杜?續(xù),宋?康,謝?輝
無人碾壓機(jī)軌跡跟蹤算法及能耗規(guī)律研究
杜?續(xù),宋?康,謝?輝
(天津大學(xué)機(jī)械工程學(xué)院,天津 300072)
無人碾壓機(jī)是降低人工作業(yè)負(fù)擔(dān)、改善碾壓作業(yè)品質(zhì)和效率的重要技術(shù)途徑.但由于碾壓機(jī)噸位高、轉(zhuǎn)向阻力大,且經(jīng)常工作在復(fù)雜非結(jié)構(gòu)路面上,常規(guī)車輛軌跡跟蹤算法的能耗通常較高,影響了控制系統(tǒng)的電能平衡、可靠性以及綜合能效.為此,針對運(yùn)行在水電大壩上的無人碾壓機(jī),提出了面向節(jié)能的串級抗擾軌跡跟蹤算法,同時研究了該算法對無人碾壓機(jī)轉(zhuǎn)向系統(tǒng)能耗的影響規(guī)律.首先,針對碾壓機(jī)高噸位造成的轉(zhuǎn)向速度慢、尋跡易超調(diào)的問題,采用位置預(yù)瞄算法對轉(zhuǎn)向系統(tǒng)進(jìn)行提前控制,進(jìn)而減少超調(diào)量,降低能耗.其次,針對車輛受路面起伏干擾定位量測噪聲大的問題,提出了車身姿態(tài)校正算法,通過測取無人碾壓機(jī)前、后車身的橫滾角對定位結(jié)果進(jìn)行修正,減少噪聲干擾,降低能耗.最后,為定位中小幅度高頻噪聲的干擾,通過采用非線性誤差反饋控制律控制,降低方向盤在小距離誤差帶內(nèi)的高頻波動,實(shí)現(xiàn)轉(zhuǎn)向系統(tǒng)節(jié)能.在水電大壩建設(shè)現(xiàn)場開展了實(shí)車試驗(yàn),結(jié)果表明;?①采用優(yōu)化的預(yù)瞄距離與無預(yù)瞄的標(biāo)稱控制器相比,可降低轉(zhuǎn)向電機(jī)能量消耗24.7%,使得軌跡跟蹤精度穩(wěn)定在?±0.15m;②采用姿態(tài)校正算法可以減少轉(zhuǎn)向電機(jī)29.2%的能量消耗,并將軌跡跟蹤精度改善了33.0%;③采用非線性誤差反饋控制律在距離誤差無明顯惡化(0.06m內(nèi))情況下,可降低轉(zhuǎn)向電機(jī)能耗31.8%.
無人碾壓機(jī);軌跡跟蹤;能耗優(yōu)化控制;預(yù)瞄算法;姿態(tài)校正
振動碾壓機(jī)作為一種高效的壓實(shí)機(jī)械,廣泛應(yīng)用于大壩、道路等工程建設(shè)中[1].無人碾壓機(jī)在減輕操作人員的工作量、提高壓實(shí)作業(yè)效率和質(zhì)量方面具有廣闊的應(yīng)用前景.精確和節(jié)能的軌跡跟蹤顯然對無人碾壓機(jī)的性能和效率至關(guān)重要.
然而,與乘用車相比,無人碾壓機(jī)的橫向控制更具挑戰(zhàn)性,轉(zhuǎn)向系統(tǒng)能耗更大.主要原因如下:①碾壓機(jī)的質(zhì)量通常大于20t,比乘用車重10倍以上,轉(zhuǎn)向系統(tǒng)的阻力扭矩更大;②碾壓機(jī)工作路面非結(jié)構(gòu),多數(shù)是大粒徑塊石或斜坡,這些因素一方面導(dǎo)致轉(zhuǎn)向阻力扭矩進(jìn)一步增大,另一方面造成了車身姿態(tài)變化波動,并導(dǎo)致車載GPS(全球定位系統(tǒng))天線的定位定性波動;③碾壓機(jī)是一個多體系統(tǒng)的鉸接結(jié)構(gòu)[2],其多運(yùn)動自由度特征,減弱車輛橫向剛度,進(jìn)一步增大了軌跡跟蹤控制的難度[3-4].
