飛思卡爾智能車大獎賽（電磁組2）軟件控制系統(tǒng)設計與開發(fā)

飛思卡爾智能車大獎賽（電磁組2）軟件控制系統(tǒng)設計與開發(fā),卡爾,智能,大獎賽,電磁,軟件,控制系統(tǒng),設計,開發(fā) 基于 PID和模糊邏輯車輛控制策略分析控制Hui-min Lia, *, Xiao-bo Wangb, Shang-bin Songa, Hao Lia摘要: 建立二自由度車輛動力學模型，根據(jù) PID 理論控制與模糊邏輯控制，通過使用橫擺力矩控制與不同控制策略的方法來實現(xiàn)車輛的穩(wěn)定性控制。通過對比和分析 PID 控制與模糊邏輯控制算法影響，得到如下結論：根據(jù)對比控制效果，偏移角和偏移率結合的控制優(yōu)于單一的偏移角和偏移率控制；模糊邏輯控制擁有一個更好的魯棒性而以分析控制理論 PID 控制更是簡單與實用。依據(jù)實際的控制需要，不同的控制方法可能用于相同的控制系統(tǒng)中。此結果可以提高并增強客車的可操作性與穩(wěn)定性，同時在某一程度上會給予其一定的參考。簡介車輛的穩(wěn)定性控制（VSC）在 90 年代得到發(fā)展，它是一種新的安全控制系統(tǒng)，當車輛在轉向或者是受到側向力時，它會適應與匹配輪胎縱向力，使它擁有更好的可操作性與穩(wěn)定性。對于輪胎的非線性特征，車輛的可操作性與穩(wěn)定性控制的主要方法是從四輪轉向控制（4WS）最初的直接橫擺力矩控制和主動前輪轉向開發(fā)控制（AFS） 。在特殊情況下，在車輛可操作性與穩(wěn)定性控制條件下 DYC 已經(jīng)成為一種更有效的控制方法了。根據(jù)不同的控制方法，偏移率，偏移角度，橫向加速度輪胎的滑移率與各個控制要素的結合被用作不同的控制變量?？刂评碚摰膽脧腜ID 控制，最優(yōu)控制和適應性控制發(fā)展到滑移系統(tǒng)中的多變量結構控制，模糊邏輯控制和人工神經(jīng)網(wǎng)絡控制等等。建立兩種自由度的車輛動力學模型，根據(jù)通過使用 PID 與模糊邏輯控制的DYC 控制方法，滑移角度與偏移率被用作控制變量。不同控制理論與控制方法的特征與影響會通過模擬進行對比與分析。1. 車輛模型的建立控制系統(tǒng)根據(jù)二自由度的線性車輛模型而設計。車輛運動依照不同的等式：式子中的 M 為車輛質量，V 為其速度，Yf 為其前輪所受側向力，Yr 是后輪所受到的側向力，β 是滑移角，r 為滑移率，I 是垂直軸時刻的慣性，If 與 Ir是前車軸與后車軸到車輛中心質量的距離。輪胎所受側向力公式:根據(jù)拉普拉斯變換，轉換兩個自由度線性等式方程，得到如下等式兩個自由度的線性車輛模型原生率可以被用作名義的車輛偏移率，在等式(4)中可以看到；為隨后的在輪胎附著力的限制側向力的條件，車輛理想原生率必須受到路附著系數(shù)的限制，滿足其約束。(5)當滑移角非常小時，側向加速度可以有以下公式表示（6）：車輛理論上的偏移率必須滿足下等式（7） ；名義上的偏移率可由等式（8）得到：二自由度線性車輛模型滑移角可當用于車輛滑移角，由等式（9）得到；考慮到輪胎與道路附著系數(shù)之間的密切關系，滑移角由等式（10）表示；一下的等式源于等式（10） ；名義滑移角的最小絕對值應該落在 βN 與 βnmax 之間?？刂葡到y(tǒng)的設計是基于兩個自由度線性車輛模型。正如它是基于采用橫擺力矩車輛的不同控制算法，兩個自由度線性車輛模型依據(jù)下面等式是正確的。這是 Mz 是添加的橫擺動量。2. 控制系統(tǒng)的設計2.1. PID控制角速度圖一反應了角速度的反饋控制，角速度傳感器傳輸車輛角速度與名義角速度的不同到控制器中。當輸入變量改變，控制器會調節(jié)角動量。接著制動力會作用在每個輪胎上，增強型 PID 控制算法會被使用到，有關它的等式符合以下關系：2.2. PID角度滑移控制圖二反應了角度滑移反饋控制?；平窃诖蠖鄶?shù)車輛穩(wěn)定控制系統(tǒng)中作為控制變量。理想滑移角與真實滑移角的差值被當作輸入控制器變量：2.3. 模糊邏輯控制雙公路列車的結果為留在一級公路路口模擬顯示在圖 8。模糊邏輯的反饋控制示于圖 3。在偏移率 γ 和名義偏移率 γN 之間的誤差 e 和誤差 Ec，被用作為邏輯控制器的輸入變量，輸出變量為相應的動量變化△M[6].在理想滑移角與真實滑移角之間的不同誤差 e 和誤差率 ec，被用作模糊邏輯控制的輸入變量，所對應的輸出變量是偏移調整動量△M，在圖 4 中展示。根據(jù)邏輯控制器的設計需要，模糊變量 E,EC 和 U 的定量領域被如下定義。E 和 EC 的模糊集定義為{NB, NM, NS, ZE, PS, PM, PB}.U 的模糊集定義為{ZE, PS, PM, PB}共同的全集模糊定義為：E and EC [-0.1, -0.8, -0.6, -0.4, -0.2, 0, 0.2, 0.4, 0.6, 0.8, 1]; U [0, 0.1, 0.2, 0.3,0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1]; E 與 EC [-0.1, -0.8, -0.6, -0.4, -0.2, 0, 0.2,0.4，0.6, 0.8, 1]; U [0, 0.1, 0.2, 0.3, 0.4,0.5, 0.6, 0.7, 0.8, 0.9, 1].為了使全集變量誤差 e，誤差率 ec 和控制變量 u 標準化，量子化因素和比例因子給予以下定義。把量子化誤差變量定義為 Ke=e-1把量子化誤差率定義為 Kec=ec-1控制變量的比例因子定義為 Ku=u-1采用三角函數(shù)輸入語言變量和輸出語言變量，可以有更好的模擬效果和更高的精度的操作簡單，如圖 5 ~圖 7 所示。模糊邏輯控制規(guī)則可以由條件語句“if……then……”來表達，這代表了來源于一些前提條件的改變?？傮w來說，模糊邏輯控制形式被用作表達這些規(guī)則，正如表一所示。兩個輸入各自有七個模糊語言變量，因此有 49 種方式。規(guī)則 1：如果 R 是 NR 且 RC 是 NR，則 NR 是 PR：規(guī)則 2：如果 E 是 NB 且 EC 是 NM ，則 U 是 PB……規(guī)則 49：如果 E 是 PB 且 EC 是 PB ，則 U 是 PB;NB 是負大值，NM 是負中值，NS 為負小值，ZE 幾乎為零；PS 為正小值，PM 為正中值；PB 為正大值。馬丹妮方法被用于模糊邏輯控制，用最大最小方法做推導，用二等分線方法當去模糊化處理。在 matlab 的模糊控制工具作為模糊控制器。輸入與輸出的模糊控制表面在圖 8 中2.4. 綜合控制通常情況下，依據(jù)控制參數(shù)的調整，對于外界干擾 PID 控制不能獲得理想的控制效果。在 PID 控制中，偏移率的控制變量應該要通過模糊邏輯控制器的調節(jié)來控制。實際中，角偏移與角偏移率被傳感器所估量與測量。因此偏移率的誤差被 PID 控制器作為輸入?yún)?shù)，偏移角度誤差被模糊邏輯控制器作為輸入?yún)?shù)使用。整合了模糊 PID 控制的系統(tǒng)結構在圖 9 中展示。