒng)的功能圖。由于管理員得到球和機(jī)器人的位置信息，系統(tǒng)將不需要雙向通訊。管理員僅僅需要把命令傳遞到每個(gè)機(jī)器人。在雙向通訊的情況下，需要傳送和接受命令的邏輯及其優(yōu)先權(quán)。它可能會(huì)增加在管理員和機(jī)器人之間的通訊系統(tǒng)的復(fù)雜性。</p><p>　　圖1足球機(jī)器人系統(tǒng)的組成</p><p><b>　　三、系統(tǒng)運(yùn)行</b></

17、p><p>　　系統(tǒng)由三部分組成：管理員，視覺(jué)系統(tǒng)和5個(gè)機(jī)器人。管理員是一臺(tái)奔騰處理器的PC機(jī)，它將進(jìn)行實(shí)時(shí)規(guī)劃，視覺(jué)系統(tǒng)有兩個(gè)攝像機(jī)和一個(gè)具有DSP和內(nèi)存的圖像處理板，攝像機(jī)分別有紅色和藍(lán)色的過(guò)濾器。一個(gè)機(jī)器人有一個(gè)CPU，通訊模塊，IR傳感器，電機(jī)等。這三部分是相互聯(lián)系的。下面的章節(jié)將進(jìn)行詳細(xì)描述。</p><p>　　1．單個(gè)機(jī)器人的配置</p><p>　　一個(gè)

18、機(jī)器人由機(jī)械裝置，CPU主板，通訊（接收）模塊和傳感器主板組成。它的尺寸在7.5*7.5*7.5厘米內(nèi)。</p><p>　　機(jī)器人的機(jī)械裝置有兩個(gè)電機(jī)、編碼器、齒輪傳動(dòng)裝置、輪子、撥球裝置和一個(gè)機(jī)架。機(jī)架的設(shè)計(jì)必須整體緊密結(jié)合。在選擇電機(jī)和齒輪傳動(dòng)裝置的時(shí)候，需要考慮操作電壓，內(nèi)部耐力和機(jī)械時(shí)間常數(shù)。它的操作電壓是6v。傳動(dòng)裝置的變形比為1：41。輪子的直徑是32mm。電機(jī)在空載下的速度為15200r/min。

19、可以計(jì)算出機(jī)器人的空載速度大約是62cm/sec。在實(shí)際中，我們可測(cè)量出機(jī)器人的最大速度是40cm/sec。兩臺(tái)電機(jī)是分別由主板上的CPU控制的。編碼器每一轉(zhuǎn)產(chǎn)生16個(gè)脈沖。</p><p>　　2. CPU主板和傳感器主板</p><p>　　在CPU主板上進(jìn)行數(shù)據(jù)處理和電機(jī)控制。CPU主板有兩個(gè)尺寸為7.5cm*6.0cm一樣大小的PCB。我們選擇80C196KC作為機(jī)器人的CPU。它

20、的基本行為就是根據(jù)來(lái)自于管理員通過(guò)通訊模塊的數(shù)據(jù)進(jìn)行控制電機(jī)。它還融合來(lái)自管理員和它自身傳感器的數(shù)據(jù)。80C196KC有三個(gè)PWM用來(lái)控制電機(jī)，8個(gè)信道的A/D的轉(zhuǎn)化器，用于接收自身傳感器的數(shù)據(jù)。電機(jī)驅(qū)動(dòng)是TC 4428，雙重高速M(fèi)OSFET驅(qū)動(dòng)器。在空間緊湊的環(huán)境中，我們應(yīng)用EPLD（可擦寫(xiě)編程邏輯裝置）進(jìn)行編碼計(jì)算，地址解碼和某些邏輯功能。</p><p>　　如圖2的a中所示，在機(jī)器人中，三個(gè)LEDs在CP

21、U主板上部。三個(gè)LEDs排列成一個(gè)等邊三角形。因?yàn)長(zhǎng)ED是整個(gè)框架中最耀眼的部分，因此視覺(jué)系統(tǒng)可以很容易地探測(cè)機(jī)器人的位置和旋轉(zhuǎn)。位于三角形中央的剩余的LED給管理員傳遞信息。啟動(dòng)LED意味著一個(gè)機(jī)器人利用自己的傳感裝置探測(cè)到球。</p><p>　　在傳感器主板中，4對(duì)IR傳感器組成了發(fā)送和接收裝置，其位置是固定的。在圖b)中，一對(duì)位于較高的位置來(lái)區(qū)別機(jī)器人和球。它僅僅能探測(cè)到球。另外三個(gè)位于較低的位置。它們可

