Distributed fuzzy fault detection method related to transmission of multi-unmanned ship system

Abstract

The invention aims to provide a distributed fuzzy fault detection method related to transmission of a multi-unmanned ship system, and belongs to the technical field of ship control. According to the fault detection method, for a multi-unmanned-ship system which may have actuator faults, interference caused by external factors is considered, and a T-S fuzzy model of an unmanned ship is established; and then communication topology and transmission delay between the unmanned ships are considered, and a special Lyapunov function is introduced, so that more variable matrixes are introduced, the conservative property of matrix inequality solution is reduced, namely, the possibility of finding a solution is increased, and the stability and effectiveness of a fault detection system are ensured. The problem that a matrix inequality is difficult to solve due to a coupling term brought by information transmission between unmanned ships is solved, and a distributed fault detection and judgment mechanism is established. According to the method, the internal nonlinearity of the unmanned ship and the interference of the environment where the unmanned ship is located are comprehensively considered, and fault detection under the adverse effects that the working environment is complex, and long-time working is needed is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of ship control, and in particular relates to a distributed fuzzy fault detection method related to transmission of a multi-unmanned ship system. Background Art

[0002] Inspired by many interesting clustering behaviors in nature (such as fish aggregation and bird migration), multi-agent systems have emerged with the rapid development of communication technology and automation technology. As an important tool for exploring marine resources, unmanned ships naturally become one of the important research objects of multi-agents. However, since unmanned ships are easily affected by harsh natural conditions and long working hours, they are prone to failure when performing tasks.

[0003] Judging from the existing research results, the relevant technical solutions mainly focus on the fault-tolerant and intrusion-tolerant control methods of unmanned ships. [1][2] There are few studies on fault detection. Of course, there are already some fault detection schemes based on fuzzy systems, such as [3][4] The fault detection filtering problem of nonlinear dynamic systems under the TS fuzzy framework was discussed in

[11] , but these studies only focus on single agents. However, in the process of multiple unmanned ships working together, if one or even several unmanned ships fail, the fault information is likely to be transmitted to the unmanned ships that have not failed. Therefore, if the unmanned ship system fails and is not discovered and eliminated in time, it is likely to cause system failure and mission failure, resulting in huge economic losses. In addition, since the unmanned ship receives information from neighboring nodes, the state information between the unmanned ships generates coupling terms, so the fault detection method of a single unmanned ship is not suitable for distributed fault detection. Therefore, it is of great significance to study the fault detection algorithm of multi-unmanned ship systems.

[0004] [1] Jiao Yuhang. Research on decision-making and control of unmanned ship swarm[D]. Dalian Maritime University, 2022. DOI: 10.26989 / d.cnki.gdlhu.2022.000667.

[0005] [2] Zhang Ao. Fault-tolerant control of unmanned boats with actuator failures based on fuzzy logic system[J]. Journal of Shenyang Normal University (Natural Science Edition), 2020, 38(03): 214-219.

[0006] [3]Y.Wen,X.Ye and

[0007] [4] Q. Liu, Y. Long, T. Li, JH Park and CLP Chen, "Fault Detection for Unmanned Marine Vehicles Under Replay Attack," in IEEE Transactions on FuzzySystems, pp.1716-1728. Summary of the invention

[0008] In view of the problems existing in the background technology, the purpose of the present invention is to provide a distributed fuzzy fault detection method related to the transmission of a multi-unmanned ship system. This fault detection method is aimed at a multi-unmanned ship system where actuator failures may occur, and takes into account the interference caused by external factors such as wind and waves, and establishes a TS fuzzy model of the unmanned ship; then the communication topology and transmission delay between the unmanned ships are considered, and a special Lyapunov function is introduced, thereby introducing more variable matrices, reducing the conservatism of solving matrix inequalities, that is, increasing the possibility of finding a solution, ensuring the stability and effectiveness of the fault detection system, dealing with the problem that the matrix inequality caused by the coupling terms brought about by the information transmission between the unmanned ships is difficult to solve, and establishing a distributed fault detection judgment mechanism. The method of the present invention comprehensively considers the internal nonlinearity of the unmanned ship and the interference of the environment in which it is located, and realizes the detection of faults under the adverse influence of complex working environment and long working time.

