張聞?dòng)?，張智剛，?帆，丁 凡，胡 煉，羅錫文

·農(nóng)業(yè)裝備工程與機(jī)械化·

水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)策略與試驗(yàn)

張聞?dòng)?,2,3，張智剛1,2,3，張 帆1,3，丁 凡1,3，胡 煉1,2,3，羅錫文1,2,3※

（1. 華南農(nóng)業(yè)大學(xué)南方農(nóng)業(yè)機(jī)械與裝備關(guān)鍵技術(shù)教育部重點(diǎn)實(shí)驗(yàn)室，廣州 510642；2. 嶺南現(xiàn)代農(nóng)業(yè)科學(xué)與技術(shù)廣東省實(shí)驗(yàn)室，廣州 510642；3. 廣東省農(nóng)業(yè)人工智能重點(diǎn)實(shí)驗(yàn)室，廣州 510642）

針對水稻收獲機(jī)與轉(zhuǎn)運(yùn)車雙機(jī)協(xié)同自主作業(yè)環(huán)節(jié)多、糧食轉(zhuǎn)運(yùn)過程復(fù)雜等問題，該研究設(shè)計(jì)了一種基于有限狀態(tài)機(jī)（Finite State Machine，F(xiàn)SM）的水稻收獲機(jī)與轉(zhuǎn)運(yùn)車協(xié)同作業(yè)策略，分析了水稻收獲機(jī)與轉(zhuǎn)運(yùn)車協(xié)同作業(yè)模式，建立有限狀態(tài)機(jī)模型。首先，基于作業(yè)環(huán)節(jié)設(shè)計(jì)觸發(fā)條件、評估方法和執(zhí)行流程等基礎(chǔ)模塊；然后，根據(jù)雙機(jī)協(xié)同的各項(xiàng)狀態(tài)建立狀態(tài)信息矩陣；最后，依據(jù)協(xié)同觸發(fā)事件與狀態(tài)轉(zhuǎn)移的邏輯設(shè)計(jì)狀態(tài)轉(zhuǎn)移鏈。構(gòu)建協(xié)同作業(yè)時(shí)分復(fù)用控制邏輯框架，并運(yùn)用Stateflow軟件進(jìn)行仿真分析，為驗(yàn)證所設(shè)計(jì)策略的田間實(shí)際作業(yè)效果，搭建了履帶式水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同試驗(yàn)系統(tǒng)，收獲速度為0.8 m/s，收割幅寬1.9 m，共28條收獲邊，協(xié)同路徑選擇在短邊的機(jī)耕道上，連續(xù)協(xié)同工作時(shí)間大于等于120 min，采用套圈路徑自主收獲0.7 hm2水稻，期間共進(jìn)行6次自動(dòng)協(xié)同轉(zhuǎn)運(yùn)作業(yè)，將所收獲的糧食轉(zhuǎn)運(yùn)到卡車上。試驗(yàn)結(jié)果表明，該策略可以實(shí)現(xiàn)水稻收獲/卸糧轉(zhuǎn)運(yùn)自主作業(yè)，收獲效率為0.35 hm2/h，為實(shí)現(xiàn)水稻收獲雙機(jī)智能轉(zhuǎn)運(yùn)協(xié)同功能奠定了基礎(chǔ)，可為水稻無人農(nóng)場建設(shè)提供技術(shù)支持。

農(nóng)業(yè)機(jī)械；水稻；收獲；無人農(nóng)場；轉(zhuǎn)運(yùn)；雙機(jī)協(xié)同策略

0 引 言

水稻是主要糧食作物，保證水稻高產(chǎn)穩(wěn)產(chǎn)保障國家糧食安全有重要意義。智能化機(jī)械生產(chǎn)能夠減輕農(nóng)業(yè)勞動(dòng)強(qiáng)度、提高生產(chǎn)效率和質(zhì)量[1-2]。水稻智能化收獲作業(yè)復(fù)雜程度相對較高，涉及到定位測姿、路徑跟蹤、收獲控制和協(xié)同轉(zhuǎn)運(yùn)等多個(gè)部分[3-5]。路徑規(guī)劃[6-7]和路徑跟蹤[8]是智能農(nóng)機(jī)共性技術(shù)，國內(nèi)外學(xué)者對農(nóng)機(jī)直線跟蹤進(jìn)行了一系列研究，包括自適應(yīng)純追蹤路徑跟蹤方法[9-11]、模糊控制方法[12]、復(fù)合循環(huán)魯棒控制[13]、防側(cè)翻適應(yīng)控制算法[14]、虛擬阿克曼轉(zhuǎn)向模型跟蹤算法[15]和預(yù)瞄追蹤模型算法[16]，跟蹤精度可達(dá)2.5 cm。

水稻收獲時(shí)，由于收獲機(jī)糧倉容量有限，收獲機(jī)糧倉快滿時(shí)需開至田邊卸糧，影響了收獲機(jī)自動(dòng)作業(yè)效率。因此，研究水稻收獲機(jī)與轉(zhuǎn)運(yùn)車的協(xié)同作業(yè)系統(tǒng)，采用智能轉(zhuǎn)運(yùn)車轉(zhuǎn)運(yùn)糧食可以有效提高收獲機(jī)的作業(yè)效率。

國內(nèi)外研究者對農(nóng)機(jī)協(xié)同作業(yè)進(jìn)行了相關(guān)研究。Noguchi等[17]提出了2種典型農(nóng)機(jī)協(xié)同模式，一種是主機(jī)指定給從機(jī)作業(yè)目標(biāo)，一是從機(jī)跟隨主機(jī)作業(yè)。Iida等[18]研發(fā)了一農(nóng)機(jī)自動(dòng)跟隨系統(tǒng)，采用超聲波傳感器控制隨車的相對位置和偏角。白曉平等[19]提出了一種聯(lián)合收獲機(jī)群協(xié)同作業(yè)的控制方法，在1.0 m/s的速度下，車輛跟隨平均跟蹤誤差為5.93 cm。張聞?dòng)畹萚20]設(shè)計(jì)了縱向相對位置位速耦合控制器，實(shí)現(xiàn)了從機(jī)跟隨卸糧，在1.0 m/s的速度下，收獲機(jī)與運(yùn)糧車相對縱向距離控制誤差標(biāo)準(zhǔn)差為9.2 cm。王猛[21]建立了多農(nóng)機(jī)協(xié)同作業(yè)靜態(tài)任務(wù)分配模型，使作業(yè)成本降低了29.48%～55.00%。

曹如月等[22]提出了基于改進(jìn)A*和Bezier曲線的多機(jī)協(xié)同全局路徑規(guī)劃，有效降低了轉(zhuǎn)彎次數(shù)。宮金良等[23]提出一種復(fù)雜環(huán)境下異質(zhì)農(nóng)業(yè)機(jī)器人群的任務(wù)分配及全區(qū)域覆蓋策略，4個(gè)試驗(yàn)機(jī)器人可以共同完成場地的覆蓋，遍歷重復(fù)率小于7%。姚竟發(fā)等[24]針對同種機(jī)型的協(xié)同進(jìn)行了路徑規(guī)劃研究，以總作業(yè)時(shí)間和作業(yè)時(shí)長為綜合優(yōu)化目標(biāo)，進(jìn)行多機(jī)無沖突協(xié)同作業(yè)路徑優(yōu)化，矩形農(nóng)田總作業(yè)時(shí)間和作業(yè)時(shí)長平均分別下降了2.45%、2.29%。翟志強(qiáng)等[25]提出一種面向主從跟隨模式的多機(jī)協(xié)同作業(yè)導(dǎo)航路徑規(guī)劃方法，基于方向包圍盒算法和分離軸定理建立農(nóng)機(jī)安全狀態(tài)檢測模型，路徑規(guī)劃算法用時(shí)最小值為0.453 s、最大值為1.563 s、平均值為0.951 s，可為小麥、青貯收獲等主從跟隨式多機(jī)協(xié)同自主作業(yè)提供有效的全局作業(yè)路徑。上述研究集中在弱耦合的多機(jī)空間協(xié)同路徑規(guī)劃，減少了作業(yè)時(shí)間和提高了覆蓋率。多智能體協(xié)同在機(jī)器人領(lǐng)域也是研究的熱點(diǎn)之一[26-32]，對農(nóng)機(jī)協(xié)同作業(yè)有借鑒作用。