22、以探測(cè)球和機(jī)器人。CPU可以識(shí)辯出它自己傳感器探測(cè)到的障礙——球和機(jī)器人。因此傳感器主板在機(jī)器人的前端，它能在機(jī)器人前面進(jìn)行區(qū)域查找。</p><p><b>　　3．通訊</b></p><p>　　從管理員到機(jī)器人，我們采用單向通訊。一般為了分享更多的信息，雙向通訊比較有利。但是這將會(huì)需要更大的空間和增大機(jī)器人和管理員執(zhí)行的任務(wù)的復(fù)雜性。在我們的系統(tǒng)中，我們把視覺(jué)

23、系統(tǒng)作為全局監(jiān)測(cè)器。對(duì)于機(jī)器人來(lái)說(shuō)不需要把它的數(shù)據(jù)傳送給管理員，因此我們采用了單向通訊。</p><p>　　圖2 CPU 主板和傳感器主板</p><p>　　用于機(jī)器人位置和旋轉(zhuǎn)的LEDs在CPU主板中的位置</p><p>　　在傳感器主板中的IR傳感器的位置</p><p>　　圖3數(shù)字化編碼數(shù)據(jù)形式圖</p><

24、p>　　圖4通訊信號(hào)組成示意圖</p><p>　　有兩種常用的通訊方法IR和R/F。IR通訊有一個(gè)問(wèn)題，就是會(huì)受光影響。因此在實(shí)際使用中，它可能出現(xiàn)故障。因此我們使用了商業(yè)R/F通訊模塊并介紹了一種高精度和高可靠性的信息傳遞數(shù)字化方法。我們把傳輸頻率設(shè)置為4kHz。兩部分產(chǎn)生一個(gè)數(shù)字化數(shù)據(jù)。在數(shù)字化數(shù)據(jù)中，如果在這兩部分之間產(chǎn)生一個(gè)狀態(tài)改變，這個(gè)數(shù)據(jù)位為1。如果沒(méi)有改變，則為0。在兩個(gè)數(shù)字化數(shù)據(jù)中，狀態(tài)改

25、變經(jīng)常發(fā)生。如圖3所示。因此數(shù)據(jù)傳輸率為2000bit/sec。我們定義了一個(gè)信道作為基本單元，一個(gè)信道有9位，一位是起始位，其余的8位是數(shù)字位。因?yàn)楣芾韱T給每個(gè)機(jī)器人發(fā)送位置和旋轉(zhuǎn)信息，每個(gè)機(jī)器人需要3個(gè)信道。我們定義了一個(gè)塊作為基本的命令單元。如圖4所示，一個(gè)塊包含17個(gè)信道，由于每個(gè)機(jī)器人需要3個(gè)信道，存在一個(gè)起始信道和一個(gè)為了將來(lái)使用的額外信道。因此，傳輸率為大約13個(gè)字每秒。這就意味著管理員能夠在每秒中給5個(gè)機(jī)器人傳送信息。&

26、lt;/p><p><b>　　4.視覺(jué)系統(tǒng)</b></p><p>　　在實(shí)時(shí)應(yīng)用中，辨別出機(jī)器人和足球的位置是很重要的。我們使用了兩個(gè)單色照相機(jī)。一個(gè)有紅色過(guò)濾器，另一個(gè)有藍(lán)色。如上所述，機(jī)器人中的LED配置形成一個(gè)等邊三角形。視覺(jué)系統(tǒng)可以很容易地探測(cè)到機(jī)器人。對(duì)于足球來(lái)說(shuō)，視覺(jué)系統(tǒng)可以通過(guò)比較兩幅圖像來(lái)探測(cè)。在比賽之前，我們調(diào)整照相機(jī)圖像的LUTs（查找表），來(lái)消除