增強型 PID 控制算法可以用作 PI 控制。PID 控制參數(shù) Kp 和 Ki 通過即時的模糊控制器調節(jié)2.5. 臨界值控制根據(jù) PID 偏移率控制算法和結合了邏輯控制的滑移角度模糊控制，建立起關于偏移率與偏移角模型，在圖 10 中展現(xiàn)。該控制器并不影響其獨立性，但只有增加的閾值規(guī)則的控制。當車輛需要制動時，控制器控制滑移角滿足約束和滑移率 PID 控制器控制車輛的行駛。當滑移角不滿足約束，這是需要通過滑移角模糊邏輯控制器控制車輛的駕駛。所有上述是基于穩(wěn)定的公差。3. 模擬分析中型車輛被選為模擬試驗車輛。車輛速度為 70 公里每小時和方向盤角在角度步驟輸入 70 度設置。仿真曲線的實線是原來沒有控制，虛線由控制得到的。 3.1. PID 控制PID 控制的偏移率和滑移角的模擬展示在圖 11，響應指數(shù)的比較在表二。PID 控制偏移率穩(wěn)定值下降 10.07％，超調量明顯下降。響應間也減少了顯然，很明顯，反應速度增加。 PID 控制的滑移角穩(wěn)定值下降 12.52％，略超調增大。響應時間明顯增加，反應速度明顯下降。3.2. 模糊邏輯控制雙公路列車右轉上二級公路路口模擬結果，顯示在圖 12。橫擺率和滑移角的模糊邏輯控制模擬，示于圖 12 與響應的比較指數(shù)示于表 3。模糊邏輯控制的偏移率穩(wěn)定值下降 14.50％，超調明顯增大。 響應時間明顯下降，響應速度明顯增加。模糊邏輯的車輛穩(wěn)態(tài)值的偏移角速度控制下降 14.75％，超調略增大。響應時間略有增加，響應速度明顯增加。3.3. 綜合模擬雙公路列車轉向對二級公路交叉口的結果仿真如圖 13 所示和響應指數(shù)的比較，如表 4 所示。 組合控制下偏移率穩(wěn)定值降 17.04％，超調明顯增大。響應時間明顯下降，響應速度明顯增加。在聯(lián)合控制模糊邏輯控制的側滑角穩(wěn)定值下降 17.07％，超調量明顯下降。響應時間略有增加，響應速度略有下降。3.4. 臨界值控制PID 控制臨界值和模糊邏輯控制偏移率與模糊邏輯控制滑移角在圖 14 中，響應指數(shù)的比較在表 5。在邏輯臨界下，PID 控制中的偏移率穩(wěn)定值下降 19.08％，超調量略有下降。 響應時間明顯下降，響應速度明顯增加。在邏輯臨界下，模糊邏輯的偏角穩(wěn)態(tài)值下降 14.75％，超調略增大。 響應時間明顯增加，響應速度明顯下降。4. 結論結論可以從方向盤角度階躍輸入模擬得出，沒有可以滿足車輛行駛的所有條件，并取得了良好的效果的控制理論與控制策略。但是在一般的操作條件下，有必要研究更有效的理論方法和控制策略。例如，當車輛打滑和不穩(wěn)定時，有必要施加橫擺力矩的控制。它可以控制滑移角的大小和有效的偏移率。在這種情況下，橫向加速度不會超過后表面附著的極限。如果充分滿足上述條件時，控制較好；若基本滿足，控制效果一般；若不滿足，該控制是不太有效。當車輛是中高速行駛時，滑移角才是相對大的，車輛顯示主要動態(tài)特點?？刂频闹饕康氖菫榉€(wěn)定，因此橫擺率控制可以實現(xiàn)更好的控制。隨著角度的增加，當單獨控制偏移率時候，其結果將會變得糟糕。在高附著地面系數(shù)下，聯(lián)合控制能夠有效地控制滑移角和滑移率并且滿足車輛動態(tài)特征。參考書目；[1]Zhou, H.N., 2007. Study on Vehicle Stability Control Strategy, Journal of Hubei Automotive Industries Institute, 21,26-31.[2]Yu, Z.S., 2000. Automobile theory (The third edition). Beijing: Machinery Industry Press[3]Abe, M., 1998 Vehicle movement and manipulation, Machinery Industry Press[4]Yao, S.Y., 2008. Study on vehicle stability control system, Xihua University, pp.16-27[5]Yang, S.Z., Yang K.C., 2005. Mechanical Engineering Control Basis, Huazhong science university Press[6]Zhu, J., 2005. Fuzzy Control Theory and System Principle, Machinery Industry Press外文二：Vehicle Control Strategies Analysis Based on PID and Fuzzy Logic ControlHui-min Li a, *, Xiao-bo Wang b , Shang-bin Song a , Hao Li aa Research Institute of Highway Ministry of Transport, Beijing 100088, Chinab China National Construction Machinery Quality Supervision Testing Center, No.55 Dong Wai Street YanQing City Beijing 102100, ChinaAbstractTwo degrees of freedom vehicle dynamic model is established. Based on theories of PID control and fuzzy logic control, controller of vehicle stability is designed by using the method of direct yaw moment control and the different control strategies. By comparing and analyzing control effect of PID control and fuzzy logic control, the result shows as follows: slip angle and yaw rate combined control is better than slip angle and yaw rate controlled individually by comparing control effect; fuzzy logic control have a better robustness and PID control is simple and practical by analyzing control theories. Different control methods can be used in the same control systems according to the need of