針對鉸接式車輛的軌跡跟蹤算法,大量學(xué)者提出了不同的控制方案.Uzunsoy等[5]以及后來的Khalaji[6]提出了基于模糊邏輯的PID(比例-積分-微分)控制器,用于分析橫向誤差和航向誤差.但是因?yàn)榭刂破鲉栴}是高度非線性的,所以基于PID的控制算法響應(yīng)速度慢,并且通常需要繁瑣的控制參數(shù)整定[7].
為解決PID控制響應(yīng)速度慢等問題,基于模型的軌跡跟蹤算法受到廣泛關(guān)注.邵俊愷等[8]基于地下礦車的運(yùn)動學(xué)模型利用強(qiáng)化學(xué)習(xí)算法對PID參數(shù)進(jìn)行自整定,提高控制系統(tǒng)的魯棒性和快速性.Khalaji等[9]和Erkan等[10]對載重汽車(LHD)的非線性運(yùn)動學(xué)模型、動力學(xué)模型以及魯棒控制進(jìn)行討論.Nayl?等[11]分析了運(yùn)動學(xué)參數(shù)對模型預(yù)測控制器(MPC)的影響.之后,Yue等[12]利用MPC實(shí)現(xiàn)了對拖掛車系統(tǒng)軌跡的穩(wěn)定跟蹤.近年來,針對碾壓機(jī)基于模型的軌跡控制逐漸形成熱點(diǎn).Bian等[13]對碾壓機(jī)運(yùn)動學(xué)建模,并基于Lyapunov穩(wěn)定性設(shè)計(jì)了控制律.Xu?等[14]基于碾壓機(jī)運(yùn)動學(xué)模型提出了一種分層的主動干擾抑制控制器,顯然,對于碾壓機(jī)與地面間存在的側(cè)滑效應(yīng)未考慮.而Nayl等[15]提出了一種考慮時變滑移角的MPC(模型預(yù)測控制)解決方案.Yang等[16]提出了一種具有側(cè)滑補(bǔ)償?shù)哪雺簷C(jī)軌跡跟蹤控制方法.但是上述方案都需要較高的計(jì)算量和精確的模型.
然而,上述研究的重點(diǎn)在軌跡跟蹤精度以及穩(wěn)定性能上,關(guān)于控制過程中的能耗問題卻鮮有研究.對于無人駕駛工程車輛,高電能消耗使得控制系統(tǒng)需要匹配高扭矩轉(zhuǎn)向電機(jī)、大容量電池和大功率發(fā)電機(jī),這不僅增大了成本,也影響了可靠性,成為工程應(yīng)用中的難題.
為此,本文提出了一種面向節(jié)能的串級抗擾軌跡跟蹤算法.在串級主動抗擾控制器框架下,引入預(yù)瞄算法主動適應(yīng)碾壓機(jī)轉(zhuǎn)向速度慢的問題,減小軌跡跟蹤控制過程中的超調(diào)量;添加姿態(tài)校正算法,修正定位坐標(biāo),抑制噪聲,降低能耗;在距離誤差控制器環(huán)節(jié)設(shè)置非線性誤差反饋控制律來降低小距離誤差帶內(nèi)方向盤的高頻波動.在水電大壩填筑現(xiàn)場實(shí)車試驗(yàn),對算法進(jìn)行了試驗(yàn)驗(yàn)證,量化分析了節(jié)能效果.
如圖1所示,無人碾壓機(jī)是一個典型的由前壓輥和后車身組成的多體系統(tǒng).本研究使用的無人碾壓機(jī)試驗(yàn)平臺采用的主要傳感器的功用以及關(guān)鍵參數(shù)如表1所示,執(zhí)行機(jī)構(gòu)-轉(zhuǎn)向電機(jī)的關(guān)鍵參數(shù)如表2?所示.