[0009] To achieve the above object, the technical solution of the present invention is as follows:

[0010] A distributed fuzzy fault detection method for transmission-related multi-unmanned ship systems comprises the following steps:

[0011] Step 1: Based on the actual operation of unmanned ships, determine the dynamic model of a single unmanned ship, including kinematic equations and dynamic equations;

[0012] Step 2: According to the dynamic model of a single unmanned ship in step 1, determine the state space model, consider the communication topology and transmission delay between unmanned ships, and obtain the state equation and measurement equation of the unmanned ship under the TS fuzzy system through the fuzzy modeling method;

[0013] Step 3: According to the state equation of a single unmanned ship under the TS fuzzy system obtained in step 2, a corresponding distributed fault detection filter is constructed and the residual is obtained;

[0014] Step 4: In order to improve the performance of the distributed fault detection method, a known fault weight matrix is ​​introduced;

[0015] Step 5: Substitute the measured output equation of the unmanned ship into the distributed fault detection filter equation, and consider the communication topology between the unmanned ships to obtain the global fault detection system equation;

[0016] Step 6: Design a method to calculate the filter gain, solve the gain matrix of the fault detection filter constructed in step 3, and make the global fault detection system equation have a specified H for disturbance ∞ performance;

[0017] Step 7: Based on the gain matrix of the fault detection filter obtained in step 6 and the residual constructed in step 3, design the residual evaluation function J r (t);

[0018] Step 8: According to actual needs, based on the residual evaluation function obtained in step 6, a threshold and alarm strategy are formulated, that is, if the residual evaluation function value obtained by real-time detection is greater than the pre-set residual evaluation function threshold, an alarm is issued; otherwise, no alarm is issued, thereby completing fault detection.

[0019] Furthermore, the specific process of step 1 is:

[0020] The dynamic model of the mth unmanned ship includes kinematic equations and dynamic equations, specifically,

[0021] Kinematic equations:

[0022] Kinetic equation:

[0023] Among them, φ m (t) is the position information of the unmanned ship in the earth coordinate system, is the heading angle information, v m (t) is the ship's own information, matrix M m ,N m ,Z m Represent the inertia matrix, damping matrix and mooring force matrix of the ship respectively, μ m (t) is the control signal generated by the controller, d m (t) is the interference signal caused by external factors such as the environment, represents the derivative, and J(·) is the transformation matrix from the UAV body coordinate system to the ground coordinate system.

[0024] Furthermore, in step 1, the position information φ of the unmanned ship m (t) contains coordinate information (x mp (t),y mp (t)) and heading angle information Ship's own information m (t) includes the sway speed v of the unmanned ship m1 (t), sway speed v m2 (t) and the bow speed v m3 (t), v m (t) = col{v m2 (t),v m2 (t),v m2 (t)}.

[0025] Furthermore, the specific process of step 2 is:

[0026] make The kinetic equation can be expressed as:

[0027]

[0028] Among them, E 1m is the known failure coefficient matrix, f m (t) is the fault signal occurring on the unmanned ship;

[0029] The state space model of the unmanned ship system includes the state equation and the measurement output equation, which are:

[0030]

[0031] y m (t) = C m x m (t)

[0032] Among them, C m is the coefficient matrix, A m , D m and E m is the augmented matrix;

[0033]

[0034] The fuzzy model of the unmanned ship is obtained through the fuzzy modeling method:

[0035] Plant Rule i:If p m1 (δ m (t))is and p m2 (δ m(t))is THEN

[0036]

[0037] Among them, i is the i-th fuzzy subsystem, d m (t) is external interference, τ mn (t) represents the delay of information transmission from the nth unmanned ship to the mth unmanned ship, and its upper bound is h. The matrix A mi ,E mi ,D mi ,K mn are all known coefficient matrices, y(t) is the measurement equation of the system, is the set of all neighboring unmanned ships of the mth unmanned ship, is the filter state on the neighboring node of the unmanned ship m;

[0038] Then the state equation and measurement equation of the global TS fuzzy system of the unmanned ship are:

[0039]

[0040] Among them, ρ i (δ m (t)) is the membership function corresponding to different fuzzy rules, λ i is the membership function after normalization, δ m (t) is the prerequisite variable of the fuzzy rule.

[0041] Furthermore, in step 2, That is, δ m (t) Membership function The sum of is 1, and

[0042] Furthermore, the specific form of the distributed fuzzy fault detection filter constructed in step 3 is:

[0043]

[0044] in, r m (t) is the state vector, measurement output vector and residual signal of the distributed fuzzy fault detection filter, is the normalized membership function of the fault detection filter; the matrix is the gain matrix of the fault detection filter to be designed, l mn is the element in the adjacency matrix, j is the sequence number of the filter fuzzy rule, and s is the number of filter fuzzy rules.