practical application. The result can improve and enhance passenger car maneuverability and stability control and also can give some reference in a way. Peer-review under responsibility of the Department of Transportation Engineering, Beijing Institute of Technology.Keywords: Vehicle dynamic; Control strategies; PID control; Auto Turn; Fuzzy logic control.1. IntroductionVehicle Stability Control (VSC) is developed during the 90’s. It’s a new active safety control system with a better maneuverability and stability by regulating and matching tire longitudinal force when vehicle is steering or under the lateral force. For tire nonlinear characteristic, main methods of vehicle maneuverability and stability control are developed from four-wheel steering control (4WS) initial to direct yaw moment control and active front steering control (AFS). In particular, DYC has become a more effective method in vehicle maneuverability and stability control. According to the difference control strategies, yaw rate, slip angle, lateral acceleration tire slip ratio and combination of them are used as control variables [1]. The application of control theories is developed from PID control, optimal control and adaptive control to variable-structure control system with sliding mode, fuzzy logic control and artificial neural network control and so on. Two degrees of freedom (2 DOF) vehicle dynamic model is established. Slip angle and yaw rate are used as control variables based on control method of DYC by using control theories of PID and fuzzy logic control. The characteristics and effect of different control theory and control strategy are compared and analyzed by simulation.2. Establishment of Vehicle ModelControl system is designed based on two degrees of freedom liner vehicle model [2,3]. Vehicle movement differential equations are given as follows:Where, M is mass of vehicle, V is vehicle velocity, Yf is lateral force of front tire, Yr is the lateral force of rear tire, βis side slip angle, r is yaw rate, I is inertia of vehicle around the vertical axis moments, lf and lr are the distance between the center of mass with front axle and rear axle.Equations about lateral force of tire are given as follows:Transfer functional equations of two degrees of freedom liner vehicle model are given though Laplace transformation shown as Equ. (1).Where,Raw rate of two degrees of freedom liner vehicle model can be used as vehicle nominal yaw rate, as shown in Equ. (4).Vehicle ideal raw rate must be restricted by road adhesion coefficient and satisfied the constraints as followed on the condition of lateral force in tire adhesion limitation.When slip angle is very small, lateral acceleration can be expressed as followsVehicle ideal yaw rate must be satisfied the following equation.Nominal yaw rate is shown as Equ. (8).Slip angle of two degrees of freedom liner vehicle model can be used as vehicle nominal slip angle, as shown in Equ. (9).Considering the restriction between tire and maximum road adhesion coefficient, slip angle is expressed in Equ.