27、除了球顏色之外的其它顏色的影響。在我們的實(shí)驗(yàn)中，只有一個(gè)單色照相機(jī)的視覺(jué)系統(tǒng)在檢測(cè)球時(shí)魯棒性差。因此，我們使用了兩個(gè)照相機(jī)。對(duì)于對(duì)手機(jī)器人的探測(cè)，MIROSOT制定了一個(gè)規(guī)則，在頂端有可以辨別隊(duì)伍顏色的一個(gè)3.5*3.5cm的顏色模塊。我們也可以通過(guò)調(diào)節(jié)LUTs使視覺(jué)系統(tǒng)能探測(cè)對(duì)手機(jī)器人。在實(shí)驗(yàn)中，視覺(jué)系統(tǒng)可以在每秒種探測(cè)到足球和10個(gè)機(jī)器人5次。</p><p>　　5．機(jī)器人的位置控制</p>

28、<p>　　機(jī)器人位置控制的方塊圖如下所示。圖5顯示了整個(gè)分塊系統(tǒng)：規(guī)劃和執(zhí)行。全局監(jiān)測(cè)環(huán)表示規(guī)劃是一個(gè)反饋環(huán)：視覺(jué)系統(tǒng)探測(cè)機(jī)器人的位置和管理員制定命令。一旦視覺(jué)系統(tǒng)探測(cè)到機(jī)器人和球的位置，管理員將根據(jù)當(dāng)前位置數(shù)據(jù)作出每個(gè)機(jī)器人的路徑規(guī)劃。這是一個(gè)緩慢的反饋過(guò)程。在表示執(zhí)行的局部控制環(huán)中，每個(gè)機(jī)器人利用解碼信號(hào)和理想位置信息進(jìn)行位置和速度控制。它的理想位置信息是由管理員提供，進(jìn)行位置和速度的局部控制。</p>&

29、lt;p>　　圖5 控制系統(tǒng)的組成圖</p><p>　　我們?cè)趫D6中給出了系統(tǒng)的配置。圖7表示一個(gè)實(shí)際機(jī)器人可以很輕松的識(shí)別出一個(gè)高爾夫球。</p><p>　　圖6系統(tǒng)配置的詳細(xì)說(shuō)明</p><p>　　圖7 實(shí)際機(jī)器人和高而夫球</p><p>　　四、機(jī)器人足球賽的協(xié)作和路徑規(guī)劃算法</p><p>　　

30、我們從兩點(diǎn)進(jìn)行了算法的考慮。一是協(xié)作，另一個(gè)是每個(gè)機(jī)器人的路徑規(guī)劃。在MIROSOT96中，規(guī)定三個(gè)機(jī)器人參加比賽。因此我們考慮三個(gè)機(jī)器人為一個(gè)多智能體系統(tǒng)。關(guān)于協(xié)作，我們采用一個(gè)機(jī)器人是守門(mén)員，其余的機(jī)器人根據(jù)不同的模式進(jìn)行角色分配。存在幾種模式。我們實(shí)驗(yàn)了4種。圖8表示4種模式?；旧?，我們采用分割—征服啟發(fā)式策略。我們把場(chǎng)地分成兩部分，分配機(jī)器人到兩個(gè)場(chǎng)地上。每個(gè)機(jī)器人根據(jù)分配的區(qū)域承擔(dān)不同的角色。模式(a)和模式(b)表示場(chǎng)地的

31、自然分割。模式(c)表示有更大的進(jìn)攻區(qū)域。模式(d)表示兩個(gè)機(jī)器人的角色不定。這些模式隨著對(duì)手策略進(jìn)行改變。我們將驗(yàn)證這些已建立的模式和開(kāi)發(fā)更有效的模式。</p><p>　　對(duì)于每個(gè)機(jī)器人的路徑規(guī)劃，我們使用了帶參量的三次樣條。價(jià)值函數(shù)有三部分組成。一是最小化曲率的變化，另一個(gè)是最小化時(shí)間，第三是避障。應(yīng)用一個(gè)機(jī)器人的實(shí)驗(yàn)表示我們的算法可以應(yīng)用于機(jī)器人足球賽。</p><p>　　我們將

32、開(kāi)發(fā)更多有效的機(jī)器人足球賽的算法。</p><p><b>　　圖8測(cè)試模式</b></p><p>　　五、結(jié)論和將來(lái)的研究</p><p>　　在本文中，我們給出了采用集中式在線系統(tǒng)的理由和應(yīng)用硬件的標(biāo)準(zhǔn)。從系統(tǒng)結(jié)構(gòu)的角度出發(fā)考慮了多種多智能體系統(tǒng)。利用兩種標(biāo)準(zhǔn)，我們對(duì)多智能體系統(tǒng)進(jìn)行了分類(lèi)?？紤]到機(jī)器人足球賽的特征，我們采用了集中式在線系