圖1?無人碾壓機(jī)測控系統(tǒng)布置
表1?主要傳感器關(guān)鍵參數(shù)
Tab.1?Key parameters of the main sensors
表2?轉(zhuǎn)向電機(jī)關(guān)鍵參數(shù)
Tab.2?Key parameters of the steering motor
圖2?碾壓機(jī)運(yùn)動學(xué)模型
該模型為被控對象,在該模型上開展面向節(jié)能的串級抗擾軌跡跟蹤算法的設(shè)計(jì)和分析研究.
基于第1.2節(jié)中的運(yùn)動學(xué)模型,設(shè)計(jì)面向節(jié)能的串級抗擾軌跡跟蹤算法,其控制架構(gòu)如圖3所示.分別以橫向距離誤差和航向角控制作為串級控制的內(nèi)環(huán)和外環(huán),最終計(jì)算所需的方向盤轉(zhuǎn)角,實(shí)現(xiàn)目標(biāo)軌跡跟蹤.
圖3?面向節(jié)能的串級抗擾軌跡跟蹤算法架構(gòu)
將式(1)帶入到式(3)簡化后得
根據(jù)式(4)設(shè)計(jì)如下抗擾控制器:
由此,將式(5)改造為積分器:
針對式(6)設(shè)計(jì)比例控制器:
首先計(jì)算俯仰產(chǎn)生后對無人碾壓機(jī)橫向運(yùn)動造成的距離誤差,即
圖4?姿態(tài)校正位姿解算示意
針對轉(zhuǎn)向和車輛航向動態(tài)變化慢的問題,設(shè)計(jì)預(yù)瞄控制算法[21],實(shí)現(xiàn)對方向盤轉(zhuǎn)角的提前控制,以減少控制波動.預(yù)瞄算法中的位置重構(gòu)方法為
車輛前剛體定位點(diǎn)距離目標(biāo)軌跡線的垂直距離計(jì)算式為
圖5?非線性誤差反饋控制律
控制律的計(jì)算式為
圖6?無人碾壓機(jī)碾壓作業(yè)示意
為了研究算法的節(jié)能效果,現(xiàn)建立如下幾個評價指標(biāo).
針對不同預(yù)瞄距離的節(jié)能效果,在相同的試驗(yàn)軌跡上分別對預(yù)瞄距離為0m、1m、2m、3m、4m、5m、6m進(jìn)行了碾壓作業(yè)試驗(yàn),碾壓試驗(yàn)倉面縱向長度為100m.試驗(yàn)結(jié)果如下.
如圖7(a)和(b)所示,預(yù)瞄算法的使用,使得控制過程超調(diào)量減小,無人碾壓機(jī)運(yùn)行軌跡趨于穩(wěn)定收斂,軌跡誤差波動降低,距離誤差穩(wěn)定在0.20m以內(nèi).其次,如圖7(c)所示,預(yù)瞄距離為3m時與無預(yù)瞄算法相比,距離誤差更為集中于小距離誤差帶.
如圖8所示,=3m、=6m與無預(yù)瞄對比,轉(zhuǎn)向電機(jī)調(diào)節(jié)的頻率變快,調(diào)節(jié)的幅度降低了49.1%,驗(yàn)證了預(yù)瞄算法主動適應(yīng)轉(zhuǎn)向速度慢問題的有效性.
根據(jù)表3和圖9可知,采用3m的預(yù)瞄距離可以兼顧轉(zhuǎn)向電機(jī)能耗與軌跡跟蹤精度,轉(zhuǎn)向電機(jī)能耗較無預(yù)瞄(=0m)降低24.7%,且距離誤差穩(wěn)定在0.15m以內(nèi).
圖8?不同預(yù)瞄距離下無人碾壓機(jī)轉(zhuǎn)向電機(jī)性能
表3?不同預(yù)瞄距離下無人碾壓機(jī)軌跡跟蹤控制的性能
Tab.3 Path-following control performance of unmanned roller under different preview distances
圖9?不同預(yù)瞄距離下無人碾壓機(jī)運(yùn)行性能
為了研究姿態(tài)校正算法對無人碾壓機(jī)性能改善情況,在相同試驗(yàn)軌跡上分別打開或關(guān)閉姿態(tài)校正算法在縱向長度為150m的碾壓試驗(yàn)倉面上進(jìn)行碾壓作業(yè).試驗(yàn)結(jié)果如下.
