Robust estimation method for articulated angle of traction-type AGV (Automatic Guided Vehicle) for multi-vehicle collaborative operation

Abstract

The invention provides a multi-vehicle collaborative operation-oriented traction type AGV hinge angle robust estimation method, and belongs to the technical field of intelligent transportation AGV state observation. Firstly, in order to improve the adaptive capacity of an observer when facing the uncertain parameters of the traction-type AGV, especially for the time-varying longitudinal speed of the traction-type AGV and the trailer mass, a T-S fuzzy model is constructed by adopting a T-S fuzzy method, so that all possible uncertain parameters can be covered; secondly, in order to save communication resources and reduce the calculation burden of an AGV controller, a DETM is provided between a sensor and an observer. Finally, on the basis, an H-infinity robust hinge angle observer is provided, the observer and DETM are designed in a collaborative mode in combination with a fuzzy Lyapunov functional method and an LMI technology, and therefore the robustness of the observer is remarkably improved. According to the method, the robustness problem in hinged angle estimation of the traction type AGV can be effectively solved, and the reliability and stability of multi-vehicle collaborative operation are further improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic guided vehicle (AGV) state observation, and relates to a robust estimation method for the articulation angle of a towing AGV for multi-vehicle collaborative operation, which can be widely applied to scenarios such as warehousing logistics, production line transportation, and automated warehouses. Background Art

[0002] With the rapid development of automated warehousing and logistics environments, traction AGVs play a vital role in scenarios such as warehouses and production lines. In modern warehousing and logistics systems, traction AGVs have extremely high performance requirements when performing tasks in complex environments, especially the coordination between the tractor and the trailer. When traction AGVs perform tasks, accurately estimating the articulation angle (that is, the angle between the tractor and the trailer) plays a key role in reasonable path planning, stability control, and collision avoidance. At present, common articulation angle estimation methods face problems such as high equipment cost, sensor failure, and excessive communication burden. How to improve the estimation accuracy and real-time performance of the system through effective control and observation methods, and play a role in the group intelligent collaboration of multiple AGVs, has become a technical problem that needs to be solved urgently.

[0003] The paper "Hitch Angle Estimation for Trailer Backup System—An Object Detection and Tracking Approach," published in IEEE Transactions on Instrumentation and Measurement, Vol. 71, 2022, proposes a method for observing the articulation angle based on computer vision and machine learning. This method is applied to advanced trailer backup assistance systems to achieve semi-autonomous or fully autonomous reversing maneuvers. The method combines kinematic models of the trailer and tractor, and uses deep learning object detection and computer vision tracking models to process image frames acquired from a rear-facing camera, detect and track the marker lights on the leading edge of the trailer, and estimate the articulation angle. However, this method faces challenges in practical applications, such as high cost, installation complexity, and sensor failure.

[0004] To overcome these shortcomings, model-based estimation methods have become a research hotspot. The paper "Hitch Angle Estimation of a Towing Vehicle With Arbitrary Configuration," published in IEEE Transactions on Intelligent Transportation Systems, Vol. 23, No. 7, 2022, proposes an articulation angle estimation method based on ultrasonic sensors, kinematics, and dynamics equations. First, the articulation angle is directly calculated using distance measurements from four ultrasonic sensors and prior geometric information. For each possible sensor pair, six articulation angle estimates are generated, and a single estimate is derived through a voting algorithm. Second, a recursive articulation angle estimation method based on the least squares method and a Kalman filter, respectively, is proposed using the kinematic and dynamic models of the tractor-trailer system. Finally, a more reliable articulation angle estimation scheme is proposed, integrating the algorithms of the previous three methods by switching data fusion logic. Experimental results demonstrate that this integrated scheme can effectively estimate the articulation angle at both low speeds (using the kinematic model) and high speeds (using the dynamic model), without requiring prior knowledge of the trailer parameters. However, traditional estimation methods often assume that system parameters are known and accurate, but in practical applications, the system parameters may be uncertain or changeable.

[0005] In summary, developing a robust estimation method for the articulation angle of towed AGVs for multi-vehicle collaborative operation considering the uncertainty of system parameters has important practical significance. Summary of the Invention

[0006] To address the above problems, the present invention proposes a robust estimation method for the articulation angle of towed AGVs for multi-vehicle collaborative operation considering the uncertainty of system parameters, and introduces a dynamic event trigger mechanism (DETM) to save communication resources and reduce the computational burden of the vehicle controller.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] A robust estimation method for the articulation angle of a towed AGV for multi-vehicle collaborative operation. First, in order to improve the adaptability of the observer in the face of uncertainty in the parameters of the towed AGV, especially for the time-varying longitudinal speed of the towed AGV and the mass of the trailer, the TS (Takagi-Sugeno) fuzzy method is used to construct a TS fuzzy model, so that all possible uncertainty parameters can be covered. Secondly, in order to save communication resources and reduce the computational burden of the AGV controller, the present invention proposes a DETM between the sensor and the observer. Finally, based on the above, an H∞ robust articulation angle observer is proposed. Combining the fuzzy Lyapunov functional method with the LMI (Linear Matrix Inequality) theory, the observer and DETM are collaboratively designed, thereby significantly improving the robustness of the observer. Through this method, the present invention can effectively solve the robustness problem in the estimation of the articulation angle of the towed AGV and improve the reliability and stability of the system in multi-vehicle collaborative operation. Specifically comprising the following steps:

[0009] Step 1: Build a dynamic model of the traction AGV. The details are as follows:

[0010] Step 1.1: The towed AGV in this invention consists of a two-axle tractor and a single-axle trailer, which are articulated via a fifth wheel. This articulation is off-axis, with the hinge point located behind the centerline of the tractor's rear axle. First, the following assumptions are made for the towed AGV's dynamic model: 1) The steering angle δ, tire slip angle α, and articulation angle θ of the towed AGV are small; 2) The longitudinal velocities of the tractor and trailer are equal and both equal to v x The following degrees of freedom are considered for this dynamic model: 1) the lateral velocity v of the tractor y1 ; 2) the yaw angular velocity ω1 of the tractor; 3) the articulation angle θ between the tractor and trailer. Furthermore, because the pitch and roll motions of a towed AGV have little impact on yaw stability, they are ignored in this invention. Based on the above, the dynamic relationship for a towed AGV is as follows:

[0011]