33、統(tǒng)。本文給出了我們的系統(tǒng)在應(yīng)用于機(jī)器人足球賽的合理性的理由，詳細(xì)描述了系統(tǒng)。為了分配計(jì)算負(fù)載，管理員進(jìn)行了路徑規(guī)劃，機(jī)器人執(zhí)行控制環(huán)——全局監(jiān)視和局部控制。為了局部查找，在機(jī)器人的前端有4個(gè)IR傳感器。4個(gè)LEDs放置在CPU主板中，成等邊三角形，利于視覺(jué)系統(tǒng)檢測(cè)位置和旋轉(zhuǎn)。我們?cè)O(shè)計(jì)了數(shù)字化編碼數(shù)據(jù)格式用于通訊。在分別測(cè)試后，這些很容易結(jié)合起來(lái)。我們系統(tǒng)的詳細(xì)說(shuō)明對(duì)于那些打算參加足球比賽的和想要制作相似系統(tǒng)的有一定的幫助。</p&

34、gt;<p>　　利用已經(jīng)建立的系統(tǒng)，我們將應(yīng)用模糊、神經(jīng)網(wǎng)絡(luò)、遺傳算法等工具來(lái)建立多機(jī)器人路徑規(guī)劃系統(tǒng)。目前，我們已經(jīng)建立了分割—征服啟發(fā)式策略。這需要開(kāi)發(fā)更多的有效足球賽算法并把實(shí)際系統(tǒng)與算法結(jié)合起來(lái)測(cè)試。</p><p><b>　　附錄2 外文原文</b></p><p>　　Development of a multi-agent system

35、 for robot soccer game</p><p>　　1、 Introduction</p><p>　　As modern industrial society progresses, the needs for useful robots are growing. Especially, mobile robots are special issue that gradua

36、lly expands its realm in industrial and studying topics. Researches on mobile robots have been mainly concentrated on single mobile robot. But, the development of multi-agent system is strongly needed by the growth of co

37、mplexity of tasks for robots to perform.</p><p>　　The multi-agent systems have been studied by many researchers[l-3]. Generally, multi-agent system is defined as the system composed of more than 2 robots [4]

38、 and performs the given task by cooperation. The system has some different factors compared with single robot system. First, the environment for robots to confront is dynamic. In multi-agent system, the robots themselves

39、 constitute dynamic environment, because each robot should recognize the other robots as moving obstacles. Many previous re</p><p>　　Robot soccer is an interesting domain for studying the multi-agent system.

40、 The players must work together: It means a sort of cooperation. Also, they play the game in dynamic environment: predictable and unpredictable environment - our robots and opponent's robots, respectively. The main o

41、bject is to put the ball in opponent's goal as frequently as possible in presence of opponent's robots which have the same task. So, according to a situation, our robots decide which action they take -defense o&l

42、t;/p><p>　　One of the advantages of robot soccer game is direct comparison of different systems. Many robot soccer systems are gathered in some competitions. We participated in a Soccer robot competition in Tae

43、jon, Korea called MIROSOT96 [8]. MIROSOT makes some rules. The rules describe precise specification for soccer game. The playground is rectangular with its length 130cm, its width 90cm. An orange golf ball is selected as

44、 the play ball. The size of a robot is restricted within 7.5*7.5*7.5cm. One team </p><p>　　In this paper, we explain some factors to be considered in establishing complete system. First, since the system arc

45、hitecture is very important, we decide the overall system as a centralized on-line system on the basis of surveys of multi-agent systems. Second, the overall system can be divided into three parts - A robot, communicatio

46、n and vision system. We describe the specifications of components of each part and the reasons to decide them. It would be helpful for later improvement.</p><p>　　The rest of the paper is organized as follow

47、s. Section 2 gives some surveys of system architectures and selection of our soccer robot system. Section 3 gives detailed descriptions of implemented hardware, especially mini robots. Section 4 gives cooperation and pat

48、h plan algorithm for robot soccer game. Section 5 gives conclusion of this paper and presents further works.</p><p>　　2、Categorization of multi-agent system and selection of soccer robot system</p>&l

49、t;p>　　The survey on this issue is closely related in Arai's work[4]. We can categorize multi- agent system based on two criteria. One is "Who makes decision and orders?" -Centralized / Decentralized , th