如圖10所示,姿態(tài)校正算法的引入主要是對橫向坐標(biāo)進(jìn)行了校正,即在橫向運(yùn)動上進(jìn)行了補(bǔ)償,減緩了定位坐標(biāo)波動的幅度,降低了倉面起伏對定位的噪聲干擾.
如圖11和表4所示,姿態(tài)校正算法在對定位補(bǔ)償后,距離誤差的標(biāo)準(zhǔn)差提高了40.0%,距離誤差絕對值的平均值減少了0.04m,整個試驗(yàn)周期的83.9%,距離誤差都穩(wěn)定在0.10m以內(nèi),轉(zhuǎn)向電機(jī)小幅度調(diào)節(jié),轉(zhuǎn)向電機(jī)瞬時功率得到改善,轉(zhuǎn)向電機(jī)的能耗降低了29.2%.
圖10?姿態(tài)校正算法對定位坐標(biāo)的校正
圖11?姿態(tài)校正前后碾壓機(jī)各項(xiàng)指標(biāo)性能
表4 姿態(tài)校正算法對無人碾壓機(jī)軌跡跟蹤控制性能的影響
Tab.4 Influence of attitude correction controller on the path-following control performance of unmanned roller
針對非線性誤差反饋控制律的節(jié)能效果,在相同試驗(yàn)軌跡上分別對距離死區(qū)為0m、0.03m、0.05m、0.10m、0.13m和0.18m進(jìn)行了碾壓作業(yè)試驗(yàn),建立碾壓試驗(yàn)倉面縱向長度為180m.試驗(yàn)結(jié)果如下.
圖12?不同距離死區(qū)下無人碾壓機(jī)的循跡情況
表5 不同距離死區(qū)下無人碾壓機(jī)軌跡跟蹤控制的性能
Tab.5 Path-following control performance of unmanned roller under different distances of dead zones
圖13?不同距離死區(qū)下無人碾壓機(jī)轉(zhuǎn)向電機(jī)性能
圖14?不同距離死區(qū)下無人碾壓機(jī)運(yùn)行性能總圖
本文針對典型的鉸接車輛碾壓機(jī)在復(fù)雜倉面上的節(jié)能控制問題進(jìn)行了研究分析.首先搭建了無人碾壓機(jī)運(yùn)動學(xué)模型,其次,針對無人碾壓機(jī)軌跡跟蹤算法能耗高的問題,設(shè)計(jì)了面向節(jié)能的串級抗擾軌跡跟蹤算法,通過大壩填筑現(xiàn)場實(shí)車試驗(yàn),驗(yàn)證了該算法在典型鉸鏈?zhǔn)杰囕v上的節(jié)能控制效果.結(jié)論如下.
(1) 采用預(yù)瞄算法,利用未來道路信息主動適應(yīng)轉(zhuǎn)向速度慢的問題,通過優(yōu)化預(yù)瞄距離,轉(zhuǎn)向系統(tǒng)電能消耗比無預(yù)瞄的標(biāo)稱控制器降低了24.7%,且能夠使得距離誤差穩(wěn)定在0.15m內(nèi).
(2) 采用姿態(tài)校正算法,消除了位置和方位測量中的噪聲干擾,同時可以消除碾壓機(jī)在大巖石上或在有斜坡的道路上壓實(shí)時,碾壓機(jī)的擺動而引起的定位偏差問題,從而降低29.2%的轉(zhuǎn)向系統(tǒng)能量消耗,距離誤差降低了33.0%.
(3) 通過使用非線性誤差反饋控制律,降低方向盤在小距離誤差帶內(nèi)的高頻波動,可以在距離誤差無明顯惡化(0.06m以內(nèi))的情況下,將轉(zhuǎn)向系統(tǒng)的能耗降低31.8%.