(10).The following equation can be derived from Equ. (10)Nominal slip angle should be the minimum absolute value βN between and βnmax.The design of control system is based on two degrees of freedom liner vehicle model. As it is based on different control algorithms applying yaw moment to vehicle, two degrees of freedom liner vehicle model is corrected as follows.Where Mz is additional yaw moment.3. Design of Control System3.1. PID control of yaw rateFeedback control of yaw rate is shown in Fig. 1. Yaw rate sensor is used to transit the difference between vehicle yaw rate and nominal yaw rate to controller [4,5].When input variables changed, yaw moment will be adjusted by controller. Then brake force is distributed to every wheel. Increasing PID control algorithm is used and its equations are given as follows:3.2 PID control of slip angleFeedback control of slip angle is shown in Fig. 2. Slip angle is used as control variable in most vehicle stability control systems. The difference between ideal slip angle and the actual one is used as input to controller.3.3. Fuzzy logic controlResult of double road train turns left on first class highway intersection simulation is shown in Fig. 8.Feedback control of fuzzy logic is shown in Fig. 3.The error e and its rate ec of difference between yaw rate γ and nominal yaw rate γN are used as input variable of fuzzy logic controller, output variable is yaw adjusted moment△M[6].The error e and its rate ec of difference between ideal slip angle and the actual one are used as input variable of fuzzy logic controller, output variable is yaw adjusted moment △M, as shown in Fig.4 (Zhu, 2005).According to the need of fuzzy logic controller design, quantificational field of fuzzy variables E、EC and U is defined as follows [4].The fuzzy sets of E and EC are defined as {NB, NM, NS, ZE, PS, PM, PB}.The fuzzy set of U is defined as {ZE, PS, PM, PB}.Universes of them are defined as: E and EC [-0.1, -0.8, -0.6, -0.4, -0.2, 0, 0.2, 0.4, 0.6, 0.8, 1]; U [0, 0.1, 0.2, 0.3,0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1]; E and EC [-0.1, -0.8, -0.6, -0.4, -0.2, 0, 0.2, 0.4?0.6, 0.8, 1]; U [0, 0.1, 0.2, 0.3, 0.4,0.5, 0.6, 0.7, 0.8, 0.9, 1].In order to make universe of variable error e, error rate ec and control variable u to correspond the standardization, quantization factor and scale factor are defined as follows.The quantization factor of variable error is defined as Ke=e-1.The quantization factor of the error rate is defined as Kec=ec-1.The scale factor of control variable is defined as Ku=u-1.Triangular membership functions are adopted to input linguistic variable and output linguistic variable, which operates simply with a better effect of simulation and a higher precision, as shown from Fig. 5 to Fig.7.The rules of fuzzy logic control can be expressed by conditional statement of “if…then…”, which represent decision result derived from many change premises. In general, fuzzy logic control form is used to express those rules, as shown in table 1.Two inputs have seven fuzzy linguistic variables