上述協(xié)同研究大多集中于多機(jī)協(xié)同的協(xié)同精度控制方法和路徑規(guī)劃等，而農(nóng)機(jī)協(xié)同作業(yè)系統(tǒng)中系統(tǒng)邏輯框架和協(xié)同控制策略是保證系統(tǒng)可靠性的關(guān)鍵。雙機(jī)協(xié)同作業(yè)模式具有環(huán)節(jié)多、交互復(fù)雜等非順序性特點(diǎn)，本文擬根據(jù)水稻收獲雙機(jī)協(xié)同作業(yè)需求，分析協(xié)同收獲模式，建立系統(tǒng)有限狀態(tài)機(jī)模型，基于該模型構(gòu)建時(shí)分復(fù)用系統(tǒng)邏輯框架，并進(jìn)行協(xié)同邏輯仿真驗(yàn)證和田間試驗(yàn)，以驗(yàn)證水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)策略的可靠性和實(shí)用性。

1 水稻收獲智能轉(zhuǎn)運(yùn)雙機(jī)協(xié)同模式

Noguchi等[17]將農(nóng)機(jī)協(xié)同模式分為“GOTO”（定點(diǎn)協(xié)同）和“FOLLOW”（跟隨協(xié)同）。水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)也可以按照糧食轉(zhuǎn)運(yùn)過程分為2種模式，模式1為召喚轉(zhuǎn)運(yùn)模式，模式2為隨車轉(zhuǎn)運(yùn)模式。模式1是全程智能收獲作業(yè)的基礎(chǔ)，本文針對模式1多環(huán)節(jié)協(xié)同作業(yè)和控制的特點(diǎn)，并將模式1分為4個(gè)基本環(huán)節(jié)（Basic Elements，BE）的循環(huán)操作：

收獲機(jī)自主收獲環(huán)節(jié)1：收獲機(jī)提高發(fā)動(dòng)機(jī)轉(zhuǎn)速、啟動(dòng)主離合和割臺離合，跟蹤預(yù)設(shè)收獲路徑，實(shí)現(xiàn)自主收獲功能；

轉(zhuǎn)運(yùn)車等待召喚環(huán)節(jié)2：轉(zhuǎn)運(yùn)車根據(jù)規(guī)劃的路徑行駛到待協(xié)同路段，停車、等待收獲機(jī)召喚。

收獲機(jī)向轉(zhuǎn)運(yùn)車卸糧環(huán)節(jié)3：收獲機(jī)到達(dá)待卸糧位，發(fā)送信號召喚轉(zhuǎn)運(yùn)車，轉(zhuǎn)運(yùn)車根據(jù)通訊信息實(shí)現(xiàn)召喚對位協(xié)同卸糧；

轉(zhuǎn)運(yùn)車卸糧環(huán)節(jié)4：轉(zhuǎn)運(yùn)車根據(jù)規(guī)劃的路徑駛往運(yùn)糧卡車卸糧。

1、2和4為弱耦合環(huán)節(jié)，兩車保持通訊各自完成自主規(guī)劃作業(yè)任務(wù)。3為強(qiáng)耦合環(huán)節(jié)，收獲機(jī)和轉(zhuǎn)運(yùn)車必須實(shí)現(xiàn)精準(zhǔn)定位和協(xié)同作業(yè)。

針對水稻自主收獲運(yùn)糧協(xié)同作業(yè)的要求，制定智能協(xié)同流程，收獲機(jī)首先根據(jù)規(guī)劃路徑出庫下田，執(zhí)行1自主收獲，轉(zhuǎn)運(yùn)車隨后出庫，進(jìn)入2在機(jī)耕道等待召喚，收獲機(jī)基于糧倉狀態(tài)停車召喚轉(zhuǎn)運(yùn)車，進(jìn)入3協(xié)同卸糧，協(xié)同卸糧完成后收獲機(jī)回到1繼續(xù)收獲，轉(zhuǎn)運(yùn)車行駛至田邊道路進(jìn)入4，完成卸糧后返回機(jī)耕道進(jìn)入2等待收獲機(jī)再次召喚。全部收獲工作結(jié)束后收獲機(jī)執(zhí)行3將糧倉中的糧食卸至轉(zhuǎn)運(yùn)車中，然后自動(dòng)駛回機(jī)庫，轉(zhuǎn)運(yùn)車完成B后回庫，最終完成智能協(xié)同作業(yè)。

2 收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同過程建模

收獲機(jī)與轉(zhuǎn)運(yùn)車智能協(xié)同作業(yè)包括自主作業(yè)和協(xié)同轉(zhuǎn)運(yùn)幾個(gè)環(huán)節(jié)，本文根據(jù)協(xié)同模式將各個(gè)動(dòng)作分為若干子狀態(tài)，采用有限狀態(tài)機(jī)（Finite State Machine，F(xiàn)SM）進(jìn)行建模，有限狀態(tài)機(jī)能夠有效表示系統(tǒng)的各個(gè)狀態(tài)和狀態(tài)間轉(zhuǎn)移關(guān)系，在商品生產(chǎn)和航天航空等領(lǐng)域廣泛應(yīng)用[33-34]。

根據(jù)收獲機(jī)和轉(zhuǎn)運(yùn)車2個(gè)實(shí)體將系統(tǒng)分為2個(gè)模塊狀態(tài)機(jī)（Module State Machine，MSM），收獲機(jī)為1，轉(zhuǎn)運(yùn)車為2。

2.1 狀態(tài)機(jī)功能模塊定義

收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同模式狀態(tài)機(jī)的組成部分包括：

1）觸發(fā)器（State Generator，）。觸發(fā)器為狀態(tài)的外部觸發(fā)輸入，功能是觸發(fā)狀態(tài)轉(zhuǎn)移。1的狀態(tài)觸發(fā)器包括：轉(zhuǎn)運(yùn)車對位完成信號G1，轉(zhuǎn)運(yùn)車駛離信號G2。2的狀態(tài)觸發(fā)器包括收獲機(jī)就位信號G1，收獲機(jī)卸糧結(jié)束信號G2。

2）評估器（State Assessor，）。評估器用于評估是否能進(jìn)入狀態(tài)。1的狀態(tài)評估包括收獲機(jī)糧倉滿評估A1，收獲機(jī)糧倉空評估A2，收獲機(jī)直線跟蹤質(zhì)量評估A3，收獲機(jī)目標(biāo)路徑狀態(tài)評估A4；2的狀態(tài)評估包括轉(zhuǎn)運(yùn)車糧倉滿評估A1，轉(zhuǎn)運(yùn)車糧倉空評估A2，轉(zhuǎn)運(yùn)車直線跟蹤質(zhì)量評估A3，轉(zhuǎn)運(yùn)車目標(biāo)路徑狀態(tài)評估A4。

