摘" 要:多無人車的編隊(duì)控制和編隊(duì)避障是無人車領(lǐng)域研究的重要問題之一。針對多無人車的編隊(duì)控制,提出一種基于無人車之間的距離和角度控制的領(lǐng)航跟隨編隊(duì)控制算法,并對算法的收斂性進(jìn)行驗(yàn)證。針對多無人車的編隊(duì)避障,提出一種自適應(yīng)編隊(duì)避障方法,由領(lǐng)航無人車規(guī)劃出編隊(duì)的行駛路徑,對路徑中的大尺寸障礙物采用動(dòng)態(tài)隊(duì)形變換實(shí)現(xiàn)編隊(duì)避障,對路徑中的小尺寸障礙物采用改進(jìn)動(dòng)態(tài)窗口法實(shí)現(xiàn)跟隨無人車的自主避障。最后,在Matlab中對編隊(duì)算法和自適應(yīng)編隊(duì)避障方法進(jìn)行實(shí)驗(yàn)驗(yàn)證,實(shí)驗(yàn)結(jié)果表明所提算法能實(shí)現(xiàn)無人車的編隊(duì)控制和編隊(duì)避障,且取得較好的隊(duì)形穩(wěn)定性和有效性。
關(guān)鍵詞:無人車編隊(duì);領(lǐng)航跟隨法;自適應(yīng)編隊(duì)避障;動(dòng)態(tài)隊(duì)形變換;動(dòng)態(tài)窗口法
中圖分類號:TP242.6" " " 文獻(xiàn)標(biāo)志碼:A" " " " " 文章編號:2095-2945(2023)25-0001-05
Abstract: The formation control and obstacle avoidance of multi-unmanned vehicles is one of the important issues in the field of unmanned vehicles. Aiming at the formation control of multi-unmanned vehicles, a pilot following formation control algorithm based on distance and angle control between unmanned vehicles is proposed, and the convergence of the algorithm is verified. Aiming at the formation obstacle avoidance of many unmanned vehicles, an adaptive formation obstacle avoidance method is proposed. The driving path of the formation is planned by the pilot unmanned vehicle, and the large size obstacles in the path are realized by dynamic formation transformation. For the small size obstacles in the path, the improved dynamic window method is used to realize the autonomous obstacle avoidance of following the unmanned vehicle. Finally, the experimental verification of the formation algorithm and adaptive formation obstacle avoidance method is carried out in Matlab, and the experimental results show that the proposed algorithm can achieve formation control and formation obstacle avoidance of unmanned vehicles, and achieve good formation stability and effectiveness.
Keywords: unmanned vehicle formation; pilot following method; adaptive formation obstacle avoidance; dynamic formation transformation; dynamic window method
無人車是一種具有自主決策規(guī)劃能力的智能設(shè)備。隨著無人車在區(qū)域巡檢、災(zāi)難救援、軍事偵查、生活服務(wù)等領(lǐng)域的應(yīng)用,人們對無人車的需求也趨向于多元化,尤其對多無人車系統(tǒng)的需求愈發(fā)強(qiáng)烈。多無人車系統(tǒng)通過相互之間的協(xié)調(diào)與配合可提高無人車完成既定任務(wù)的效率與可靠性。此外,多無人車系統(tǒng)在適應(yīng)性、時(shí)效性、協(xié)調(diào)性等方面具有顯著優(yōu)勢,因而成為近些年無人車領(lǐng)域研究的重點(diǎn)方向之一。