[0012] Among them, v x represents the longitudinal speed of the traction AGV and trailer (hereinafter referred to as vehicle speed), m1 represents the mass of the tractor, m2 represents the mass of the trailer, and I z1 represents the yaw moment of inertia of the tractor, I z2 represents the trailer's yaw moment of inertia, v y1 Indicates the lateral speed of the tractor, represents the lateral acceleration of the tractor, v y2Indicates the lateral speed of the trailer, represents the lateral acceleration of the trailer, ω1 represents the yaw angular velocity of the tractor, represents the yaw acceleration of the tractor, ω2 represents the yaw velocity of the trailer, represents the trailer's yaw angular acceleration, F yf Indicates the lateral tire force on the front axle of the tractor, F yr Indicates the lateral tire force on the rear axle of the tractor, F yt Indicates the lateral tire force on the trailer axle, F yP represents the lateral articulation force of the articulation point, a1 represents the distance from the front axle of the tractor to the center of mass of the tractor, a2 represents the distance from the rear axle of the tractor to the center of mass of the tractor, a3 represents the distance from the articulation point to the center of mass of the tractor, b1 represents the distance from the articulation point to the center of mass of the trailer, b2 represents the distance from the trailer axle to the center of mass of the trailer, where b1+b2=l2 represents the distance from the trailer axle to the articulation point.

[0013] Given the linear articulated coupling equations, the trailer's lateral velocity and trailer's yaw rate are obtained by the following relations:

[0014]

[0015] Where θ represents the hinge angle, represents the hinge angular velocity.

[0016] Based on formula (2), v in formula (1) y2 and ω2 can be replaced.

[0017] Step 1.2: Model the tire model of the traction AGV.

[0018] Since the lateral tire force is proportional to the tire slip angle when the tire slip angle is small, the lateral tire force acting on each axle of the traction AGV is expressed as:

[0019]

[0020] Among them, C f Indicates the tire cornering stiffness of the front axle of the tractor, C r Indicates the tire cornering stiffness of the tractor's rear axle, C t represents the tire cornering stiffness of the trailer axle, α f represents the tire slip angle of the front axle of the tractor, α r represents the tire slip angle of the rear axle of the tractor, α t Indicates the tire slip angle of the trailer axle.

[0021] In formula (3), the tire side slip angle of each axle of the traction AGV is expressed as:

[0022]

[0023] Where δ represents the steering angle of the traction AGV.

[0024] Step 1.3: Based on the contents of steps 1.1 and 1.2, express the state space equation form of the traction AGV dynamic model. Specifically:

[0025] First, define the state vector as According to formula (1)-formula (4), the state space equation form of the traction AGV dynamic model is expressed as:

[0026]

[0027] in, represents the derivative of the state vector; A represents the state matrix and B represents the input matrix, which are expressed as: Among them, A 11 、A 12 、A 13 、A 14 、A 21 、A 22 、A 23 、A 24 、A 41 、A 42 、A 43 、A 44 、B 11 、B 21 and B 41 denote the coefficients of the system matrix, which are expressed as:

[0028]

[0029] Among them, κ 11 , κ 12 , κ 13 , κ 21 , κ 22 , κ 23 and Ω represent coefficients, which are expressed as:

[0030] κ 11 =a1m1+a1m2+a3m2

[0031] κ 12 =a2m1+a2m2-a3m2

[0032] κ 13 =C f a1-C r a2-C t a3

[0033]

[0034] Then, assuming that the lateral velocity and yaw rate of the tractor in the state vector can be measured by sensors, and taking into account the system modeling error, uncertainty, and measurement error, the dynamic model of the tractor AGV is finally written as:

[0035]

[0036] Where x represents the state vector of the dynamic model of the traction AGV, w1 represents the energy-bounded system modeling error disturbance, w2 represents the energy-bounded measurement error disturbance, y represents the output vector of the dynamic model of the traction AGV, z represents the controlled output vector of the dynamic model of the traction AGV, C represents the output matrix, and D represents the controlled output matrix. The output matrix and the controlled output matrix are expressed as:

[0037] At this point, the dynamic model of the traction AGV has been established.

[0038] Step 2: Based on the TS fuzzy method and the dynamic model of the towed AGV established in step 1 as shown in formula (6), a TS fuzzy model is established to include all possible uncertain parameters to further improve the observer's ability to adapt to the parameter uncertainty of the towed AGV. The details are as follows:

[0039] Step 2.1: The speed of the towing AGV varies during its driving process. Considering the functional attributes of the towing AGV, the mass of the trailer also varies greatly. Therefore, the speed v x and the mass m2 of the trailer also represent uncertainty parameters in the dynamic model of the towing AGV. However, the vehicle speed v x and the mass m2 of the trailer are bounded and are respectively bounded by [v xmin ,v xmax ] and [m 2min ,m 2max ], where v xmin Indicates the minimum longitudinal speed of the traction AGV, v xmax Indicates the maximum longitudinal speed of the traction AGV, m 2min Indicates the minimum mass of the trailer, m 2max Indicates the maximum mass of the trailer. To this end, the present invention uses the TS fuzzy method to describe the dynamic model of the towing AGV with uncertainty parameters of vehicle speed and trailer mass. From the dynamic model of the towing AGV, it can be seen that the uncertainty parameter v x , Both m and m2 exist in the dynamic model of the traction AGV, so the definition is:

[0040]

[0041] in, and represents the reference coefficient.

[0042] Step 2.2: Uncertainty parameter v x , and m2 can be described as:

[0043]

[0044] in, and Denotes membership functions that satisfy:

[0045]

[0046] Among them, the membership function and Expressed as:

[0047]

[0048] Step 2.3: Based on steps 2.1 and 2.2, the TS fuzzy model of the towed AGV with uncertain parameters is expressed as:

[0049]

[0050] Among them, A j Represents the state matrix under different fuzzy rules, B j Represents the input matrix under different fuzzy rules, C j represents the output matrix under different fuzzy rules, D j represents the controlled output matrix under different fuzzy rules, j represents the sequence number of the fuzzy rule, represents the set of fuzzy rules, ρ j represents the fuzzy weighting function under different fuzzy rules, ρ j is represented as:

[0051]

[0052] From (12), we can see that the TS fuzzy model has 16 fuzzy rules. And, according to the fuzzy weighting function ρ under different fuzzy rules defined by formula (10) and formula (12), j The following conditions are met:

[0053]

[0054] The present invention assumes that the vehicle speed and the mass of the trailer can be accurately obtained and are constant within a given range. Therefore, the corresponding fuzzy weighting function ρ can be obtained j .