50、e other is "When does the system make plans?" - On line / Offline.</p><p>　　Centralized system means that a supervisor integrates all available data, plans the behaviors of all the robots and makes

51、 commands. Since a supervisor considers all the robots simultaneously, the system can achieve the optimization of the motions of all the robots. But, as the number of robot increases, more computational power of a superv

52、isor is needed. If the supervisor makes any fault, there is no way for the robots to correct it.</p><p>　　Decentralized system means that each robot makes plan for itself on the basis of collected informatio

53、n from other robots and its own sensors. In the system, there is not considerable increasing computational load as the number of robots increases. Even if one robot fails to work, other robots work well. But, the system

54、cannot guarantee the optimization of the motions of all the robots.</p><p>　　Off-line system means that all the paths are planned before all the robots move. Because of no restriction of time and computing p

55、ower, the system can achieve optimization. But, since the system assumes static environment, it is not robust to small variation of environment. In real world, it may malfunction with some variances.</p><p>

56、　　On-line system means real-time planning. It is robust to dynamic environment. But, it needs large computational power and effective algorithm.</p><p>　　In a lot of researches, the above two categorizations

57、 are interrelated each other. We summarize the researches in Table 1.</p><p>　　Table 1. Categorization of multi-agent system</p><p>　　In MIROSOT96, The size of a robot is restricted. Therefore,

58、it is difficult to implement a robot with large computational power. Also, it is important to decide what a robot must equip. Basically, a robot has to equip actuator module and communication module. Additional equipment

59、s are selected and implemented considering the space of a robot. Soccer game needs global information of our robots' position. So, we decide the centralized system as our system. From the viewpoint of path planning t

60、ime,</p><p>　　In centralized on-line system, a supervisor acquires all available information of whole environment and the robots. Simultaneously, a supervisor should plan all the robots' paths in real-ti

61、me. Therefore it requires fast computing power. To decrease the burden of a supervisor, we choose partitioned system which separate main planning and executing. This is somewhat similar to the works of Shakey[15], Firby[

62、16]. In our soccer robot system, a supervisor makes plans of all the robots on the basis of</p><p>　　Figure 1. The constitution of our soccer robot system</p><p>　　3、System implementation</p&

63、gt;<p>　　System is composed of three parts - a supervisor, vision and 5 robots. A supervisor is a PC - pentium processor -which makes plan in real time, and vision system has two cameras which have red and blue fi

64、lter respectively, and image processing board which has a DSP and memory. A robot has a CPU, communication module, IR sensors, motors, etc. These three parts are related one another. Detailed descriptions are given in fo

65、llowing chapters.</p><p>　　1. Configuration of individual robot</p><p>　　A robot is consisted of mechanical part, CPU board, communication(receiver) module and sensor board. Its size is within 7

66、.5*7.5*7.5 cm.</p><p>　　2. Mechanical part</p><p>　　Mechanical part of a robot is consisted of two motors, encoders, gearheads, wheels, a ball caster, and a frame. The frame is designed for easy

67、, compact and hardy integration. Motors and gearheads are selected in consideration for operating voltage, internal resistance, mechanical time constant. Its operating voltage is 6V. Reduction ratio of gearhead is 1:41.

68、A diameter of a wheel is 32 mm. The no-load speed of a motor is 15200 rev/min. So, no-load speed of robots can be calculated as about 62 c</p><p>　　3. CPU board and sensor board</p><p>　　Figure

69、2. The CPU board and sensor board</p><p>　　the locations of LEDs in CPU board forrobot's position and rotation</p><p>　　the locations of IR sensors in sensor board</p><p>　　In C

70、PU board, data processing and motor control are carried out. CPU board consists of two PCBs - same size of 7.5cm*6.0cm. We choose 80C196KC as robot's CPU. Its basic operation is to control motors according to the dat

71、a from a supervisor via communication module. Also, it fuses the data from a supervisor and sensors equipped in itself. 80C196KC has three PWM generators which are used to control motors and 8 channel A/D converters whi

72、ch are used to receive the data from its own sensors. Motor </p><p>　　As can be seen in Figure 2(a), 4 LEDs are in CPU board located above in a robot. Three LEDs make an isosceles triangle. Because LED is th

73、e most brightest thing in playground, vision system detects the position and rotation of a robot easily. The other LED located at the center of triangle is used to provide some information to a supervisor. Turning on the