3）執(zhí)行器（State Executor，）。執(zhí)行器為當(dāng)前狀態(tài)的執(zhí)行任務(wù)，1的狀態(tài)執(zhí)行器包括收獲狀態(tài)執(zhí)行器E1、召喚狀態(tài)執(zhí)行器E2和卸糧狀態(tài)執(zhí)行器E3。2的狀態(tài)執(zhí)行器包括對位狀態(tài)執(zhí)行器E1，等待狀態(tài)執(zhí)行器E2和卸車狀態(tài)執(zhí)行器E3。

4）確認(rèn)器（State Verification，）。確認(rèn)器用于確認(rèn)當(dāng)前狀態(tài)的結(jié)果，1的狀態(tài)確認(rèn)器包括：收獲機(jī)卸糧點(diǎn)就位確認(rèn)V1、收獲機(jī)卸糧完畢確認(rèn)V2。2的狀態(tài)確認(rèn)器包括轉(zhuǎn)運(yùn)車與收獲機(jī)對位成功確認(rèn)V1、轉(zhuǎn)運(yùn)車與運(yùn)糧卡車對位成功確認(rèn)V2。

2.2 狀態(tài)信息矩陣定義

分別將收獲機(jī)1和轉(zhuǎn)運(yùn)車2的觸發(fā)器、評估器、執(zhí)行器和確認(rèn)器各部件組成系統(tǒng)狀態(tài)矩陣，收獲機(jī)狀態(tài)矩陣為1、轉(zhuǎn)運(yùn)車狀態(tài)矩陣為2，1包括收獲路徑和糧倉監(jiān)控等信息，收獲機(jī)的狀態(tài)1表達(dá)式如下：

式中11為路徑狀態(tài)機(jī)，12為主離合狀態(tài)機(jī)，13為割臺離合狀態(tài)，14為卸糧離合狀態(tài)，15為糧倉監(jiān)測狀態(tài)，16為糧筒位置狀態(tài)，17為定位定向狀態(tài)，18為雙機(jī)對位狀態(tài)，19為發(fā)動(dòng)機(jī)轉(zhuǎn)速狀態(tài)，110為車速狀態(tài)，為收獲機(jī)狀態(tài)個(gè)數(shù)。

2包括轉(zhuǎn)運(yùn)路徑和糧倉監(jiān)控等信息，表達(dá)式如下：

式中21為路徑狀態(tài)機(jī)，22為卸糧離合狀態(tài)，23為糧倉監(jiān)測狀態(tài)，24為糧筒位置狀態(tài)，S為定位定向狀態(tài)，26為運(yùn)移對位狀態(tài)，27為發(fā)動(dòng)機(jī)轉(zhuǎn)速狀態(tài)，28為車速狀態(tài)，為轉(zhuǎn)運(yùn)車狀態(tài)個(gè)數(shù)。

上述每個(gè)狀態(tài)都涉及觸發(fā)器、評估器、執(zhí)行器、確認(rèn)器，集合關(guān)系式表達(dá)式為

11需要評估收獲機(jī)是否在正常直線跟蹤狀態(tài)；收獲作業(yè)時(shí)12、13、19和110按順序結(jié)合主離合和割臺離合加油門起步，停車等待時(shí)按則逆序操作。卸糧時(shí)14、15和16基于糧倉的狀態(tài)先伸出糧筒、提高油門、結(jié)合卸糧離合，停止卸糧則逆序操作，18和19進(jìn)行對位過程執(zhí)行和評估。轉(zhuǎn)運(yùn)車除了沒有主離合和割臺離合狀態(tài)機(jī)，其他的狀態(tài)機(jī)功能與收獲機(jī)相同。收獲機(jī)和轉(zhuǎn)運(yùn)車狀態(tài)集合如表1。

表1 收獲機(jī)和轉(zhuǎn)運(yùn)車狀態(tài)集合表Table 1 State parameter table of harvester and transport vehicle

注：E1為收獲狀態(tài)執(zhí)行器，E2為召喚轉(zhuǎn)運(yùn)車狀態(tài)執(zhí)行器，E3為卸糧狀態(tài)執(zhí)行器，A1為收獲機(jī)糧倉滿評估器，A2為收獲機(jī)糧倉空評估器，A3為收獲機(jī)直線跟蹤質(zhì)量評估器，V1為收獲機(jī)卸糧點(diǎn)就位確認(rèn)器，V2為收獲機(jī)卸糧完畢確認(rèn)器，G1為轉(zhuǎn)運(yùn)車與收獲機(jī)對位完成信號，G2為轉(zhuǎn)運(yùn)車駛離信號，E1為對位狀態(tài)執(zhí)行器，E2為等待狀態(tài)執(zhí)行器，E3為轉(zhuǎn)運(yùn)車卸糧狀態(tài)執(zhí)行器，A1為轉(zhuǎn)運(yùn)車糧倉滿評估器，A2為轉(zhuǎn)運(yùn)車糧倉空評估器，A3為轉(zhuǎn)運(yùn)車直線跟蹤質(zhì)量評估器，V1為運(yùn)糧車與收獲機(jī)對位成功確認(rèn)器，V2為轉(zhuǎn)運(yùn)車與運(yùn)糧卡車對位成功確認(rèn)器，G1為收獲機(jī)就位信號，G2為收獲機(jī)卸糧結(jié)束信號。

Note:E1is the state executor of harvest,E2is the state executor of calling transport,E3is the state executor of grain unloading,A1is the state assessor of harvester granary full,A2is the state assessor of harvester granary empty,A3is the state assessor of harvester line tracking quality,V1is the state verification of harvester in place unloading point,V2is the state verification of harvester finished grainunloading,G1is the state generator of transportalignment with harvest,G2is the state generator of transport leaving,E1is the state executor of alignment,E2is the state executor of waiting,E3is the state executor of the transport grain unloading,A1is the state assessor of transport granary full,A2is the state assessor of transport granary empty,A3is the state assessor of transport line tracking quality,V1is the state verification of alignment transport with harvest,V2is the state verification of alignment transport with grain transporting truck,G1is the state generator of harvester in place,G2is the state generator of harvest unloading over.

2.3 基本動(dòng)作執(zhí)行器

根據(jù)水稻收獲轉(zhuǎn)運(yùn)協(xié)同作業(yè)過程中的操作，可以分為固定的執(zhí)行流程：E1、E2、E3和E1、E2、E3。

E1順序改變18、13、12、110和11的狀態(tài)，使得收獲機(jī)從停車模式進(jìn)入收獲1。

E2順序改變110、18、12和13的狀態(tài)，使得收獲機(jī)從收獲模式進(jìn)入召喚等待，召喚轉(zhuǎn)運(yùn)車前來運(yùn)糧。

E3分2組動(dòng)作，E3_1開始卸糧和E3_2停止卸糧操作。

E1分2類功能，路徑跟蹤運(yùn)移和協(xié)同對位。路徑跟蹤運(yùn)移E1_1，沿21的規(guī)劃路徑進(jìn)行路徑運(yùn)移。協(xié)同對位E1_2進(jìn)行對位操作，E2使車輛從運(yùn)移模式進(jìn)入2。由于2卸糧系統(tǒng)與1相同，E3的功能和動(dòng)作與E3相同，分為E3_1和E3_2。