[0055] Step 3: In order to save limited CAN communication resources, the present invention introduces DETM (Dynamic Event Triggering Mechanism), so that the data measured by the sensor is transmitted through the CAN communication main line only when the event triggering condition is met. The event triggering time is defined as t k ( Where k represents the time series, represents the set of positive integers including 0), then y(t k ) represents the output signal that satisfies DETM, and y(t) represents the real-time output signal measured by the sensor. Then, DETM is constructed as:

[0056]

[0057] Where t0 represents the initial time; η(t) represents the dynamic variable; ( where R + represents a positive real number) represents the coordination factor; σ(σ∈[0,1)) represents a given constant; η0(η0∈R + ) represents known parameters; e y (t) represents the output error, which can be expressed as e y (t) = y(t k )-y(t)(t∈[t k ,t k+1 ), where t represents time, t k+1 represents the next event triggering moment); y(t) represents the real-time output signal measured by the sensor; e y (t - ) represents the output error of the left limit at time t; Indicates that when t>t k η(0) represents the value of the dynamic variable at the initial moment; represents the derivative of the dynamic variable; β (β>0) represents the parameter to be designed. Considering the TS fuzzy model, it is expressed as:

[0058]

[0059] Among them, β j represents β under different fuzzy rules.

[0060] From formula (14), we can see that when , you will get the static ETM (Event Trigger Mechanism):

[0061]

[0062] This means that the static ETM shown in formula (16) is the DETM shown in formula (14) Due to the introduction of the dynamic variable η(t), the data transmission amount under DETM can be further reduced compared with the static ETM.

[0063] According to the properties of DETM shown in formula (14), the following lemma is obtained:

[0064] Lemma 1: For any t∈[0,∞), there exists:

[0065]

[0066] η(t)>0 (18)

[0067] Step 4: An important issue in designing DETM is that the Zone phenomenon cannot exist, that is, any two triggering times t k and t k+1 The distance between them is positive. This step will exclude the Zeno behavior of DETM. The following lemma is introduced to exclude the Zeno behavior of DETM:

[0068] Lemma 2: Assuming x, δ, w1 and w2 are bounded, then any two triggering times t k and t k+1 The distance between them is positive.

[0069] According to Lemma 2, the DETM shown in formula (14) proposed in this invention does not have Zeno behavior.

[0070] Step 5: Design a DETM-based H∞ robust joint angle observer based on the TS fuzzy model of the traction AGV shown in formula (11) obtained in step 2 and the DETM shown in formula (14) obtained in step 3. The details are as follows:

[0071] Step 5.1: First, based on the TS fuzzy model and DETM of the towed AGV, design the Luenberger state observer as shown below:

[0072]

[0073] in, represents the estimated state vector, represents the derivative of the estimated state vector, represents the estimated output vector, represents the estimated controlled output vector, L j represents the observation gain of the observer under different fuzzy rules, Represents the output vector estimated at time t, y(t k ) represents t k Output vector of trigger moment estimates.

[0074] Step 5.2: Define the estimation error and According to formula (19) and formula (11), the closed-loop error system can be obtained as follows:

[0075]

[0076] Among them, e x represents the estimation error of the state vector; represents the derivative of the estimated error of the state vector; e z represents the estimated error of the controlled output vector; e y (t) represents the output error.

[0077] Step 5.3: Considering the system modeling error disturbance, measurement error disturbance and parameter uncertainty disturbance in the TS fuzzy model of the towing AGV as shown in formula (11), in order to achieve accurate estimation of the vehicle driving state under the DETM shown in formula (11), the observation gain L of the Luenberger state observer shown in formula (19) under different fuzzy rules is calculated based on the H∞ theory. j and DETM are designed. Lemma 3 and Theorem 1 are introduced below to analyze the observation gain L under different fuzzy rules. j and DETM are designed, and Lemma 3 is a necessary condition for the proof of Theorem 1.

[0078] Lemma 3: Given real matrices E and F, there exists a positive number ε>0 satisfying the following inequality:

[0079]

[0080] Theorem 1: Given constants γ1>0, γ2>0, γ3>0 and σ∈[0,1), if there exists a positive definite symmetric matrix P 1j and P 2j , matrix Q j , constant α j >0 and parameter β j >0 makes all Satisfies the following LMI (Linear Matrix Inequality):

[0081]

[0082] Among them, Q j represents the matrix under different fuzzy rules; I represents the unit matrix of appropriate dimension; P 1j and P2j represents the positive definite symmetric matrix under different fuzzy rules; γ1, γ2 and γ3 represent constants; β j represents the parameters of the DETM to be designed under different fuzzy rules; j Represents the matrix under different fuzzy rules; Π 11j Represents the matrix π j The variable in the first row and first column of 22j Represents the matrix π j The variable in the second row and second column; specifically:

[0083]

[0084] Among them, α j represents the parameters under different fuzzy rules; σ represents a given constant;

[0085] Then, the closed-loop error system shown in formula (20) has the following H∞ performance:

[0086]

[0087] Among them, ||.|| represents the norm of the parameter; λ max (P 1j ) indicates P 1j The maximum eigenvalue of ; x(0) represents the state vector at the initial moment; λ max (P 2j ) indicates P 2j The maximum eigenvalue of x (0) represents the estimated error of the state vector at the initial moment; δ represents the steering angle of the traction AGV; w1 is the energy-bounded system modeling error disturbance; w2 is the energy-bounded measurement error disturbance. Moreover, the modeling error disturbance w1 and the measurement error disturbance w2 satisfy w1,w2∈L2[0,∞), and the steering angle δ of the traction AGV satisfies δ∈L2[0,∞). Finally, based on Theorem 1, the observation gain of the observer under different fuzzy rules can be obtained as follows: and the parameters β of the DETM to be designed under different fuzzy rules j Therefore, the H∞ robust articulation angle observer based on DETM as shown in formula (19) can achieve accurate and robust estimation of the articulation angle of the towing AGV.

[0088] Furthermore, in step 3, in the DETM of the present invention, and σ are pre-set parameters, and their reasonable selection will affect the performance indicators of the observer. First, the lower bound of the trigger interval is Secondly, the convergence speed of the observer is directly related to σ. By properly selecting σ, the decay rate can be adjusted, and the number of event triggering can also be controlled by changing the value of σ. In short, by adjusting and σ, the observer can achieve a more ideal observation performance.