74、 LED means that a robot detects the ball by its own sensors.</p><p>　　In sensor board, 4 pairs of IR sensor consisted of transmitter and receiver are located in fixed positions. As can be seen in Figure 2(b)

75、, one pair is located in higher position to distinguish a robot from the ball. It can detect only the ball. The other three pairs are located in the lower position. They can detect the ball and robots. So, a CPU can reco

76、gnize the obstacle detected by its own sensors - a ball and a robot. Since sensor board is in front of a robot, it can perform local searching </p><p>　　Figure 3. The shape of digitally coded data</p>

77、<p>　　Figure 4. The constitution of communication signal</p><p>　　3. Communication</p><p>　　We choose unidirectional communication - from a supervisor to robots. Generally, to share more i

78、nformation, bi-directional communication is better. But, it needs more space and increases the complexity of tasks which robots and a supervisor carry out. In our system, we use vision system as a global monitor. So, it

79、is not necessary for a robot to transmit its data to a supervisor. Therefore, we adopt unidirectional communication.</p><p>　　There are two prevalent communication methods - IR and R/F. IR communication has

80、a problem such that it is affected by light. So, in real competition, it may malfunction. Therefore, we modify the commercial R/F communication module and introduce a digital method for high precision and good reliabilit

81、y of information transfer. We set the carrier frequency as 4kHz. Two sections make one digital data. If there is a state change between two sections in the digital data, this digital data means " 1" </p>

82、<p>　　future purpose. So, the transfer rate is about 13 block per second. It means that a supervisor can transfer information to five robots 13 times per second.</p><p>　　4. Vision system</p>&l

83、t;p>　　It is very important to recognize the positions of robots and the ball in real-time. We use two monochrome cameras. One has red filter and the other has blue filter. As mentioned above, a robot has an isosceles

84、triangle shaped LED configuration. Vision system can easily detect a robot. As for a ball, vision system compare two image for detecting the ball. We regulate LUTs( Look Up Table ) of image data from two cameras to elimi

85、nate other colors except a ball color before competition. In our expe</p><p>　　5．Position control of a robot</p><p>　　The block diagram of position control of a robot is presented in Figure 5. T

86、he figure shows the partitioned system : Planning and executing. Global Monitoring loop representing the Planning is a feedback loop in which vision system detect a robot position and a supervisor makes commands. As soon

87、 as vision system detects positions of robots and the ball, a supervisor makes path planning of each robot according to the current position data. It is slow feedback. In Local Control loop representing e</p><

88、p>　　Figure 5. Constitution of Control system</p><p>　　Figure 6. The configuration of our system.a detailed description.</p><p>　　Figure 7. The actual robot with a golf ball</p><p&g

89、t;　　4、Cooperation and path plan algorithm for robot soccer game</p><p>　　We are making algorithms in consideration for two points. One is for cooperation; the other is for path plan of each robot. In MIROSOT

90、96, three robots were permitted to play game. Therefore, we make algorithms considering three robots as a multi-agent system. For viewpoint of cooperation, one robot is a goal keeper, the others takes the roles according

91、 to the modes. There can be several modes. We tested 4 modes. Figure 8 shows 4 modes. Basically, we use divide-and-conquer heuristics. We divide t</p><p>　　For path plan of each robot, we use parametric cubi

92、c spline. The cost function is consisted of three term. One is for minimization of variations of curvature, the other is for minimization of time, the third is for obstacle avoidance. The results of one robot experiment

93、show that our algorithm can be applied to robot soccer game[17].We are going to develop more effective algorithm for robot soccer game. </p><p>　　Figure 8. The 4 tested modes</p><p>　　5、Conclusi

94、ons and further works</p><p>　　In this paper, we provide the reasons why we adopt centralized on-line system as our robot soccer system and specification of implemented hardware. We survey many multi-agent s

95、ystems from the viewpoint of system architecture. Using two criteria, we categorize the multi-agent systems. Then, we select the centralized on-line system considering the nature of robot soccer game. This paper explains

96、 suitability of our system to robot soccer game for many reasons. Also, we explain our system in detail.</p><p>　　Using established system, we are going to build multiple robot path planning system using sev

97、eral techniques - fuzzy, neural network, genetic algorithm and so on. At present, We have established divide-and-conquer heuristics. It is needed to develop more effective soccer algorithms and test our algorithm combine