無人車的編隊(duì)控制是多無人車系統(tǒng)最典型的協(xié)同方法之一。目前,無人車編隊(duì)控制方法有領(lǐng)航跟隨法[1]、基于行為法[2]、人工勢場法[3]、虛擬結(jié)構(gòu)法[4]等。Jonathan等[5]提出了一種基于無姿態(tài)測量的領(lǐng)航跟隨法;Duan等[6]通過引入貪心算法、基于行為法、虛擬結(jié)構(gòu)法提出了一種基于領(lǐng)航跟隨的編隊(duì)變換方法;Trinh等[7]提出了一種基于角度信息的無人車的編隊(duì)控制算法;賈海峰[8]提出了一種基于距離角度信息的無人車編隊(duì)算法;郝金玉等[9]以圖理論為基礎(chǔ)提出了基于領(lǐng)航跟隨法協(xié)同和人工勢場避障相結(jié)合的無人艇協(xié)同編隊(duì)控制算法。
編隊(duì)避障也是無人車編隊(duì)控制需要考慮的問題,相比于單無人車的避障,多無人車編隊(duì)不僅要避開障礙物,而且還要保證編隊(duì)隊(duì)形的完整。付雷等[10]使用改進(jìn)人工勢場法實(shí)現(xiàn)了無人車的編隊(duì)避障;Wu等[11]利用縮小觀測角的方法實(shí)現(xiàn)了無人車的編隊(duì)避障;Yan等[12]利用構(gòu)造勢函數(shù)法實(shí)現(xiàn)了無人車的編隊(duì)避障。
本文在前人研究的基礎(chǔ)上,根據(jù)差速無人車的運(yùn)動(dòng)模型和無人車的位置關(guān)系提出了一種基于距離和角度控制的領(lǐng)航跟隨編隊(duì)算法,算法可通過改變跟隨無人車與領(lǐng)航無人車之間的理想距離和理想觀測角實(shí)現(xiàn)編隊(duì)隊(duì)形變換。針對無人車的編隊(duì)避障提出了一種自適應(yīng)編隊(duì)避障方法,該方法將動(dòng)態(tài)隊(duì)形變換與跟隨無人車自主避障2種避障方式結(jié)合,提高了無人車編隊(duì)的避障能力。最后在Matlab進(jìn)行了仿真驗(yàn)證,結(jié)果驗(yàn)證了所提算法的有效性與合理性。
由1.2分析可知,eθ保持穩(wěn)定即可,不必最終收斂到0,由于ωl光滑有界,可被視為有界擾動(dòng)項(xiàng),將式(11)線性化可知其局部穩(wěn)定,因此系統(tǒng)的零動(dòng)態(tài)穩(wěn)定,則整個(gè)編隊(duì)系統(tǒng)穩(wěn)定,可實(shí)現(xiàn)無人車的編隊(duì)控制。
2" 自適應(yīng)編隊(duì)避障方法設(shè)計(jì)
多無人車編隊(duì)在執(zhí)行任務(wù)時(shí),編隊(duì)的行駛路徑通常由領(lǐng)航無人車規(guī)劃。現(xiàn)存的無人車編隊(duì)避障方案多為編隊(duì)整體避障和跟隨無人車自主避障。編隊(duì)整體避障可以保持編隊(duì)隊(duì)形,但避障效率低,靈活性不好;無人車自主避障在對大尺寸障礙物避障時(shí)會(huì)出現(xiàn)避障時(shí)間長等問題。針對以往編隊(duì)避障存在的問題,提出一種自適應(yīng)編隊(duì)避障方法,方法的核心是根據(jù)障礙物的尺寸動(dòng)態(tài)選擇避障方法。對大尺寸障礙物采用動(dòng)態(tài)隊(duì)形變換實(shí)現(xiàn)編隊(duì)避障;對小尺寸障礙物采用改進(jìn)動(dòng)態(tài)窗口法實(shí)現(xiàn)跟隨無人車自主避障。
2.1" 動(dòng)態(tài)隊(duì)形變換避障
無人車編隊(duì)的常見隊(duì)形有柱形、三角形、一字形等。無人車編隊(duì)若要使用動(dòng)態(tài)隊(duì)形變換避障,如何選取最優(yōu)變換隊(duì)形是問題的關(guān)鍵。
無人車編隊(duì)的仿真結(jié)果如圖3所示。
從圖3可見,無人車沿直線-曲線組合軌跡運(yùn)動(dòng)時(shí),保持了比較理想的編隊(duì)。
3.2" 自適應(yīng)編隊(duì)避障方法仿真驗(yàn)證
3輛無人車的初始位姿分別為(0 0 π/4),(-2 0 π/4),(0 -2 π/4)。無人車編隊(duì)隊(duì)形數(shù)據(jù)庫有3種隊(duì)形,分別是三角形、收縮三角形、柱形,3種隊(duì)形的理想距離和理想觀測角見表3,表中數(shù)據(jù)以(d,φ)形式表示。
為觀測無人車編隊(duì)在行進(jìn)過程中動(dòng)態(tài)隊(duì)形變換和跟隨無人車自主避障。在圖4中設(shè)置了一些障礙物。
無人車編隊(duì)首先以三角形隊(duì)形行進(jìn),通過障礙物1和障礙物2之間的通道時(shí),領(lǐng)航無人車判斷ρgt;1,編隊(duì)保持隊(duì)形繼續(xù)行進(jìn);通過障礙物3和障礙物4之間通道時(shí),領(lǐng)航無人車判斷ρ?1,隊(duì)形變換為收縮三角形;通過障礙物5和障礙物6之間通道時(shí),領(lǐng)航無人車判斷ρ?1,隊(duì)形變換為柱形,編隊(duì)駛出障礙物5與障礙物6之間的通道后,隊(duì)形恢復(fù)最初的三角形編隊(duì)。隨后,跟隨無人車遇到小尺寸障礙物,跟隨無人車自主避障完成后恢復(fù)編隊(duì)隊(duì)形。從無人車編隊(duì)避障的軌跡可以看出,動(dòng)態(tài)隊(duì)形變換以及改進(jìn)動(dòng)態(tài)窗口法均能實(shí)現(xiàn)無人車的編隊(duì)避障。