[0089] Furthermore, in step 5, the observation gains under all fuzzy rules are calculated offline. Therefore, the computational complexity of the DETM-based H∞ robust joint angle observer is very small.

[0090] Furthermore, in step 5, the DETM-based H∞ robust articulation angle observer still has good robustness in the face of system modeling error disturbance, measurement error disturbance and parameter uncertainty disturbance.

[0091] The beneficial effects of the present invention are:

[0092] (1) Taking into account the parameter uncertainties of vehicle speed and trailer mass, the present invention constructs a TS fuzzy model of the traction AGV dynamics based on the TS fuzzy method to improve the adaptability of the H∞ robust articulation angle observer in the face of parameter uncertainties.

[0093] (2) The present invention proposes a continuous-time DETM between the sensor and the observer, which can effectively save limited communication resources and reduce the computational burden of the vehicle controller.

[0094] (3) This paper proposes an H∞ robust hinge angle observer based on DETM, and uses the fuzzy Lyapunov functional method and LMI theory to collaboratively design the observation gain and DETM of the observer to improve the asymptotic stability and robustness of the observer against disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is a schematic diagram of the dynamic model of the traction AGV;

[0096] Figure 2 This is a flowchart of a robust estimation method for articulation angles of traction-type AGVs for multi-vehicle collaborative operations;

[0097] Figure 3 This is the estimation effect diagram of the robust estimation method for the articulation angle of the traction-type AGV for multi-vehicle collaborative operation proposed in this invention when m2 = 6665kg;

[0098] Figure 4 This is the estimation effect diagram of the robust estimation method of the articulation angle of the traction-type AGV for multi-vehicle collaborative operation proposed in the present invention when m2 = 9000kg. DETAILED DESCRIPTION

[0099] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0100] like Figure 1-Figure 4 As shown, the present invention includes the following steps:

[0101] Step 1: Build a dynamic model of the traction AGV. The details are as follows:

[0102] Step 1.1: The traction AGV in the present invention includes a two-axle tractor and a single-axle trailer, which are articulated through a fifth wheel. The articulation is an off-axis articulation with the articulation point located behind the centerline of the tractor's rear axle. The dynamic model diagram of the traction AGV is shown in the figure below. Figure 1 First, the following assumptions are made for the dynamic model of the towing AGV: 1) the steering angle δ, tire side slip angle α, and articulation angle θ of the towing AGV are small; 2) the longitudinal velocities of the tractor and trailer are equal and both equal to v x The following degrees of freedom are considered for this dynamic model: 1) the lateral velocity v of the tractor y1 ; 2) the yaw angular velocity ω1 of the tractor; 3) the articulation angle θ between the tractor and trailer. Furthermore, because the pitch and roll motions of a towed AGV have little impact on yaw stability, they are ignored in this invention. Based on the above, the dynamic relationship for a towed AGV is as follows:

[0103]

[0104] Among them, v x represents the longitudinal speed of the traction AGV and trailer (hereinafter referred to as vehicle speed), m1 represents the mass of the tractor, m2 represents the mass of the trailer, and I z1 represents the yaw moment of inertia of the tractor, I z2 represents the trailer's yaw moment of inertia, v y1 Indicates the lateral speed of the tractor, represents the lateral acceleration of the tractor, v y2 Indicates the lateral speed of the trailer, represents the lateral acceleration of the trailer, ω1 represents the yaw angular velocity of the tractor, represents the yaw acceleration of the tractor, ω2 represents the yaw velocity of the trailer, represents the trailer's yaw angular acceleration, F yf Indicates the lateral tire force on the front axle of the tractor, F yr Indicates the lateral tire force on the rear axle of the tractor, F yt Indicates the lateral tire force on the trailer axle, F yPrepresents the lateral articulation force of the articulation point, a1 represents the distance from the front axle of the tractor to the center of mass of the tractor, a2 represents the distance from the rear axle of the tractor to the center of mass of the tractor, a3 represents the distance from the articulation point to the center of mass of the tractor, b1 represents the distance from the articulation point to the center of mass of the trailer, b2 represents the distance from the trailer axle to the center of mass of the trailer, where b1+b2=l2 represents the distance from the trailer axle to the articulation point.

[0105] Given the linear articulated coupling equations, the trailer's lateral velocity and trailer's yaw rate are obtained by the following relations:

[0106]

[0107] Where θ represents the hinge angle, represents the hinge angular velocity.

[0108] Based on formula (2), v in formula (1) y2 and ω2 can be replaced.

[0109] Step 1.2: Model the tire model of the traction AGV.

[0110] Since the lateral tire force is proportional to the tire slip angle when the tire slip angle is small, the lateral tire force acting on each axle of the traction AGV is expressed as:

[0111]

[0112] Among them, C f Indicates the tire cornering stiffness of the front axle of the tractor, C r Indicates the tire cornering stiffness of the tractor's rear axle, C t represents the tire cornering stiffness of the trailer axle, α f represents the tire slip angle of the front axle of the tractor, α r represents the tire slip angle of the rear axle of the tractor, α t Indicates the tire slip angle of the trailer axle.

[0113] In formula (3), the tire side slip angle of each axle of the traction AGV is expressed as:

[0114]

[0115] Where δ represents the steering angle of the traction AGV.

[0116] Step 1.3: Based on the contents of steps 1.1 and 1.2, express the state space equation form of the traction AGV dynamic model. Specifically:

[0117] First, define the state vector as According to formula (1)-formula (4), the state space equation form of the traction AGV dynamic model is expressed as:

[0118]

[0119] in, represents the derivative of the state vector; A represents the state matrix and B represents the input matrix, which are expressed as: Among them, A 11 、A 12 、A 13 、A 14 、A 21 、A 22 、A 23 、A 24 、A 41 、A 42 、A 43 、A 44 、B 11 、B 21 and B 41 denote the coefficients of the system matrix, which are expressed as:

[0120]

[0121] Among them, κ 11 , κ 12 , κ 13 , κ 21 , κ 22 , κ 23 and Ω represent coefficients, which are expressed as:

[0122] κ 11 =a1m1+a1m2+a3m2

[0123] κ 12 =a2m1+a2m2-a3m2

[0124] κ 13 =C f a1-C r a2-C t a3

[0125]

[0126] Then, assuming that the lateral velocity and yaw rate of the tractor in the state vector can be measured by sensors, and taking into account the system modeling error, uncertainty, and measurement error, the dynamic model of the tractor AGV is finally written as:

[0127]

[0128] Where x represents the state vector of the dynamic model of the traction AGV, w1 represents the energy-bounded system modeling error disturbance, w2 represents the energy-bounded measurement error disturbance, y represents the output vector of the dynamic model of the traction AGV, z represents the controlled output vector of the dynamic model of the traction AGV, C represents the output matrix, and D represents the controlled output matrix. The output matrix and the controlled output matrix are expressed as:

[0129] At this point, the dynamic model of the traction AGV has been established.