2.4 協(xié)同作業(yè)狀態(tài)轉(zhuǎn)移鏈構(gòu)建

根據(jù)4個(gè)基本環(huán)節(jié)設(shè)計(jì)狀態(tài)機(jī)的狀態(tài)轉(zhuǎn)移鏈。1環(huán)節(jié)主要由執(zhí)行器E1組成，2環(huán)節(jié)主要由執(zhí)行器E2組成，3環(huán)節(jié)主要由執(zhí)行器E2、E3和E1組成，4環(huán)節(jié)主要E3組成。1、2和4環(huán)節(jié)相對獨(dú)立，3為強(qiáng)耦合環(huán)節(jié)。

當(dāng)收獲機(jī)1處于E1收獲狀態(tài)，A1評估如糧倉中的水稻儲量達(dá)到閾值，1則按照11的規(guī)劃在預(yù)定路徑點(diǎn)上進(jìn)入E2等待狀態(tài)。同時(shí)V2確認(rèn)到達(dá)指定位置，通過無線網(wǎng)絡(luò)發(fā)送G1信號標(biāo)志位至2轉(zhuǎn)運(yùn)車。如果2此時(shí)處于E2狀態(tài)則進(jìn)入E1執(zhí)行，如果處于E3狀態(tài)則完成卸糧后前往協(xié)同地點(diǎn)后經(jīng)由E2轉(zhuǎn)入E1執(zhí)行。2執(zhí)行E1，完成后，V1確認(rèn)成功后發(fā)送G1觸發(fā)1的E3卸糧執(zhí)行。A2監(jiān)測到如糧倉已經(jīng)卸空則發(fā)送G2信號給2，接著1結(jié)束卸糧執(zhí)行E1。此時(shí)2根據(jù)A2評估結(jié)果，如倉儲達(dá)到閾值則進(jìn)入E3卸糧流程，否則重新進(jìn)入E2繼續(xù)等待。E3依據(jù)A2判斷糧空，然后依據(jù)路徑規(guī)劃到達(dá)協(xié)同位置進(jìn)入E2。

依據(jù)上述流程設(shè)計(jì)水稻自動(dòng)收獲/卸糧轉(zhuǎn)運(yùn)雙機(jī)協(xié)同的狀態(tài)轉(zhuǎn)移鏈，如圖1所示。

注：M1為收獲機(jī)狀態(tài)機(jī)，M2為轉(zhuǎn)運(yùn)機(jī)狀態(tài)機(jī)。

3 水稻自動(dòng)收獲/卸糧轉(zhuǎn)運(yùn)雙機(jī)協(xié)同作業(yè)邏輯

3.1 邏輯框架設(shè)計(jì)

基于上文的雙機(jī)協(xié)同模型、狀態(tài)參數(shù)和狀態(tài)轉(zhuǎn)移鏈分別設(shè)計(jì)收獲機(jī)和轉(zhuǎn)運(yùn)車的邏輯框架。

雙機(jī)協(xié)同轉(zhuǎn)運(yùn)邏輯框架是雙機(jī)協(xié)同模式從理論模型到技術(shù)實(shí)現(xiàn)的橋梁，其將路徑跟蹤控制系統(tǒng)、縱向?qū)ξ豢刂扑惴?、作業(yè)動(dòng)作流程控制模塊和自動(dòng)卸糧模塊等具體的技術(shù)細(xì)節(jié)進(jìn)行有效整合，使得收獲運(yùn)糧系統(tǒng)能夠在自主控制條件下完成水稻收獲轉(zhuǎn)運(yùn)作業(yè)。

嵌入式導(dǎo)航駕駛控制器中同時(shí)有多個(gè)不同任務(wù)在運(yùn)行，所以設(shè)計(jì)協(xié)同邏輯框架時(shí)采用時(shí)分復(fù)用模式，以主循環(huán)重復(fù)檢索的方式為協(xié)同框架，收獲機(jī)協(xié)同邏輯流程如圖2a所示，轉(zhuǎn)運(yùn)車協(xié)同邏輯流程如圖2b所示，該流程是由雙機(jī)通訊所驅(qū)動(dòng)的狀態(tài)轉(zhuǎn)移鏈流程，2組邏輯通過觸發(fā)信號相互配合運(yùn)行，實(shí)現(xiàn)水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)的功能。

3.2 協(xié)同邏輯時(shí)序仿真分析

仿真分析主要研究雙機(jī)信號傳遞與協(xié)同邏輯框架的有效性，針對雙機(jī)的順序執(zhí)行器進(jìn)行了簡化，引入延時(shí)替換車輛內(nèi)部順序執(zhí)行器。采用MATLAB中的Stateflow建立仿真模型[34]?；趨f(xié)同邏輯，設(shè)計(jì)收獲機(jī)和轉(zhuǎn)運(yùn)車對應(yīng)的Stateflow模型，分別為Harvester模塊和Transfer模塊。其中Output和Input分別代表模塊當(dāng)前狀態(tài)的輸出和下一時(shí)刻的狀態(tài)輸入，中間的延時(shí)函數(shù)Delay、Delay1代表執(zhí)行的步驟所用的時(shí)間。2個(gè)模塊之間通信信號包括G1、G1和G2觸發(fā)器，傳遞方向如圖3a所示，信號間的Delay2、Delay3、Delay4代表通訊延時(shí)。其中State表示農(nóng)機(jī)當(dāng)前狀態(tài)，其他符號與上文定義一致。State的狀態(tài)初始值都設(shè)為1，G1、G1和G2的信號初值都設(shè)為0，時(shí)序仿真結(jié)果如圖 3b 和圖3c所示，圖中收獲機(jī)和轉(zhuǎn)運(yùn)車的狀態(tài)按照設(shè)計(jì)在1至4之間有序變化，如圖3b1和2進(jìn)入1和2，當(dāng)1進(jìn)入3，2進(jìn)入3，3結(jié)束后，2進(jìn)入4，1重回1，2結(jié)束4回到2，進(jìn)入下一個(gè)循環(huán)。如圖3c1進(jìn)入3時(shí)發(fā)出G1信號召喚2，2就位后發(fā)出G1信號，卸糧結(jié)束1發(fā)出G2信號，進(jìn)入下一個(gè)循環(huán)。試驗(yàn)結(jié)果表明收獲轉(zhuǎn)運(yùn)流程邏輯框架有效。

圖2 雙機(jī)協(xié)同邏輯流程

注：收獲機(jī)的狀態(tài)1-4分別為Eh1、Eh2、Eh3_1和Eh3_2執(zhí)行狀態(tài)，轉(zhuǎn)運(yùn)車的狀態(tài)1～6分別為Et1_1、Et2、Et1_2、Et3、Et3_1和Et3_2執(zhí)行狀態(tài)，輸入1、輸入2分別為收獲機(jī)和轉(zhuǎn)運(yùn)車的輸入狀態(tài)，輸出1、輸出2分別為收獲機(jī)和轉(zhuǎn)運(yùn)車的輸出狀態(tài)。

4 水稻自動(dòng)收獲/卸糧轉(zhuǎn)運(yùn)雙機(jī)協(xié)同作業(yè)田間試驗(yàn)

4.1 雙機(jī)協(xié)同導(dǎo)航試驗(yàn)系統(tǒng)