最后,本文的研究內(nèi)容主要關(guān)注車輛轉(zhuǎn)向系統(tǒng)的能耗,尚未涉及整車的能耗分析.因此,后續(xù)的研究工作將會結(jié)合鉸接式車輛的整車動力學(xué)開展進(jìn)一步分析,評估控制算法對整車能耗的影響規(guī)律.
[1] Zhang Q,Liu T,Zhang Z,et al. Unmanned rolling compaction system for rockfill materials[J]. Automation in Construction,2019,100:103-117.
[2] Gao Y,Cao D,Shen Y. Path-following control by dynamic virtual terrain field for articulated steer vehicles[J/OL]. Vehicle System Dynamics,https://doi. org/10. 1080/00423114. 2019. 1648837,2019-09-31.
[3] Zhang Y,Khajepour A,Hashemi E,et al. Reconfigurable model predictive control for articulated vehicle stability with experimental validation[J]. IEEE Transactions on Transportation Electrification,2020,6(1):308-317.
[4] Liu S,Hou Z,Tian T,et al. Path tracking control of a self-driving wheel excavator via an enhanced data-driven model-free adaptive control approach[J]. IET Control Theory & Applications,2020,14(2):220-232.
[5] Uzunsoy E,Erkilic V. Development of a trajectory following vehicle control model[J]. Advances in Mechanical Engineering,2016,8(5):1-11.
[6] Khalaji A K. PID-based target tracking control of a tractor-trailer mobile robot[J]. ARCHIVE Proceedings of the Institution of Mechanical Engineers Part C:Journal of Mechanical Engineering Science,2019,233(13):4776-4787.
[7] Fang X,Bian Y,Yang M,et al. Development of a path following control model for an unmanned vibratory roller in vibration compaction[J]. Advances in Mechanical Engineering,2018,10(5):1-16.
[8] 邵俊愷,趙?翾,楊?玨,等. 無人駕駛鉸接式車輛強(qiáng)化學(xué)習(xí)路徑跟蹤控制算法[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2017,48(3):376-382.
Shao Junkai,Zhao Xuan,Yang Jue,et al. Reinforcement learning algorithm for path following control of articulated vehicle[J]. Transactions of the Chinese Society of Agricultural Machinery,2017,48(3):376-382(in Chinese).
[9] Khalaji A K,Jalalnezhad M. Control of a tractor-trailer robot subjected to wheel slip[J]. Proc IMechE Part K:J Multi-Body Dynamics,2019,233(4):956-967.
[10] Erkan K,Wouter S,Herman R,et al. Experimental validation of linear and nonlinear MPC on an articulated unmanned ground vehicle[J]. Transactions on Mechatronics,IEEE/ASME,2018,23(5):2023-2030.
[11] Nayl T,Nikolakopoulos G,Gustafsson T,et al. Design and experimental evaluation of a novel sliding mode controller for an articulated vehicle[J]. Robotics and Autonomous Systems,2018,103:213-221.
[12] Yue M,Wu X,Guo L,et al. Quintic polynomial-based obstacle avoidance trajectory planning and tracking control framework for tractor-trailer system[J]. International Journal of Control Automation and Systems,2019,17(3):2634-2646.
[13] Bian Yongming,Yang Meng,F(xiàn)ang Xiaojun,et al. Kinematics and path following control of an articulated drum roller[J]. Chinese Journal of Mechanical Engineering,2017,30(4):888-899.
[14] Xu Quanzhi,Song Kang,Xie Hui. The impact of control structure on the path-following control of unmanned compaction rollers[C]//SAE 2019 Intelligent and Connected Vehicles Symposium. Paris,F(xiàn)rance,2020:2020-01-5030.
[15] Nayl T,Nikolakopoulos G,Gustafsson T. Switching model predictive control for an articulated vehicle under varying slip angle[C]// 2012 20th Mediterranean Conference on Control & Automation. Barcelona,Spain,2012:884-889.
[16] Yang M,Bian Y M,Liu G J,et al. Path tracking control of an articulated road roller with sideslip compensation[J/OL]. IEEE Access,DOI:10.1109/ACCESS.2020. 3008455,2020-09-20.