[0130] Step 2: Based on the TS fuzzy method and the dynamic model of the towed AGV established in step 1 as shown in formula (6), a TS fuzzy model is established to include all possible uncertain parameters to further improve the observer's ability to adapt to the parameter uncertainty of the towed AGV. The details are as follows:

[0131] Step 2.1: The speed of the towing AGV varies during its driving process. Considering the functional attributes of the towing AGV, the mass of the trailer also varies greatly. Therefore, the speed v x and the mass m2 of the trailer also represent uncertainty parameters in the dynamic model of the towing AGV. However, the vehicle speed v x and the mass m2 of the trailer are bounded and are respectively bounded by [v xmin ,v xmax ] and [m 2min ,m 2max ], where v xmin Indicates the minimum longitudinal speed of the traction AGV, v xmax Indicates the maximum longitudinal speed of the traction AGV, m 2min Indicates the minimum mass of the trailer, m 2max Indicates the maximum mass of the trailer. To this end, the present invention uses the TS fuzzy method to describe the dynamic model of the towing AGV with uncertainty parameters of vehicle speed and trailer mass. From the dynamic model of the towing AGV, it can be seen that the uncertainty parameter v x , Both m and m2 exist in the dynamic model of the traction AGV, so the definition is:

[0132]

[0133] in, and represents the reference coefficient.

[0134] Step 2.2: Uncertainty parameter v x , and m2 can be described as:

[0135]

[0136] in, and Denotes membership functions that satisfy:

[0137]

[0138] Among them, the membership function and Expressed as:

[0139]

[0140] Step 2.3: Based on steps 2.1 and 2.2, the TS fuzzy model of the towed AGV with uncertain parameters is expressed as:

[0141]

[0142] Among them, A j Represents the state matrix under different fuzzy rules, B j Represents the input matrix under different fuzzy rules, C j represents the output matrix under different fuzzy rules, D j represents the controlled output matrix under different fuzzy rules, j represents the sequence number of the fuzzy rule, represents the set of fuzzy rules, ρ j represents the fuzzy weighting function under different fuzzy rules, ρ j is represented as:

[0143]

[0144] From (12), we can see that the TS fuzzy model has 16 fuzzy rules. And, according to the fuzzy weighting function ρ under different fuzzy rules defined by formula (10) and formula (12), j The following conditions are met:

[0145]

[0146] The present invention assumes that the vehicle speed and the mass of the trailer can be accurately obtained and are constant within a given range. Therefore, the corresponding fuzzy weighting function ρ can be obtained j .

[0147] Step 3: In order to save limited CAN communication resources, the present invention introduces DETM (Dynamic Event Triggering Mechanism), so that the data measured by the sensor is transmitted through the CAN communication main line only when the event triggering condition is met. The event triggering time is defined as t k ( Where k represents the time series, represents the set of positive integers including 0), then y(t k ) represents the output signal that satisfies DETM, and y(t) represents the real-time output signal measured by the sensor. Then, DETM is constructed as:

[0148]

[0149] Where t0 represents the initial time; η(t) represents the dynamic variable; ( where R + represents a positive real number) represents the coordination factor; σ(σ∈[0,1)) represents a given constant; η0(η0∈R + ) represents known parameters; e y (t) represents the output error, which can be expressed as e y (t) = y(t k )-y(t)(t∈[t k ,t k+1 ), where t represents time, t k+1 represents the next event triggering moment); y(t) represents the real-time output signal measured by the sensor; e y (t - ) represents the output error of the left limit at time t; Indicates that when t>t k η(0) represents the value of the dynamic variable at the initial moment; represents the derivative of the dynamic variable; β (β>0) represents the parameter to be designed. Considering the TS fuzzy model, it is expressed as:

[0150]

[0151] Among them, β j represents β under different fuzzy rules.

[0152] From formula (14), we can see that when , you will get the static ETM (Event Trigger Mechanism):

[0153]

[0154] This means that the static ETM shown in formula (16) is the DETM shown in formula (14) Due to the introduction of the dynamic variable η(t), the data transmission amount under DETM can be further reduced compared with the static ETM.

[0155] According to the properties of DETM shown in formula (14), the following lemma is obtained:

[0156] Lemma 1: For any t∈[0,∞), there exists:

[0157]

[0158] η(t)>0 (18)

[0159] Step 3.1: Prove Lemma 1.

[0160] The DETM shown in formula (14) ensures that for any t∈[0,∞),

[0161]

[0162] Due to e y (t) is continuous and satisfies for any t∈[0,∞) Therefore, inequality (17) is proved. If Inequality (18) is a special case of inequality (17), so in this case η(t)>0. According to inequality (17), we can get:

[0163]

[0164] According to formula (14), we can get:

[0165]

[0166] By comparison and analysis, we can see that for any t∈[0,∞), η(t)>0. Therefore, inequality (18) is proved. The proof is complete.

[0167] Step 4: An important issue in designing DETM is that the Zone phenomenon cannot exist, that is, any two triggering times t k and t k+1 The distance between them is positive. This step will exclude the Zeno behavior of DETM. The following lemma is introduced to exclude the Zeno behavior of DETM:

[0168] Lemma 2: Assuming x, δ, w1 and w2 are bounded, then any two triggering times t k and t k+1 The distance between them is positive.

[0169] According to Lemma 2, the DETM shown in formula (14) proposed in this invention does not have Zeno behavior.

[0170] Step 4.1: Prove Lemma 2.

[0171] Notice You can get:

[0172]

[0173] in, represents y(t k )’s derivative; Indicates e y The derivative of (t); represents the derivative of y(t).