為驗(yàn)證水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)策略的實(shí)際作業(yè)效果和穩(wěn)定性，構(gòu)建了履帶式水稻協(xié)同收獲系統(tǒng)。收獲協(xié)同系統(tǒng)包括履帶式水稻收獲機(jī)（濰柴雷沃重工RG70V4G-014）和履帶式水稻轉(zhuǎn)運(yùn)車（濰柴雷沃重工RG70V4G-015），兩機(jī)均采用全電控的底盤，主要參數(shù)如表2所示，具有線控的離合、割臺、糧筒和履帶行駛系統(tǒng)，采用CAN（Controller Area Network）總線發(fā)送線控協(xié)議報(bào)文。雙機(jī)均安裝雙天線BDS定位系統(tǒng)（司南K726），定位信息獲取頻率為10 Hz，水平定位精度為±（10+1×10-6×）mm，其中為基站到移動(dòng)站的距離(km)。水稻收獲協(xié)同試驗(yàn)的雙機(jī)通訊采用有人云的4/5G DTU(Data Transfer unit)，型號為USR-G781，雙工通訊串口透傳頻率為1 Hz，采用16位CRC（Cyclic Redundancy Check）校驗(yàn)設(shè)計(jì)串口通訊協(xié)議減少通訊誤碼率。導(dǎo)航控制模塊與控制終端通過RS-232通訊；雙機(jī)分別安裝了導(dǎo)航控制終端（eAgri-800-RS）用于收獲智能控制，通過CAN總線與雙機(jī)的底盤電控單元（Electronic Control Unit，ECU）通訊。雙機(jī)協(xié)同控制系統(tǒng)基于c語言運(yùn)用Keil uVision5開發(fā)，協(xié)同策略邏輯嵌入其中，架構(gòu)形式與文中邏輯流程一致，以雙機(jī)之間通訊驅(qū)動(dòng)狀態(tài)轉(zhuǎn)移。導(dǎo)航系統(tǒng)直線路徑跟蹤算法采用預(yù)瞄跟隨控制方法[16]跟蹤標(biāo)準(zhǔn)差小于5 cm，雙機(jī)對位控制算法采用位置誤差PID（Proportion Integration Differentiation）方法[20]，縱向誤差小于20 cm，控制系統(tǒng)結(jié)構(gòu)和試驗(yàn)車輛如圖4所示。

表2 收獲機(jī)和轉(zhuǎn)運(yùn)車主要結(jié)構(gòu)參數(shù)Table 2 Main structural parameters of harvester and transport vehicle

注：CAN為控制器局域網(wǎng)，RTK-GNSS為全球?qū)Ш叫l(wèi)星定位系統(tǒng)實(shí)時(shí)動(dòng)態(tài)測量技術(shù)。

4.2 水稻自動(dòng)收獲/卸糧轉(zhuǎn)運(yùn)雙機(jī)協(xié)同作業(yè)試驗(yàn)設(shè)計(jì)與分析

試驗(yàn)設(shè)計(jì)：協(xié)同作業(yè)試驗(yàn)在華南農(nóng)業(yè)大學(xué)增城試驗(yàn)基地進(jìn)行。試驗(yàn)地長120 m，寬60 m，約0.7 hm2。由于本試驗(yàn)中沒有安裝糧倉位傳感器，所以策略中的倉滿、倉空采用產(chǎn)量預(yù)估和定時(shí)器計(jì)算確定，及時(shí)性和準(zhǔn)確性不高，通過仿真驗(yàn)證符合邏輯運(yùn)行需求。由于轉(zhuǎn)運(yùn)邊是固定在機(jī)耕道上，必須完成收獲一整圈才能卸糧轉(zhuǎn)運(yùn)?？梢酝ㄟ^試驗(yàn)前一次長邊的水稻收獲量估算需要2圈進(jìn)行轉(zhuǎn)運(yùn)。實(shí)際作業(yè)中一條長邊作業(yè)大約可以收獲20%糧倉容積的糧食，所以估算收獲機(jī)快完成2圈作業(yè)時(shí)（4條長邊收獲）提前通知轉(zhuǎn)運(yùn)車做好轉(zhuǎn)運(yùn)準(zhǔn)備。設(shè)計(jì)收獲機(jī)和轉(zhuǎn)運(yùn)車先后從基地機(jī)庫出發(fā)自主前往田間，收獲機(jī)到田間后啟動(dòng)收獲流程開始收獲，收獲速度設(shè)置為0.8 m/s。轉(zhuǎn)運(yùn)車在機(jī)耕道等待，按照所設(shè)計(jì)的協(xié)同作業(yè)策略，將收獲的糧食轉(zhuǎn)運(yùn)到路邊的運(yùn)糧卡車中。整塊農(nóng)田收獲完成之后，順序回到機(jī)庫中，圖5為協(xié)同收獲試驗(yàn)現(xiàn)場。

圖5 增城試驗(yàn)基地水稻試驗(yàn)田智能收獲試驗(yàn)現(xiàn)場

路徑設(shè)計(jì)：按照試驗(yàn)計(jì)劃，設(shè)計(jì)收獲路徑和轉(zhuǎn)運(yùn)路徑。收獲路徑根據(jù)田塊大小采用先收外圈再平行套圈的方式實(shí)現(xiàn)覆蓋，收割路徑幅寬設(shè)為1.9 m，共28條作業(yè)邊，協(xié)同路徑選擇在短邊的機(jī)耕道上。收獲作業(yè)時(shí)短邊行駛直線路徑重合，轉(zhuǎn)運(yùn)車協(xié)同只需要規(guī)劃1條復(fù)用的路徑，在該條路徑上前進(jìn)或倒車就可以完成全田的糧食對位轉(zhuǎn)運(yùn)工作，具體轉(zhuǎn)運(yùn)點(diǎn)由收獲機(jī)發(fā)出的坐標(biāo)計(jì)算獲得，都位于該路徑上，協(xié)同路徑如圖6a所示。

試驗(yàn)分析：水稻收獲試驗(yàn)中，智能收獲機(jī)與轉(zhuǎn)運(yùn)車按照預(yù)定的路徑完成協(xié)同作業(yè)流程，作業(yè)軌跡如圖6b，圖中的幾處軌跡異常是因?yàn)閿?shù)據(jù)記錄通訊延時(shí)導(dǎo)致。規(guī)劃了12組套圈路徑，每2組轉(zhuǎn)運(yùn)一次，按照設(shè)計(jì)共進(jìn)行了6次轉(zhuǎn)運(yùn)協(xié)同，圖6c、6d和6e分別是記錄的網(wǎng)絡(luò)通訊邏輯觸發(fā)信號G1、G2和G1按照預(yù)定邏輯先后正常觸發(fā)，協(xié)同全過程對位準(zhǔn)確，連續(xù)協(xié)同工作時(shí)間不少于120 min，完成協(xié)同收獲作業(yè)并返回機(jī)庫。試驗(yàn)表明水稻收獲雙機(jī)協(xié)同作業(yè)策略可以實(shí)現(xiàn)的水稻收獲/卸糧轉(zhuǎn)運(yùn)自主作業(yè)，收獲效率為0.35 hm2/h，能夠?yàn)樗局悄芨采w收獲協(xié)同作業(yè)提供支撐。