[0045] Furthermore, in step 4, a known fault weight matrix is ​​introduced so that

[0046]

[0047] Among them, A wm ,B wm ,C wm is a known matrix.

[0048] Furthermore, the global fault detection system equation constructed in step 5 is:

[0049]

[0050] in,

[0051]

[0052] is the Kronecker product and ⊙ is the Hadamard product.

[0053] Furthermore, the specific process of step 6 is:

[0054] (1) Based on the design principle of stability and robustness to disturbances, it is necessary to ensure that the error augmentation system is asymptotically stable and has the specified H ∞ Performance γ, so that for the disturbance signal ω(t), the following equation holds:

[0055]

[0056] γ is a performance indicator to measure the disturbance attenuation capability, and the superscript T indicates the transition rank;

[0057] (2) When H ∞ When a γ-distributed fuzzy fault detection filter with good performance exists, the matrix containing its gain information is directly obtained through the formulated solution conditions. And the correlation matrix L2, through the following operation, the gain matrix of the fault detection filter is obtained

[0058]

[0059] Furthermore, the residual evaluation function R in step 7 m (r m (t)) The specific form is:

[0060]

[0061] Among them, t is the detection time and t0 is the initial time.

[0062] Furthermore, the specific process of fault detection in step 8 is as follows:

[0063] Set the threshold form of the residual evaluation function:

[0064]

[0065] The strategy of distributed fuzzy fault detection of multi-unmanned ship system is: if the residual evaluation function value obtained by real-time detection is greater than the threshold of the residual evaluation function, an alarm is given; otherwise, no alarm is given; its expression is:

[0066]

[0067] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0068] In terms of system modeling, the fault detection method of the present invention utilizes the TS fuzzy system to process the nonlinear terms in the coordinate transformation matrix, takes into account the communication topology between unmanned ships and the inevitable delay phenomenon during information transmission, and develops a multi-unmanned ship system under the TS fuzzy model framework. At the same time, in terms of system security, external interference is considered, and a distributed fuzzy fault detection filter is designed that can detect actuator failures occurring on the unmanned ships of the current node and neighboring nodes. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 is a flow chart of the fault detection method of the present invention.

[0070] Figure 2 The topological structure of the unmanned ship.

[0071] Figure 3 is the residual signal generated by the four filters.

[0072] Figure 4 is the residual evaluation function corresponding to the four residual signals. DETAILED DESCRIPTION

[0073] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the implementation modes and the accompanying drawings.

[0074] The present invention discloses a distributed fuzzy fault detection method for multi-unmanned ship system transmission related, the flow chart of the method is as follows Figure 1 As shown, the specific steps include:

[0075] Step 1: The dynamic model of the mth unmanned ship is as follows, including kinematic equations and dynamic equations.

[0076] Kinematic equations:

[0077] Kinetic equation:

[0078] Among them, φm (t) is the position information of the unmanned ship in the earth coordinate system, is the heading angle information, v m (t) is the ship's own information, matrix M m ,N m ,Z m Represent the inertia matrix, damping matrix and mooring force matrix of the ship respectively. The position information of the unmanned ship φ m (t) contains coordinate information (x mp (t),y mp (t)) and heading angle information Ship's own information m (t) includes the sway speed v of the unmanned ship m1 (t), sway speed v m2 (t) and the bow speed v m3 (t), v m (t) = col{v m2 (t),v m2 (t),v m2 (t)}.

[0079] Step 2: Order The kinetic equation can be expressed as:

[0080]

[0081] Among them, E m is the known failure coefficient matrix;

[0082] Furthermore, the state equation and measurement output equation of the unmanned ship system are obtained:

[0083]

[0084] y m (t) = C m x m (t)

[0085] in,

[0086]

[0087] The fuzzy model of the unmanned ship is obtained through the fuzzy modeling method:

[0088] Plant Rule i:If p m1 (δ m (t))is and p m2 (δ m (t))is THEN

[0089]

[0090] Among them, i is the i-th fuzzy subsystem, d m (t) is external interference, τ mn (t) represents the delay of information transmission from the nth unmanned ship to the mth unmanned ship, and its upper bound is h. The matrix A mi ,E mi ,D mi ,K mn are all known coefficient matrices; y(t) is the measurement equation of the system;