[17] Corke P,Ridley P. Load haul dump vehicle kine-matics and control[J]. Journal of Dynamic Systems Measurement & Control,2003,125(1):54-59.
[18] Iida M,F(xiàn)ukuta M,Tomiyama H. Measurement and analysis of side-slip angle for an articulated vehicle[J]. Engineering in Agriculture Environment & Food,2010,3(1):1-6.
[19] 謝?輝,趙龍同,阮迪望. 智能振動碾壓機(jī)的自抗擾循跡控制方法[J]. 天津大學(xué)學(xué)報(bào):自然科學(xué)與工程技術(shù)版,2020,53(9):900-909.
Xie Hui,Zhao Longtong,Ruan Diwang. Path following control method with active disturbance rejection for an intelligent vibration roller[J]. Journal of Tianjin University:Science and Technology,2020,53(9):900-909(in Chinese).
[20] 韓京清. 從PID技術(shù)到“自抗擾控制”技術(shù)[J]. 控制工程,2002(3):13-18.
Han Jingqing. From PID technique to active disturbances rejection control technique[J]. Control Engineering of China,2002(3):13-18(in Chinese).
[21] 孟?宇,甘?鑫,白國星. 基于預(yù)瞄距離的地下礦用鉸接車路徑跟蹤預(yù)測控制[J]. 工程科學(xué)學(xué)報(bào),2019,41(5):662-671.
Meng Yu,Gan Xin,Bai Guoxing. Path following control of underground mining articulated vehicle based on the preview control method[J]. Chinese Journal of Engineering,2019,41(5):662-671(in Chinese).
Path-Following Algorithm and Energy Consumption Law of Unmanned Roller
Du Xu,Song Kang,Xie Hui
(School of Mechanical Engineering,Tianjin University,Tianjin 300072,China)
The unmanned roller is an important technique to reduce the burden of manual work and improve the quality and efficiency of rolling operation. However,because of the high tonnage,high steering resistance,and frequent working on complex unstructured road surfaces,the energy consumption of the conventional vehicles’path-following algorithms is usually high,which affects the power balance,reliability,and comprehensive energy efficiency of the control system. Hence,an energy-saving cascade disturbance rejection path-following algorithm is proposed for an unmanned roller running on a hydropower dam. At the same time,the influence of the algorithm on the energy consumption of the steering system of the unmanned roller is studied. First,a position preview algorithm is used to control the steering system in advance to reduce the overshoot caused by the heavy load and to reduce energy consumption. Second,a vehicle body attitude correction algorithm is proposed to resolve the problem of disturbances in the vehicle caused by road surface undulation. By measuring the roll angle of the front and rear body of the unmanned roller compactor,the positioning results are corrected to reduce the noise interference and energy consumption. Finally,to reduce the high-frequency fluctuation of the steering wheel in the small range error band,a nonlinear error feedback control law is used to save energy in the steering system. A real vehicle test was conducted at the hydropower dam’s construction site. The results show that:①Compared with the nominal controller without preview,the optimized preview distance can reduce the energy consumption of the steering motor by 24.7% and make the path-following accuracy stable at ±0.15m;②Using the attitude correction algorithm,the energy consumption of the steering motor can be reduced by 29.2% and the path-following accuracy can be improved by 33.0%;③Using nonlinear error feedback control law,the energy consumption of the steering motor can be reduced by 31.8% without the usual deterioration of distance error(within 0.06m).
unmanned roller;path-following;energy consumption optimization control;preview algorithm;attitude correction
TP242.6
A
0493-2137(2021)08-0834-10
10.11784/tdxbz202009082
2020-09-28;
2020-12-09.
杜?續(xù)(1994—??),男,碩士研究生,arcgerald@163.com.
宋?康,songkangtju@tju.edu.cn.
天津市人工智能科技重大專項(xiàng)資助項(xiàng)目(19ZXZNGX00050).
Supported by the Science and Technology Major Project on Artificial Intelligence of Tianjin,China(No. 19ZXZNGX00050).
(責(zé)任編輯:金順愛)