圖6 水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同作業(yè)試驗(yàn)結(jié)果

5 結(jié) 論

針對水稻導(dǎo)航收獲協(xié)同環(huán)節(jié)多、糧食需要轉(zhuǎn)運(yùn)等問題，本文設(shè)計(jì)了一種水稻收獲機(jī)和轉(zhuǎn)運(yùn)車智能協(xié)同作業(yè)策略，將具體的技術(shù)進(jìn)行了路徑跟蹤算法、協(xié)同對位算法、定位信息融合、通訊鏈路、底盤線控等技術(shù)按照策略有機(jī)組合，最終實(shí)現(xiàn)水稻收獲機(jī)和轉(zhuǎn)運(yùn)車的智能化作業(yè)，具體論證如下。

1）建立了協(xié)同環(huán)節(jié)的有限狀態(tài)機(jī)模型，構(gòu)建了觸發(fā)器、評估器、執(zhí)行器和確認(rèn)器模塊，建立了收獲轉(zhuǎn)運(yùn)環(huán)節(jié)的狀態(tài)參數(shù)，定義了狀態(tài)信息矩陣以表示各項(xiàng)狀態(tài)的具體內(nèi)容，結(jié)合該矩陣和基礎(chǔ)模塊設(shè)計(jì)了狀態(tài)轉(zhuǎn)移鏈，編制了協(xié)同邏輯框架。采用Stateflow仿真軟件對協(xié)同邏輯框架進(jìn)行了仿真驗(yàn)證。

2）構(gòu)建了履帶式水稻協(xié)同收獲系統(tǒng)，制定了雙機(jī)通訊協(xié)議，設(shè)計(jì)了水稻收獲雙機(jī)協(xié)同試驗(yàn)，包括作業(yè)流程和作業(yè)路徑。進(jìn)行田間水稻收獲試驗(yàn)，收獲速度為0.8 m/s，收割幅寬1.9 m，共28條收獲邊，協(xié)同路徑選擇在短邊的機(jī)耕道上，連續(xù)協(xié)同工作時(shí)間大于等于120 min，采用套圈路徑自主收獲0.7 hm2水稻，期間進(jìn)行了6次自動(dòng)協(xié)同轉(zhuǎn)運(yùn)作業(yè)，將所收獲的糧食轉(zhuǎn)運(yùn)到卡車上，試驗(yàn)結(jié)果表明，可以實(shí)現(xiàn)的水稻收獲/卸糧轉(zhuǎn)運(yùn)自主作業(yè)，收獲效率為0.35 hm2/h，狀態(tài)信號正常，能夠?yàn)樗臼斋@協(xié)同作業(yè)提供支撐。

由于現(xiàn)有系統(tǒng)中未安裝糧倉傳感器，所以策略中的倉滿、倉空采用產(chǎn)量預(yù)估和定時(shí)器進(jìn)行替換，及時(shí)性和準(zhǔn)確性不高，未能實(shí)現(xiàn)策略中的所設(shè)計(jì)的全部功能。后期加入糧倉監(jiān)測系統(tǒng)后可更好地實(shí)現(xiàn)收獲策略。未來的研究可以將根據(jù)收獲實(shí)際情況設(shè)計(jì)動(dòng)態(tài)路徑規(guī)劃方法和融入隨車轉(zhuǎn)運(yùn)模式2，將進(jìn)2種模式相結(jié)合形成混合轉(zhuǎn)運(yùn)模式?；诩Z倉傳感器和已作業(yè)區(qū)域數(shù)據(jù)選擇執(zhí)行機(jī)耕道轉(zhuǎn)運(yùn)還是隨車協(xié)同轉(zhuǎn)運(yùn)。當(dāng)需要轉(zhuǎn)運(yùn)糧食時(shí)判斷是否有可協(xié)同的安全空間可以使用，如果沒有足夠的空間，自動(dòng)設(shè)置在機(jī)耕道定點(diǎn)轉(zhuǎn)運(yùn)。經(jīng)過幾圈收獲作業(yè)后將隨車轉(zhuǎn)運(yùn)空間騰出后，即可指定隨車轉(zhuǎn)運(yùn)模式，提高效率的同時(shí)確?？尚行浴?/p>

[1] 羅錫文，廖娟，胡煉，等. 我國智能農(nóng)機(jī)的研究進(jìn)展與無人農(nóng)場的實(shí)踐 [J]. 華南農(nóng)業(yè)大學(xué)學(xué)報(bào)，2021，42(6)：8-17.

Luo Xiwen, Liao Juan, Hu Lian, et al. Research progress of intelligent agricultural machinery and practice of unmanned farm in China[J]. Journal of South China Agricultural University, 2021, 42(6): 8-17. (in English with Chinese abstract)

[2] 趙春江. 智慧農(nóng)業(yè)的發(fā)展現(xiàn)狀與未來展望[J]. 華南農(nóng)業(yè)大學(xué)學(xué)報(bào)，2021，42(6)：1-7.

Zhao Chunjiang. Current situations and prospects of smart agriculture[J]. Journal of South China Agricultural University, 2021, 42(6): 1-7. (in Chinese with English abstract)

[3] Aravind K R, Raja P, Manuel Pérez-Ruiz. Task-based agricultural mobile robots in arable farming: A review[J]. Spanish Journal of Agricultural Research, 2017, 15(1): 1-16.

[4] Han Shufeng, He Yong, Fang Hui. Recent development in automatic guidance and autonomous vehicle for agriculture: A review[J]. Journal of Zhejiang University(Agriculture and Life Sciences), 2018, 44(4): 381-391, 515.

韓樹豐，何勇，方慧. 農(nóng)機(jī)自動(dòng)導(dǎo)航及無人駕駛車輛的發(fā)展綜述[J]. 浙江大學(xué)學(xué)報(bào)(農(nóng)業(yè)與生命科學(xué)版)，2018，44(4)：381-391，515. (in English with Chinese abstract)

[5] 蘭玉彬，趙德楠，張彥斐，等. 生態(tài)無人農(nóng)場模式探索及發(fā)展展望[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2021，37(9)：312-327.

Lan Yubin, Zhao Denan, Zhang Yanfei, et al. Exploration and development prospect of eco-unmanned farm modes[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(9): 312-327. (in English with Chinese abstract)

[6] 周俊，何永強(qiáng). 農(nóng)業(yè)機(jī)械導(dǎo)航路徑規(guī)劃研究進(jìn)展[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2021，52(9)：1-14.

Zhou Jun, He Yongqiang. Research progress on navigation path planning of agricultural machinery[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(9): 1-14. (in Chinese with English abstract)

[7] 黃小毛，張壘，王紹帥，等. 凸多邊形田塊下油菜聯(lián)合直播機(jī)組作業(yè)路徑規(guī)劃[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2022，53(1)：33-40，150.

Huang Xiaomao, Zhang Lei, Wang Shaoshuai, et al. Path planning of rapeseed combine seeder in field of convex boundary[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(1): 33-40, 150. (in Chinese with English abstract)

[8] 張漫，季宇寒，李世超，等. 農(nóng)業(yè)機(jī)械導(dǎo)航技術(shù)研究進(jìn)展[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2020，51(4)：1-18.

Zhang Man, Ji Yuhan, Li Shichao, et al. Research progress of agricultural machinery navigation technology[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(4): 1-18. (in Chinese with English abstract)

[9] 張朝宇，董萬靜，熊子慶，等. 履帶式油菜播種機(jī)模糊自適應(yīng)純追蹤控制器設(shè)計(jì)與試驗(yàn)[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2021，52(12)：105-114.

Zhang Chaoyu, Dong Wanjing, Xiong Ziqing, et al. Design and experiment of fuzzy adaptive pure pursuit control of crawler-type rape seeder[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(10): 105-114. (in Chinese with English abstract)

[10] 張智剛，羅錫文，趙祚喜，等. 基于Kalman濾波和純追蹤模型的農(nóng)業(yè)機(jī)械導(dǎo)航控制[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2009，40(S)：6-12.