separately, so there are 49 rules in total as follows.Rule 1：if R is NR and RC is NR then U is PRRule 2：if E is NB and EC is NM then U is PB……Rule 49：if E is PB and EC is PB then U is PBWhere, NB is Negative Big; NM is Negative Medium; NS is Negative Small; ZE is Almost Zero; PS is Positive Small; PM is Positive Medium; PB is Positive Big.Method of “Mamdani” is used for fuzzy logic control, and “max-min” is used for fuzzy reasoning, bisector of area is used for defuzzification method. Fuzzy control toolbox in Matlab is used for fuzzy logic controller. Input and out input surface of fuzzy controller is showed in Fig. 8.3.4. Combined controlIn general, PID control can’t get the ideal control effect for interfered by the method of control parameter adjustment. Control variable of yaw rate in PID control should be adjusted in fuzzy logic controller. In practice, slip angle is estimated and yaw rate can be measured by the sensor. So the error of yaw rate can be used for PID controller input parameter, error of slip angle can be used for fuzzy logic controller input parameter. The system structure of combined fuzzy PID control is showed in Fig. 9.Increasing PID control algorithm is used for PI control [5].PID control parameters KP and KI are adjusted by fuzzy controller real-time.3.5. Threshold controlBased on the algorithm of yaw rate PID control and slip angle fuzzy control, combined PID fuzzy logic controller is established about yaw rate and slip angle, as shown in Fig.10. The controller doesn’t affect their independence but only increases control of the threshold rules. When vehicle need to brake, controller control slip angle to satisfy constraints and control vehicle’s driving by yaw rate PID controller. When slip angle doesn’t satisfy constraints, it is necessary to control vehicle’s driving by slip angle fuzzy logic controller. All above is based on steady tolerance.4. Simulation AnalysisMedium vehicle is selected as simulation test vehicle. Vehicle speed of 70 km / h and the steering wheel angle of 70 deg in the angle step input are set. The solid line of simulation curves is original with no control, the dashed line is obtained by control.4.1. PID controlPID control simulation of yaw rate and slip angle is shown in Fig. 11 and the comparison of response index is shown in table 2.Yaw rate steady value of PID control decreases 10.07%, overshoot decreases obviously. Step time also decreases obviously, the reaction speed increases obviously. Slip angle steady value of PID control decreases 12.52%, overshoot increases slightly. Step time increases obviously, the reaction speed decreases obviously.4.2. Fuzzy logic controlResult of double road train turns right on second class highway intersection simulation is shown in Fig. 12.Fuzzy logic control simulation of yaw rate and slip angle is shown in Fig. 12 and the comparison of response index is shown in table 3.Yaw rate steady value of fuzzy logic control decreases 14.50%, overshoot increases obviously. Response time decreases obviously, response speed increases