[0091] Then the state equation and measurement equation of the global TS fuzzy system of the unmanned ship are:

[0092]

[0093] Among them, ρ i (δ m (t)) is the membership function corresponding to different fuzzy rules, That is, δ m (t) Membership function The sum of is 1, and

[0094] The purpose of introducing the TS fuzzy system is to approximate the nonlinearity in the unmanned ship system with multiple linear subsystems, so as to deal with the unmanned ship system by the method of dealing with linear systems; the nonlinear terms in the coordinate transformation matrix are processed by the sector nonlinear method, and the specific expression of the membership function is obtained. In the subsequent analysis, only λ is considered. i (δ m (t))≥0 and The property of , the solution of the filter gain problem in the subsequent steps will not involve the true value of the specific membership function;

[0095] Step 3: The specific form of the constructed distributed fuzzy fault detection filter is:

[0096]

[0097] in, r m (t) is the state vector, measurement output vector and residual signal of the distributed fuzzy fault detection filter, is the normalized membership function of the fault detection filter; the matrix is the gain matrix of the fault detection filter to be designed;

[0098] Step 4: Introduce the fault weight matrix to improve the performance of the fault detection system. Its specific form is:

[0099]

[0100] Among them, A wm ,B wm ,C wm is a known matrix;

[0101] Selecting different parameters in this step will make the residual have different performance. Selecting appropriate parameters will make the constructed residual sensitive to faults and robust to interference. Generally, they are selected based on the experience of experts.

[0102] Step 5: Comprehensively consider the unmanned ship fuzzy state equation, filter dynamic equation and fault weight matrix to obtain the global fault detection system equation, which is in the form of:

[0103]

[0104] in,

[0105]

[0106] It can be seen that compared with a single unmanned ship system, the distributed multi-unmanned ship system produces This brings great difficulties to solving the constructed matrix inequality in step 6. In the subsequent process, different decoupling methods are adopted to decompose the matrix P into P = diag{L1,L2,T},

[0107] Step 6: Design a solution calculation method for the filter gain, solve and obtain the gain matrix of the fault detection filter constructed in step 4, and make the residual constructed in step 5 robust to disturbances and sensitive to faults. The specific process is as follows:

[0108] (1) Based on the design principle of stability and robustness to disturbances, it is necessary to ensure that the error augmentation system is asymptotically stable and has the specified H ∞ Performance γ, so that for the disturbance signal ω(t), the following equation holds:

[0109]

[0110] (2) When H ∞ When a γ-distributed fuzzy fault detection filter with good performance exists, the matrix containing its gain information is directly obtained through the formulated solution conditions. And the correlation matrix L2, through the following operation, the gain matrix of the fault detection filter is obtained

[0111]

[0112] Since τ mn (t) is a time-varying function, so the above global fault detection system equation can be regarded as an infinite-dimensional system. Therefore, the conventional Lyapunov stability criterion cannot be used to judge the stability of the system;

[0113] The present invention adopts the following form of Lyapunov function as the stability criterion:

[0114]

[0115] If we do not consider the time-varying delay τ mn (t), the Lyapunov function can be chosen as

[0116] Therefore, the distributed fuzzy fault detection filter gain condition is obtained: ∞ A sufficient condition for the existence of a fault detection filter with performance γ is that if there exist suitable symmetric matrices L1>0, L2>0, T>0, Q1>0, Q2>0, R1>0, R2>0, Z>0, X≥0 and several scalars γ>0, 0≤τ(t)≤h, then the resulting distributed fault detection system is asymptotically stable and has H if the following conditions are met ∞ Performance Indicator γ:

[0117]

[0118]

[0119] ψ 121 =R1-S1-P1+P2

[0120] ψ 77 =h -2 (R2-2L2)

[0121]

[0122] Π 33 =-R1-Q1,Π 44 =-R2-Q2,Π 55 =-γ 2 I

[0123] ψ 56 =[DE] T (R1+Z),ψ 66 =-h -2 (R1+Z),ψ 88 =-I

[0124] Through the formulated solution conditions, the matrix containing its gain information is directly obtained And the correlation matrix L2, through the following operation, the gain matrix of the fault detection filter is obtained

[0125]

[0126] For a single system, the variables to be solved in the above matrix inequality are all 6-dimensional matrices, but for a multi-agent system, the above matrix is ​​6×N-dimensional, and each element in the matrix to be solved is an unknown variable. Therefore, as the dimension of the matrix to be solved increases, the number of unknown variables also increases exponentially, making it more difficult to implement a distributed fault detection solution than to implement a single unmanned ship fault detection solution. This further requires that the matrix P needs to have a special structure;