Zhang Zhigang, Luo Xiwen, Zhao Zouxi, et al. Trajectory tracking control method based on kalman filter and pure pursuit model for agricultural vehicle[J]. Transactions of the Chinese Society of Agricultural Engineering, 2009, 40(S): 6-12. (in Chinese with English abstract)

[11] 張聞?dòng)?，丁幼春，廖慶喜，等. 基于SVR逆向模型的拖拉機(jī)導(dǎo)航純追蹤控制方法[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2016，47(1)：29-36.

Zhang Wenyu, Ding Youchun, Liao Qingxi, et al. Pure pursuit control method based on SVR inverse-model for tractor navigation[J]. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(1): 29-36. (in Chinese with English abstract)

[12] Gxa B, Mc A, Xh C, et al. Path following control of tractor with an electro-hydraulic coupling steering system: Layered multi-loop robust control architecture[J]. Biosystems Engineering, 2021, 209: 282-299.

[13] Hsiao Y C, Farzaneh K, Stavros G. V, et al. Developing and evaluating an autonomous agricultural all-terrain vehicle for field experimental rollover simulations[J]. Computers and Electronics in Agriculture, 2022, 194, 106735.

[14] Zhang L H, Zhang R R, Li L L, et al. Research on virtual Ackerman steering model based navigation system for tracked vehicles[J]. Computers and Electronics in Agriculture, 2022, 192, 106615.

[15] Aravind K R, Raja P, Manuel Pérez-Ruiz. Task-based agricultural mobile robots in arable farming: A review[J]. Spanish Journal of Agricultural Research, 2017, 15(1): 1-16.

[16] 王輝，王桂民，羅錫文，等. 基于預(yù)瞄追蹤模型的農(nóng)機(jī)導(dǎo)航路徑跟蹤控制方法[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2019，35(4)：11-19.

Wang Hui, Wang Guimin, Luo Xiwen, et al. Path tracking control method of agricultural machine navigation based on aiming pursuit model[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(4): 11-19. (in Chinese with English abstract)

[17] Noguchi N, Will J, Reid J, et al. Development of a master- slave robot system for farm operations[J]. Computers & Electronics in Agriculture, 2004, 44(1): 1-19.

[18] Iida M, Kudou M, Ono K, et al. Automatic following control for agricultural vehicle[C]. Istanbul, Turkey, 6th International Workshop on Advanced Motion Control. Proceedings, 2000, 8494: 158-162.

[19] 白曉平，王卓，胡靜濤，等. 基于領(lǐng)航-跟隨結(jié)構(gòu)的聯(lián)合收獲機(jī)群協(xié)同導(dǎo)航控制方法[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2017，48(7)：14-21.

Bai Xiaoping, Wang Zhuo, Hu Jingtao, et al. Harvester group corporative navigation method based on leader-follower structure[J]. Transactions of the Chinese Society for Agricultural Machinery, 2017, 48(7): 14-21. (in Chinese with English abstract)

[20] 張聞?dòng)睿瑥堉莿?，羅錫文，等. 收獲機(jī)與運(yùn)糧車縱向相對位置位速耦合協(xié)同控制方法與試驗(yàn)[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2021，37(9)：1-11.

Zhang Wenyu, Zhang Zhigang, Luo Xiwen, et al. Position-velocity coupling control method and experiments for longitudinal relative position of harvester and grain truck[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(9): 1-11. (in Chinese with English abstract)

[21] 王猛. 農(nóng)機(jī)多機(jī)協(xié)同作業(yè)任務(wù)分配關(guān)鍵技術(shù)研究[D]. 北京：中國農(nóng)業(yè)機(jī)械化科學(xué)研究院，2021.

Wang Meng. Research on Key Technologies on Farm Task Allocation for Multi-Machine Cooperative Operation[D]. Bei jing: Chinese Academy of Agricultural Mechanization Sciences, 2021. (in Chinese with English abstract)

[22] 曹如月，張振乾，李世超，等. 基于改進(jìn)A*算法和Bezier曲線的多機(jī)協(xié)同全局路徑規(guī)劃[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2021，52(S)：548-554.

Cao Ruyue, Zhang Zhenqian, Li Shichao, et al. Multi-machine cooperation global path planning based on A-star algorithm and bezier curve[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(S): 548-554. (in Chinese with English abstract)

[23] 宮金良，王偉，張彥斐，等. 基于農(nóng)田環(huán)境的農(nóng)業(yè)機(jī)器人群協(xié)同作業(yè)策略[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2021，37(2)：11-19.

Gong Jinliang, Wang Wei, Zhang Yanfei, et al. Cooperative working strategy for agricultural robot groups based on farmland environment[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(2): 11-19. (in Chinese with English abstract)

[24] 姚竟發(fā)，滕桂法，霍利民，等. 聯(lián)合收割機(jī)多機(jī)協(xié)同作業(yè)路徑優(yōu)化[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2019，35(17)：12-18.

Yao Jingfa, Teng Guifa, Huo Limin, et al. Optimization of cooperative operation path for multiple combine harvesters without conflict[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(17): 12 -18. (in Chinese with English abstract)

[25] 翟志強(qiáng)，王秀倩，王亮，等. 面向主從跟隨協(xié)同作業(yè)的導(dǎo)航路徑規(guī)劃方法[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào)，2021，52(S)：542-547.

Zhai Zhiqiang, Wang Xiuqian, Wang Liang, et al. Collaborative path planning for autonomous agricultural machinery of master-slave cooperation[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(S): 542-547. (in Chinese with English abstract)

[26] Jin Y, Wei SQ, Yuan J, et al. Hierarchical and stable multiagent reinforcement learning for cooperative navigation control[J]. IEEE Transactions on Neural Networks and Learning Systems( Early Access ), 2021, 9(19): 1-14.

[27] Gxa B, Mc A, Xh C, et al. Path following control of tractor with an electro-hydraulic coupling steering system: Layered multi-loop robust control architecture[J]. Biosystems Engineering, 2021, 209: 282-299.

[28] Valery P, Lyudmila V, Inna N, et al. The problem of choice of optimal technological decisions on harvester control[J]. MATEC Web of Conferences, 2018, 226: 1-6

[29] Wang H L, Yao Z X, Guo Y H. Collaborative control of unmanned underwater vehicles[J]. Journal of Physics: Conference Series, 2021, 1887(1): 1-5.

[30] 朱良麒，丁力平，陳文亮，等.-Shape算法雙機(jī)器人協(xié)作工作空間研究[J]. 機(jī)械設(shè)計(jì)與制造雜志社，2021(10)：267-271，278.

Zhu Liangqi, Ding Liping, Chen Wenliang, et al. The study of dual-robot cooperative workspace based on α-shape 3d reconstruction theory[J]. Machinery Design & Manufacture, 2021(10): 267-271, 278. (in Chinese with English abstract)

[31] 牛作碩，宮金良，張彥斐. 基于多路線追蹤的機(jī)器人局部路徑規(guī)劃與實(shí)驗(yàn)[J]. 計(jì)算機(jī)應(yīng)用與軟件，2022，39(1)：60-64.