[0174] Then, since x, δ, w1 and w2 are bounded in formula (28), we can conclude that:

[0175]

[0176] where γ e represents a constant greater than zero; therefore, for all t∈[t k ,t k+1 ):

[0177]

[0178] Then, from inequality (30), we can see that

[0179]

[0180] Then, from the DETM shown in formula (14), we can get:

[0181]

[0182] Based on inequality (32), Lemma 1 and the fact e y (t k )=0, the following inequality can be obtained:

[0183]

[0184] Therefore, for any Inequality (33) implies that:

[0185]

[0186] Therefore, it can be concluded that two arbitrary triggering times t k and t k+1 The distance between them is positive. The proof is complete.

[0187] Step 5: Design a DETM-based H∞ robust joint angle observer based on the TS fuzzy model of the traction AGV shown in formula (11) obtained in step 2 and the DETM shown in formula (14) obtained in step 3. The details are as follows:

[0188] Step 5.1: First, based on the TS fuzzy model and DETM of the towed AGV, design the Luenberger state observer as shown below:

[0189]

[0190] in, represents the estimated state vector, represents the derivative of the estimated state vector, represents the estimated output vector, represents the estimated controlled output vector, L j represents the observation gain of the observer under different fuzzy rules, Represents the output vector estimated at time t, y(t k ) represents t k Output vector of trigger moment estimates.

[0191] Step 5.2: Define the estimation error and According to formula (19) and formula (11), the closed-loop error system can be obtained as follows:

[0192]

[0193] Among them, e x represents the estimation error of the state vector; represents the derivative of the estimated error of the state vector; e z represents the estimated error of the controlled output vector; e y (t) represents the output error.

[0194] Step 5.3: Considering the system modeling error disturbance, measurement error disturbance and parameter uncertainty disturbance in the TS fuzzy model of the towing AGV as shown in formula (11), in order to achieve accurate estimation of the vehicle driving state under the DETM shown in formula (11), the observation gain L of the Luenberger state observer shown in formula (19) under different fuzzy rules is calculated based on the H∞ theory. j and DETM are designed. Lemma 3 and Theorem 1 are introduced below to analyze the observation gain L under different fuzzy rules. j and DETM are designed, and Lemma 3 is a necessary condition for the proof of Theorem 1.

[0195] Lemma 3: Given real matrices E and F, there exists a positive number ε>0 satisfying the following inequality:

[0196]

[0197] Theorem 1: Given constants γ1>0, γ2>0, γ3>0 and σ∈[0,1), if there exists a positive definite symmetric matrix P 1j and P 2j , matrix Q j , constant α j >0 and parameter β j >0 makes all Satisfies the following LMI (Linear Matrix Inequality):

[0198]

[0199] Among them, Q j represents the matrix under different fuzzy rules; I represents the identity matrix of appropriate dimension; P 1j and P 2j represents the positive definite symmetric matrix under different fuzzy rules; γ1, γ2 and γ3 represent constants; β j represents the parameters of the DETM to be designed under different fuzzy rules; j Represents the matrix under different fuzzy rules; Π 11j Represents the matrix π j The variable in the first row and first column of 22j Represents the matrix π j The variable in the second row and second column; specifically:

[0200]

[0201] Among them, α j represents the parameters under different fuzzy rules; σ represents a given constant;

[0202] Then, the closed-loop error system shown in formula (20) has the following H∞ performance:

[0203]

[0204] Among them, ||.|| represents the norm of the parameter; T represents the time greater than the initial moment; λ max (P 1j ) indicates P 1j The maximum eigenvalue of ; x(0) represents the state vector at the initial moment; λ max (P 2j ) indicates P 2j The maximum eigenvalue of x(0) represents the estimated error of the state vector at the initial moment; δ represents the steering angle of the traction AGV; w1 is the energy-bounded system modeling error disturbance; w2 is the energy-bounded measurement error disturbance. Moreover, the modeling error disturbance w1 and the measurement error disturbance w2 satisfy w1,w2∈L2[0,∞), and the steering angle δ of the traction AGV satisfies δ∈L2[0,∞). Finally, based on Theorem 1, the observation gain of the observer under different fuzzy rules can be obtained as follows: and the parameters β of the DETM to be designed under different fuzzy rules j Therefore, the H∞ robust articulation angle observer based on DETM as shown in formula (19) can achieve accurate and robust estimation of the articulation angle of the towing AGV.

[0205] Step 5.4: Prove Theorem 1.

[0206] First, construct the Lyapunov function as:

[0207]

[0208] Where V represents the Lyapunov function.

[0209] Based on Lemma 1, V>0. The derivative of the Lyapunov function V with respect to time is:

[0210]

[0211] In order to establish the H∞ performance, the following performance index is introduced:

[0212]

[0213] Wherein, J represents the performance index.

[0214] Based on Lemma 3, we can get:

[0215]

[0216] Therefore, the performance index J satisfies the following conditions:

[0217]

[0218] in,

[0219]

[0220] Among them, Θ j represents the matrix; Θ 11j Represents the matrix Θ j The parameter of the first row and first column in ; Θ 22j Represents the matrix Θ jThe parameters in the second row and second column; specifically:

[0221]

[0222] Then, using Schur's complement lemma, inequality (40) can be transformed into:

[0223]

[0224] in, represents a matrix; Representation matrix The parameter of the first row and first column in; Representation matrix The parameters in the second row and second column; specifically:

[0225]

[0226] Define the matrix Q j =P 2j L j , and according to (22) we can get: but:

[0227]

[0228] Right now:

[0229]

[0230] in,

[0231]

[0232] in, represents the derivative of V; V(T) represents the value of V at time T; V(0) represents the value of V at the initial time;

[0233] Then, based on inequality (47), inequality (45) is equivalent to inequality (24), so the closed-loop error system shown in formula (20) has H∞ performance. The proof is complete.

[0234] Finally, the process of robust estimation of articulation angle of traction AGV for multi-vehicle collaborative operation is as follows: Figure 2 As shown. Then, satisfactory articulation angle estimation performance will be obtained. In addition, the robust estimation method for articulation angle of traction AGV for multi-vehicle collaborative operation proposed in this invention has the following effect on the estimation of articulation angle: Figure 3 and Figure 4As shown in the figure, the red solid line represents the true value of the articulation angle, the blue dotted line represents the estimation result of the robust estimation method for the articulation angle of the traction-type AGV for multi-vehicle collaborative operation proposed in this invention, and the green dotted line represents the estimation result of the traditional KF (Kalman Filter) estimation method. Figure 3 It can be seen from the figure that when the mass of the trailer is 6665 kg, the estimation result of the robust estimation method for the articulation angle of towed AGVs for multi-vehicle collaborative operation proposed in this invention can track the true value well. Even in the face of different vehicle speeds, the estimation error is very small, and the estimation error is much smaller than that of the KF estimation method. Figure 4 The estimation results for a trailer mass of 9000kg are shown. As can be seen from the figure, even in the face of variations in trailer mass, the robust estimation method for the articulation angle of towed AGVs for multi-vehicle collaborative operations proposed in this invention still achieves excellent estimation results and has significant advantages over the KF estimation method. In summary, the robust estimation method for the articulation angle of towed AGVs for multi-vehicle collaborative operations proposed in this invention achieves excellent estimation results and robustness even in the face of parameter uncertainties in vehicle speed and trailer mass, laying the foundation for the reliability and stability of multi-vehicle collaborative operations.