4" 結(jié)論
本文以差速無人車為研究對象,提出了一種基于距離與角度的領(lǐng)航跟隨編隊(duì)算法。算法可通過改變理想距離和理想觀測角實(shí)現(xiàn)編隊(duì)隊(duì)形變換。針對無人車編隊(duì)避障提出了一種自適應(yīng)編隊(duì)避障方法,方法根據(jù)障礙物的尺寸選擇避障算法,大尺寸障礙物采用動(dòng)態(tài)隊(duì)形變換避障,小尺寸障礙物采用改進(jìn)動(dòng)態(tài)窗口法避障。通過2種避障方法的結(jié)合,提高了無人車編隊(duì)的穩(wěn)定性和適應(yīng)性。最后在Matlab中對所提算法進(jìn)行了仿真實(shí)驗(yàn),仿真實(shí)驗(yàn)結(jié)果表明所提算法既可以實(shí)現(xiàn)穩(wěn)定的編隊(duì)效果,又可以較好地實(shí)現(xiàn)編隊(duì)避障。
參考文獻(xiàn):
[1] LI H Y,LI H L,LI Z Y. The design of following controller for autonomous vehicle based on leader-follower strategies[J]. MATEC Web of Conferences,2022:355.
[2] LEE G, CHWA D. Decentralized behavior-based formation control of multiple robots considering obstacle avoidance[J]. Intelligent Service Robotics, 2018,11(1):127-138.
[3] SANG H, YOU Y, SUN X, et al. The hybrid path planning algorithm based on improved A* and artificial potential field for unmanned surface vehicle formations[J]. Ocean Engineering, 2021,223(3-4):108709.
[4] CHEN X L, HUANG F H, ZHANG Y G, et al. A novel virtual-structure formation control design for mobile robots with obstacle avoidance[J].Applied Sciences,2020,10(17):5807.
[5] JONATHAN H A,JAVIER P, CéSAR C H, et al. Leader-follower formation control of wheeled mobile robots without attitude measurements[J].Applied Sciences,2021,11(12):5639.
[6] DUAN Y, GAO Z, PENG Z, et al. Formation transformation based on leader-follower algorithm[J].International journal of technology and human interaction, 2019,15(3):28-46.
[7] TRINH M H, OH K K, JEONG K, et al. Bearing-only control of leader first follower formations[J].IFAC PapersOnLine, 2016, 49(4):7-12.
[8] 賈海峰.多無人車編隊(duì)與路徑規(guī)劃[D].大連:大連理工大學(xué),2021.
[9] 郝金玉,范雪,楊少龍,等.基于動(dòng)態(tài)目標(biāo)點(diǎn)協(xié)同的無人艇編隊(duì)路徑規(guī)劃[J].兵器裝備工程學(xué)報(bào),2022,43(9):1-8.
[10] 付雷,秦一杰,何頂新,等.基于改進(jìn)人工勢場法的多機(jī)器人編隊(duì)避障[J].控制工程,2022,29(3):388-396.
[11] WU X R, WANG S S, XING M Y. Observer-based leader-following formation control for multi-robot with obstacle avoidance[J]. IEEE Access,2019,7:14791-14798.
[12] YAN J, GUAN X P, LUO X Y, et al. Formation control and obstacle avoidance for multi-agent systems based on virtual leader-follower strategy[J].International Journal of Information Technologyamp;Decision Making,2017,16(3):865-880.