Niu Zuoshuo, Gong Jinliang, Zhang Yanfei. Robot local path planning and experriment based on multi-path tracking[J]. Computer Application and Software, 2022, 39(1): 60-64. (in Chinese with English abstract)

[32] Liu L Y, Liu Q Y, Song Y, et al. A collaborative control method of dual-arm robots based on deep reinforcement learning[J]. Applied Sciences, 2021, 11(4): 1-16.

[33] 孫鵬，譚玉璽，湯磊. 基于有限狀態(tài)機(jī)的作戰(zhàn)實(shí)體模型行為規(guī)則可視化建模[J]. 指揮控制與仿真，2015，37(2)：27-30.

Sun Peng, Tan Yuxi, Tang Lei. Visual modeling of combat entities behavior model rules based on finite state machine[J]. Command Control & Simulation, 2015, 37(2): 27-30. (in Chinese with English abstract)

[34] 聞霞，任雯，賴森財(cái)，等. 基于有限狀態(tài)機(jī)模型的全自動(dòng)燙印機(jī)控制系統(tǒng)設(shè)計(jì)[J]. 工程設(shè)計(jì)學(xué)報(bào)，2020，27(6)：771-780.

Wen Xia, Ren Wen, Lai Sencai, et al. Design of control system of automatic hot stamping machine based on finite state machine model[J]. Chinese Journal of Engineering Design, 2020, 27(6): 771-780. (in Chinese with English abstract)

Cooperative autonomous operation strategy and experiment of the rice harvester together with a rice-transporting vehicle

Zhang Wenyu1,2,3, Zhang Zhigang1,2,3, Zhang Fan1,3, Ding Fan1,3, Hu Lian1,2,3, Luo Xiwen1,2,3※

(1.,,,510642,; 2.,510642,; 3(),510642,)

Many links and complex cooperative operations have posed a great challenge to the autonomous harvesting between rice harvester and transfer vehicle. In this study, a cooperative operation strategy was designed for the autonomous rice harvester and transfer vehicle using Finite State Machine (FSM). The cooperative mode was then divided into four links: independent harvesting, waiting for calls, cooperative truck unloading grain, and transportation. An FSM model was also established to construct the basic components of a collaborative harvesting state machine. After that, the state information matrix was defined to design the specific flow of basic action execution, including the harvester starts harvesting, stops harvesting, starts unloading grain and stops unloading grain. The transfer vehicle was then driven at the waiting point to cooperate with the alignment, then to start or stop grain unloading. As such, the state transition chain of collaborative work was constructed to clarify the transition relationship and trigger conditions between the states in the process of collaboration. The cooperative control logic framework involved the harvester and transfer vehicle, according to the state transition chain architecture. Stateflow tool was selected to simulate and verify the compiled logic framework in the MATLAB platform. The sequential execution was also simplified to introduce the timing and signal transmission delay of internal execution. The simulation results show that the states of the harvester and transfer vehicles were transferred orderly, particularly with the correct conversion of the trigger signal. The test results also show that the control logic strategy performed better for cooperative harvesting. A crawler-type rice collaborative harvesting system was constructed to verify the actual operating performance of the logic strategy, including the crawler rice harvester (Weichai Lovol Heavy Industry RG70V4G-014) and crawler rice transfer vehicle (Weichai Lovol Heavy Industry RG70V4G-015). Among them, the two intelligent agricultural machines adopted the fully electronically controlled chassis with the wire-controlled clutch, header, grain cylinder, and crawler driving system. The dual antenna BDS positioning system (Sina K726) was also equipped, where the data transfer unit (USR-G781) was used in the communication between two computers in the fixed-point cooperative experiment of rice harvesting. The 4/5 G Data Transfer Unit (DTU) with human cloud was adopted at the same time. The autonomous control module was communicated with the control terminal through RS-232. A Self-control terminal (eAgri-800-RS) was installed in the two computers for the harvesting self-control, which communicated with the chassis Electronic Control Unit of the two computers through CAN bus. The software system was developed by Keil uVision 5. The linear path tracking was adopted to follow the model control in the navigation system. The two-computer alignment control was adopted to deal with the position error PID. The collaborative system test was carried out in the Zengcheng Experimental Base of South China Agricultural University. The harvester and transfer vehicle were designed to independently work in the field from the hangar of the base. Specifically, the harvesting speed and width were set at 0.8 m/s, and 1.9 m, respectively. The continuous cooperative working time was not less than 120 min. About 0.7 hm2of rice were automatically harvested by the ferrule path. Six operations of the automatic cooperative transfer system were carried out to transfer the grain to the truck during this period. The transfer truck waited on the tractor road, and then transferred the harvested grain to the roadside truck, according to the designed fixed-point cooperative operation strategy. The harvester and transfer vehicle returned to the hangar in sequence after harvesting the whole farmland. The harvest and transportation path were also designed in the test plan. Among them, the specific harvesting path was designed to cover the field, where the outer ring was harvested first and then parallel rings, according to the size of the field. The cooperative path was selected on the short-side tractor road. A balance was obtained on the short-side straight-line paths of the harvesting operation. The transfer vehicles only needed to plan a reusable path. The grain alignment transfer work was completed to advance or reverse this path in the whole field. The specific transfer points were calculated from the coordinates issued by the harvester, all of which were located on this path. Consequently, the fixed-point cooperative operation was also realized using the autonomous harvester and the transfer vehicle, according to the predetermined path. The logic signals were successively recorded to normally trigger during network communication under the predetermined logic framework in the test. The whole cooperative process was aligned accurately to successfully complete the fixed-point cooperative harvesting operation and return to the hangar. Therefore, the cooperative operation strategy of double machines for rice harvesting was effective and reliable under the configuration, and the harvesting efficiency was 0.35 hm2/h. The finding can also provide strong support for the cooperative operation of autonomous full coverage harvesting in rectangular rice regions.

agricultural machinery; rice; harvesting; unmanned farm; transport; cooperation strategy between harvester and transfer vehicle

10.11975/j.issn.1002-6819.2022.15.001

S147.2

A

1002-6819(2022)-15-0001-09

張聞?dòng)?，張智剛，張帆，? 水稻收獲轉(zhuǎn)運(yùn)雙機(jī)協(xié)同自主作業(yè)策略與試驗(yàn)[J]. 農(nóng)業(yè)工程學(xué)報(bào)，2022，38(15)：1-9.doi：10.11975/j.issn.1002-6819.2022.15.001 http://www.tcsae.org

Zhang Wenyu, Zhang Zhigang, Zhang Fan, et al. Cooperative autonomous operation strategy and experiment of the rice harvester together with a rice-transporting vehicle[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(15): 1-9. (in Chinese with English abstract) doi：10.11975/j.issn.1002-6819.2022.15.001 http://www.tcsae.org

2022-03-30

2022-06-06

國家重點(diǎn)研發(fā)計(jì)劃項(xiàng)目（2021YFD2000602）；智慧農(nóng)場技術(shù)與裝備集成應(yīng)用模式與數(shù)字化展示（2130106）；廣東省基礎(chǔ)與應(yīng)用基礎(chǔ)研究基金項(xiàng)目（2019A1515111152）

張聞?dòng)睿┦?，講師，研究方向?yàn)榫珳?zhǔn)農(nóng)業(yè)、無人農(nóng)場關(guān)鍵技術(shù)、農(nóng)業(yè)機(jī)械智能控制。Email：zhangwenyu@scau.edu.cn

羅錫文，教授，中國工程院院士，研究方向?yàn)檗r(nóng)業(yè)機(jī)械化、電氣化和自動(dòng)化。Email：xwluo@scau.edu.cn