[0127] Step 7: Based on the gain matrix of the fault detection filter obtained in step 6 and the residual constructed in step 3, design the residual evaluation function R m (r m (t)), the specific form is as follows:

[0128]

[0129] Among them, t is the detection duration, t0 is the initial time;

[0130] Step 8: According to actual needs, formulate thresholds and alarm strategies, that is, if the residual evaluation function value obtained by real-time detection is greater than the preset residual evaluation function threshold, an alarm is triggered; otherwise, no alarm is triggered, thereby completing fault detection; the specific process is as follows:

[0131] Set the threshold form of the residual evaluation function:

[0132]

[0133] Where t is the detection time;

[0134] Therefore, the fault detection strategy of the networked unmanned ship system is: if the residual evaluation function value obtained by real-time detection is greater than the threshold of the residual evaluation function, an alarm is given; otherwise, no alarm is given; its expression is:

[0135]

[0136] Example 1

[0137] This method is used for fault detection. In the simulation process, it is assumed that there are 4 unmanned ships, and the topological structure of the unmanned ships is as follows: Figure 2 As shown, each unmanned ship receives the same interference. Assuming that the interference signal is in the form of,

[0138] dm1 (t) = 28sin(1.28t)e (-0.3t) ,(0<t<6),

[0139] d m2 (t) = -30sin(1.5t)e (-1.3t) ,d m3 (t) = 24sin(1.24t)e (-0.5t) ,(0<t<5.5).

[0140] Assume that only the third unmanned ship fails, the fault signal is: f3(t)=20sin(t-14), (14<t<20), and the upper limit of the transmission delay is 0.3s.

[0141] Figure 3 The dotted line indicates the situation without fault, and the solid line indicates the situation with fault. It can be seen that between 14s and 20s, the red solid line fluctuates significantly compared with the blue dotted line, so the fault signal will have a certain impact on the system.

[0142] Figure 4 It shows the residual evaluation function, threshold, and detection effect corresponding to the four filters. It can be seen that the first filter failed to detect whether the system had a fault, the second filter detected a fault at 14.46 seconds, the third filter detected a fault at 14.12 seconds, and the fourth filter detected a fault at 14.58 seconds. Therefore, the system detected a fault at 14.12 seconds and triggered an alarm.

[0143] The above description is only a specific implementation mode of the present invention. Any feature disclosed in this specification, unless otherwise stated, can be replaced by other alternative features that are equivalent or have similar purposes; all the disclosed features, or all the steps in the methods or processes, except for mutually exclusive features and / or steps, can be combined in any way.

Claims

1. A distributed fuzzy fault detection method for transmission-related multi-unmanned ship systems, characterized in that: The following steps are involved: Step 1: Based on the actual operation of unmanned ships, determine the dynamic model of a single unmanned ship, including kinematic equations and dynamic equations; Step 2: According to the dynamic model of a single unmanned ship in step 1, determine the state space model, consider the communication topology and transmission delay between unmanned ships, and obtain the state equation and measurement equation of the unmanned ship under the TS fuzzy system through the fuzzy modeling method; Step 3: According to the state equation of a single unmanned ship under the TS fuzzy system obtained in step 2, a corresponding distributed fault detection filter is constructed and the residual is obtained; Step 4: In order to improve the performance of the distributed fault detection method, a known fault weight matrix is ​​introduced; Step 5: Substitute the measured output equation of the unmanned ship into the distributed fault detection filter equation, and consider the communication topology between the unmanned ships to obtain the global fault detection system equation; Step 6: Design a method to calculate the filter gain, solve the gain matrix of the fault detection filter constructed in step 3, and make the global fault detection system equation have a specified H for disturbance ∞ performance; Step 7: Based on the gain matrix of the fault detection filter obtained in step 6 and the residual constructed in step 3, design the residual evaluation function J r (t); Step 8: According to actual needs, based on the residual evaluation function obtained in step 6, a threshold and alarm strategy are formulated, that is, if the residual evaluation function value obtained by real-time detection is greater than the pre-set residual evaluation function threshold, an alarm is issued; otherwise, no alarm is issued, thereby completing fault detection.