[0235] The above disclosure is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A robust estimation method for articulation angles of traction-type AGVs for multi-vehicle collaborative operation, characterized by: The steps of the robust estimation method for the articulation angle of the traction-type AGV are as follows: Step 1: For the time-varying longitudinal velocity of the towing AGV and the trailer mass, the TS fuzzy method is used to build the dynamic model of the towing AGV; Step 2: Based on the TS fuzzy method and the dynamic model of the towed AGV established in step 1, a TS fuzzy model is established to include all possible uncertain parameters and improve the observer's ability to adapt to the parameter uncertainty of the towed AGV; Step 3: Introduce the dynamic event trigger mechanism DETM, so that the data measured by the sensor is transmitted through the CAN communication main line only when the event trigger condition is met; Step 4: Eliminate the Zeno behavior of DETM; the absence of the Zone phenomenon refers to the absence of any two triggering moments t k and t k+1 The distance between them is positive; Step 5: Design a DETM-based H∞ robust articulation angle observer based on the TS fuzzy model of the towed AGV obtained in step 2 and the DETM obtained in step 3 to estimate the articulation angle of the towed AGV.

2. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 1 is characterized in that: The step 1 is specifically as follows: Step 1.1: First, make the following assumptions for the dynamic model of the towing AGV: 1) The steering angle δ, tire slip angle α, and articulation angle θ of the towing AGV are small; 2) Assume that the longitudinal velocities of the tractor and trailer are equal and both equal to v x The following degrees of freedom are considered for this dynamic model: 1) the lateral velocity v of the tractor y1 ; 2) the yaw angular velocity ω1 of the tractor; 3) the articulation angle θ between the tractor and trailer; and the pitch and roll motions of the towed AGV are ignored; the dynamic relationship of the towed AGV is as follows: Among them, v x Indicates the longitudinal speed of the traction AGV and trailer, referred to as vehicle speed; m1 indicates the mass of the tractor; m2 indicates the mass of the trailer; I z1 Indicates the yaw moment of inertia of the tractor; I z2 represents the trailer's yaw moment of inertia; v y1 Indicates the lateral speed of the tractor; represents the lateral acceleration of the tractor; v y2 Indicates the lateral speed of the trailer; represents the lateral acceleration of the trailer; ω1 represents the yaw angular velocity of the tractor; represents the yaw acceleration of the tractor; ω2 represents the yaw velocity of the trailer; Indicates the trailer's yaw angular acceleration; F yf Indicates the lateral tire force on the front axle of the tractor; F yr Indicates the lateral tire force on the rear axle of the tractor; F yt Indicates the lateral tire force on the trailer axle; F yP represents the lateral articulation force of the articulation point; a1 represents the distance from the front axle of the tractor to the center of mass of the tractor; a2 represents the distance from the rear axle of the tractor to the center of mass of the tractor; a3 represents the distance from the articulation point to the center of mass of the tractor; b1 represents the distance from the articulation point to the center of mass of the trailer; b2 represents the distance from the trailer axle to the center of mass of the trailer, where b1+b2=l2 represents the distance from the trailer axle to the articulation point; Step 1.2: Model the tire model of the traction AGV; The lateral tire force acting on each axle of the traction AGV is expressed as: Among them, C f Indicates the tire cornering stiffness of the front axle of the tractor, C r Indicates the tire cornering stiffness of the tractor's rear axle, C t represents the tire cornering stiffness of the trailer axle, α f represents the tire slip angle of the front axle of the tractor, α r represents the tire slip angle of the rear axle of the tractor, α t Indicates the tire slip angle of the trailer axle; Step 1.3: Based on the contents of steps 1.1 and 1.2, express the state space equation form of the traction AGV dynamic model; specifically: First, define the state vector as The state space equation form of the traction AGV dynamic model is expressed as: in, Represents the derivative of the state vector; A represents the state matrix, B represents the input matrix; Then, assuming that the lateral velocity and yaw rate of the tractor in the state vector can be measured by sensors, considering the system modeling error, uncertainty and measurement error, the dynamic model of the tractor AGV is: Where x represents the state vector of the dynamic model of the traction AGV, w1 represents the energy-bounded system modeling error disturbance, w2 represents the energy-bounded measurement error disturbance, y represents the output vector of the dynamic model of the traction AGV, z represents the controlled output vector of the dynamic model of the traction AGV, C represents the output matrix, and D represents the controlled output matrix. The output matrix and the controlled output matrix are expressed as: At this point, the dynamic model of the traction AGV has been established.

3. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 1 is characterized in that: In step 1: In step 1.1, given the linear articulated coupling equation, the trailer's lateral velocity and trailer's yaw rate are obtained by the following relationship: Where θ represents the hinge angle, represents the articulation angular velocity; based on formula (2), v in formula (1) y2 and ω2 can be replaced; In formula (3) of step 1.2, the tire side slip angle of each axle of the traction AGV is expressed as: Where, δ represents the steering angle of the traction AGV; In step 1.3, A and B are expressed as: Among them, A 11 , A 12 , A 13 , A 14 , A 21 , A 22 , A 23 , A 24 , A 41 , A 42 , A 43 , A 44 , B 11 , B 21 Japanese B 41 Display system square series.

4. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 2 is characterized in that: The step 2 is specifically as follows: Step 2.1: When the traction AGV is driving, the speed v x and the mass m2 of the trailer are bounded and are respectively bounded by [v xmin ,v xmax ] and [m 2min ,m 2max ], where v xmin Indicates the minimum longitudinal speed of the traction AGV, v xmax Indicates the maximum longitudinal speed of the traction AGV, m 2min Indicates the minimum mass of the trailer, m 2max represents the maximum mass of the trailer; the TS fuzzy method is used to describe the dynamic model of the towing AGV with uncertainty parameters of vehicle speed and trailer mass; from the dynamic model of the towing AGV, it can be seen that the uncertainty parameter v x , Both m and m2 exist in the dynamic model of the traction AGV, so the definition is: in, and represents the reference coefficient; Step 2.2: Uncertainty parameter v x , and m2 is described as: in, and represents the membership function; Step 2.3: Based on steps 2.1 and 2.2, the TS fuzzy model of the towed AGV with uncertain parameters is expressed as: Among them, A j Represents the state matrix under different fuzzy rules, B j represents the input matrix under different fuzzy rules, C j represents the output matrix under different fuzzy rules, D j represents the controlled output matrix under different fuzzy rules, j represents the sequence number of the fuzzy rule, represents the set of fuzzy rules, ρ j Represents the fuzzy weighting function under different fuzzy rules.

5. The robust estimation method of articulation angle of traction-type AGV for multi-vehicle collaborative operation according to claim 4 is characterized in that: In step 2: In step 2.1, and satisfy: Among them, the membership function and Expressed as: In step 2.3, ρ j is represented as: From (12), we can see that the TS fuzzy model has 16 fuzzy rules; and according to the fuzzy weighting function ρ under different fuzzy rules defined by formula (10) and formula (12), j The following conditions are met:

6. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 4 is characterized in that: The step 3 is specifically as follows: Define the event trigger time as t k ,in k represents the time series, represents the set of positive integers including 0, y(t k ) represents the output signal that satisfies DETM, and y(t) represents the real-time output signal measured by the sensor; DETM is: Where t0 represents the initial time; η(t) represents the dynamic variable; represents the coordination factor, R + Represents a positive real number; σ represents a given constant; η0 represents a known parameter, η0∈R + ;e y (t) represents the output error; y(t) represents the real-time output signal measured by the sensor; e y (t - ) represents the output error of the left limit at time t; Indicates that when t>t k η(0) represents the value of the dynamic variable at the initial moment; represents the derivative of the dynamic variable; β represents the parameter to be designed, β>0; From formula (14), we can see that when When , we will get the static event trigger mechanism ETM:

7. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 6 is characterized in that: In step 3: In the formula (14) of step 3: the output error e y (t) is expressed as e y (t) = y(t k )-y(t)(t∈[t k ,t k+1 ), where t represents time, t k+1 Indicates the next event triggering time; the σ∈[0,1); According to formula (14), we can get the following lemma: Lemma 1: For any t∈[0,∞), there exists: η(t)>0 (18).

8. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 7 is characterized in that: In step 4, the following lemma is introduced to exclude the Zeno behavior of DETM: Lemma 2: Assuming x, δ, w1 and w2 are bounded, then any two triggering times t k and t k+1 The distance between them is positive; according to Lemma 2, the DETM shown in the proposed formula (14) does not have Zeno behavior.

9. The robust estimation method for articulation angle of traction-type AGVs for multi-vehicle collaborative operation according to claim 8 is characterized in that: In step 5, a DETM-based H∞ robust joint angle observer is designed based on the TS fuzzy model of the traction AGV shown in formula (11) obtained in step 2 and the DETM shown in formula (14) obtained in step 3; specifically, as follows: Step 5.1: First, based on the TS fuzzy model and DETM of the towed AGV, design the Luenberger state observer as shown below: in, represents the estimated state vector, represents the derivative of the estimated state vector, represents the estimated output vector, represents the estimated controlled output vector, L j represents the observation gain of the observer under different fuzzy rules, Represents the output vector estimated at time t, y(t k ) represents t k Output vector of trigger moment estimation; Step 5.2: Define the estimation error and According to formula (19) and formula (11), the closed-loop error system is: Among them, e x represents the estimation error of the state vector; represents the derivative of the estimated error of the state vector; e z represents the estimated error of the controlled output vector; e y (t) represents the output error; Step 5.3: To achieve accurate estimation of the vehicle driving state under the DETM shown in formula (11), the observation gain L of the Luenberger state observer shown in formula (19) under different fuzzy rules is calculated based on the H∞ theory. j and DETM are designed; Lemma 3 and Theorem 1 are introduced to analyze the observation gain L under different fuzzy rules. j Design with DETM; Lemma 3: Given real matrices E and F, there exists a positive number ε>0 satisfying the following inequality: Theorem 1: Given constants γ1>0, γ2>0, γ3>0 and σ∈[0,1), if there exists a positive definite symmetric matrix P 1j and P 2j , matrix Q j , constant α j >0 and parameter β j >0 makes all If the linear matrix inequality LMI is satisfied, the closed-loop error system shown in formula (20) has the following H∞ performance: Among them, ||.|| represents the norm of the parameter; λ max (P 1j ) indicates P 1j The maximum eigenvalue of ; x(0) represents the state vector at the initial moment; λ max (P 2j ) indicates P 2j The maximum eigenvalue of x (0) represents the estimation error of the state vector at the initial moment; δ represents the steering angle of the traction AGV; w1 is the system modeling error perturbation with bounded energy; w2 represents the measurement error perturbation with bounded energy; and the modeling error perturbation w1 and the measurement error perturbation w2 satisfy w1,w2∈L2[0,∞), and the steering angle δ of the traction AGV satisfies δ∈L2[0,∞); Finally, based on Theorem 1, the observation gain of the observer under different fuzzy rules is obtained as follows: and the parameters β of the DETM to be designed under different fuzzy rules j Therefore, the H∞ robust articulation angle observer based on DETM as shown in formula (19) can achieve accurate and robust estimation of the articulation angle of the towing AGV.

10. The robust estimation method of articulation angle of traction-type AGV for multi-vehicle collaborative operation according to claim 9 is characterized in that: The linear matrix inequality LMI is: Among them, Q j represents the matrix under different fuzzy rules; I represents the unit matrix of appropriate dimension; P 1j and P 2j represents the positive definite symmetric matrix under different fuzzy rules; γ1, γ2 and γ3 represent constants; β j represents the parameters of the DETM to be designed under different fuzzy rules; j Represents the matrix under different fuzzy rules; Π 11j Represents the matrix π j The variable in the first row and first column of 22j Represents the matrix π j The variable in the second row and second column; specifically: Among them, α j represents the parameters under different fuzzy rules; σ represents a given constant.

Images