obviously. Yaw angle velocity of vehicle steady value of fuzzy logic control decreases 14.75%, overshoot increases slightly. Response step time increases slightly, response speed increases obviously.4.3. Combined simulationResult of double road train turn around on second class highway intersection simulation is shown in Fig. 13 and the comparison of response index is shown in table 4.Yaw rate steady value decreases 17.04% in combined control, overshoot increases obviously. Response time decreases obviously, response speed increases obviously. Slip angle steady value of fuzzy logic control in combined control decreases 17.07%, overshoot decreases obviously. Response time increases slightly, response speed decreases slightly.4.4. Threshold controlThreshold control of PID control and fuzzy logic control simulation of yaw rate and slip angle is shown in Fig. 14 and the comparison of response index is shown in table 5.Yaw rate steady value of PID control in logic threshold decreases 19.08%, overshoot decreases slightly. Response time decreases obviously, response speed increases obviously.Slip angle steady value of fuzzy logic in logic threshold decreases 14.75%, overshoot increases slightly.Response time increases obviously, response speed decreases obviously.5. ConclusionConclusions can be drawn from steering wheel angle step input simulation, there are no control theory and control strategy which can run all the conditions of the vehicle and achieve a good effect. But for general operation condition, it is necessary to study a more effective theoretical method and control strategy. For example, when the vehicle is skidding and instability, it’s necessary to impose yaw moment control. It can control the size of slip angle and yaw rate effectively. In this situation, lateral acceleration will not exceed the limit of later surface attachment. If fully meet the above conditions, the control is better; basically satisfied, the control effect is in general; not satisfied, the control is less effective.Slip angle is relative large when the vehicle is in the middle and high velocity, vehicle shows mainly dynamic characteristics. The main purpose of control is for stability, so yaw rate control can achieve a better control preference. With the increasing of angle, the result will become bad when control separately yaw rate. In the high adhesion coefficient road, combined control can control slip angle and yaw rate efficiently and satisfy the dynamic characteristic.References[1]Zhou, H.N., 2007. Study on Vehicle Stability Control Strategy, Journal of Hubei Automotive Industries Institute, 21,26-31.[2]Yu, Z.S., 2000. Automobile theory (The third edition). Beijing: Machinery Industry Press[3]Abe, M., 1998 Vehicle movement and manipulation, Machinery Industry Press[4]Yao, S.Y., 2008. Study on vehicle stability control system, Xihua University, pp.16-27[5]Yang, S.Z., Yang K.C., 2005. Mechanical Engineering Control Basis, Huazhong science university Press[6]Zhu, J., 2005. Fuzzy Control Theory and System Principle, Machinery Industry Press