2. The distributed fuzzy fault detection method according to claim 1, characterized in that: The specific process of step 1 is: The dynamic model of the mth unmanned ship includes kinematic equations and dynamic equations, specifically, Kinematic equations: Kinetic equation: Among them, φ m (t) is the position information of the unmanned ship in the earth coordinate system, is the heading angle information, v m (t) is the ship's own information, the matrices Mm, Nm, and Zm represent the ship's inertia matrix, damping matrix, and mooring force matrix, respectively, and μ m (t) is the control signal generated by the controller, d m (t) is the interference signal caused by external factors such as the environment, represents the derivative, J(·) is the transformation matrix from the unmanned ship body coordinate system to the ground coordinate system; the position information of the unmanned ship φ m (t) contains coordinate information (x mp (t),y mp (t)) and heading angle information Ship's own information m (t) includes the sway speed v of the unmanned ship m1 (t), sway speed v m2 (t) and the bow speed v m3 (t), v m (t) = col{v m2 (t),v m2 (t),v m2 (t)}.

3. The distributed fuzzy fault detection method according to claim 2, characterized in that: The specific process of step 2 is: make The kinetic equation can be expressed as: Among them, E 1m is the known failure coefficient matrix, f m (t) is the fault signal occurring on the unmanned ship; The state space model of the unmanned ship system includes the state equation and the measurement output equation, which are: y m (t)=C m x m (t) Among them, C m is the coefficient matrix, A m , D m and E m is the augmented matrix; The fuzzy model of the unmanned ship is obtained through the fuzzy modeling method: Among them, i is the i-th fuzzy subsystem, d m (t) is external interference, τ mn (t) represents the delay of information transmission from the nth unmanned ship to the mth unmanned ship, and its upper bound is h. The matrix A mi ,E mi ,D mi ,K mn are all known coefficient matrices, y(t) is the measurement equation of the system, is the set of all neighboring unmanned ships of the mth unmanned ship, is the filter state on the neighboring node of the unmanned ship m; Then the state equation and measurement equation of the global TS fuzzy system of the unmanned ship are: Among them, ρ i (δ m (t)) is the membership function corresponding to different fuzzy rules, λi is the membership function after normalization, δ m (t) is the prerequisite variable of the fuzzy rule.

4. The distributed fuzzy fault detection method according to claim 3, characterized in that: In step 2, That is, δ m (t) Membership function The sum of is 1, and 5. The distributed fuzzy fault detection method according to claim 3, characterized in that: The specific form of the distributed fuzzy fault detection filter constructed in step 3 is: in, r m (t) is the state vector, measurement output vector and residual signal of the distributed fuzzy fault detection filter, is the normalized membership function of the fault detection filter; the matrix is the gain matrix of the fault detection filter to be designed, l mn is the element in the adjacency matrix, j is the sequence number of the filter fuzzy rule, and s is the number of filter fuzzy rules.

6. The distributed fuzzy fault detection method according to claim 3, characterized in that: In step 4, the known fault weight matrix is ​​introduced so that Among them, A wm ,B wm ,C wm is a known matrix.

7. The distributed fuzzy fault detection method according to claim 6, characterized in that: The global fault detection system equation constructed in step 5 is: in, is the Kronecker product and ⊙ is the Hadamard product.

8. The distributed fuzzy fault detection method according to claim 7, characterized in that: The specific process of step 6 is: (1) Based on the design principle of stability and robustness to disturbances, it is necessary to ensure that the error augmentation system is asymptotically stable and has the specified H ∞ Performance γ, so that for the disturbance signal ω(t), the following equation holds: γ is a performance indicator to measure the disturbance attenuation capability, and the superscript T indicates the transition rank; (2) When H ∞ When a γ-distributed fuzzy fault detection filter with good performance exists, the matrix containing its gain information is directly obtained through the formulated solution conditions. And the correlation matrix L2, through the following operation, the gain matrix of the fault detection filter is obtained 9. The distributed fuzzy fault detection method according to claim 1, characterized in that: The residual evaluation function R in step 7 m (r m (t)) The specific form is: Among them, t is the detection time and t0 is the initial time.

10. The distributed fuzzy fault detection method according to claim 9, characterized in that: The specific process of fault detection in step 8 is as follows: Set the threshold form of the residual evaluation function: The strategy of distributed fuzzy fault detection for multi-unmanned ship systems is: if the residual evaluation function value obtained by real-time detection is greater than the threshold of the residual evaluation function, an alarm is given; otherwise, no alarm is given.

Images