A deadlock-free collision avoidance method for AUV formation based on artificial potential field method under communication delay
Through the multi-AUV formation control method based on artificial potential field method, the problem of deadlock and collision delay in communication of AUV formation is solved, and the stable convergence and collision avoidance effect of formation is achieved. Through feedback linearization and collision avoidance potential field function, the linearization and collision avoidance of the formation is achieved. The design ensures that the AUV can effectively coordinate and avoid deadlocks during communication delays.
Patent Information
- Application Number
- CN202210693357.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-17
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2042-06-17
AI Technical Summary
AUV formations are prone to deadlock and collision problems under communication delay conditions. The existing technology has failed to effectively solve these problems, especially during formation transformation. The collision avoidance algorithm may cause the AUV to fail to reach the desired position.
Using the multi-AUV formation control method based on artificial potential field method, by establishing a nonlinear model of AUV and performing feedback linearization, a formation consistency controller with time delay is designed, and a collision avoidance potential field function is introduced into the controller. and deadlock avoid components to ensure that the AUV formation converges stably under communication delay.
It effectively avoids collision and deadlock problems in AUV formations under communication delay conditions, so that the formation can converge stably in the presence of communication delay, and improves the collision avoidance effect and stability of the formation.
Smart Images

Figure CN114995469B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of ship control, and in particular relates to a deadlock-free collision avoidance method for an AUV formation based on an artificial potential field method under communication delay. Background Art
[0002] As the demand for marine missions increases, the working environment of AUVs will become more and more complex, and the difficulty of their operations will also increase. Due to the limited resources they carry and their own lack of capabilities, they are generally unable to complete these tasks independently. In recent years, research work based on multi-AUV underwater detection systems has emerged in an endless stream. The collaborative operation of AUV formations instead of single AUV operations can make up for the problem of their own lack of capabilities. The formation has higher work efficiency, especially in navigation, path planning, collaborative data collection, target search and other aspects.
[0003] The basis of AUV formation coordination control in actual operation is that members can complete information exchange, that is, they have the ability to communicate with each other. However, the transmission speed of underwater acoustic communication is slow, which limits the bandwidth of underwater acoustic communication, so compared with the radio transmission method on land, the communication delay cannot be ignored. Therefore, communication delay is an inevitable problem in AUV formation coordination operations. In the process of formation transformation, the AUV may not be able to reach the expected formation position due to the control effect of the collision avoidance algorithm. For example, the positions of AUV1 and AUV2 are their respective expected positions, the driving force and potential field force are collinear, and they cancel each other out, resulting in a deadlock phenomenon.
[0004] The existing research results on the deadlock problem were searched. The literature (Study on Multi-AUV Formation Control Based on the Hungarian Algorithm) proposed the deadlock problem but did not propose corresponding solutions to the deadlock problem. Although the probability of deadlock problem occurring in the actual environment is low, the introduction of the deadlock avoidance component can prevent the formation control from falling into deadlock in some extreme environments. At the same time, adding the deadlock avoidance component will hardly affect the effect of formation collision avoidance. Summary of the invention
[0005] The purpose of the present invention is to solve the collision and deadlock problems that may occur during the AUV formation control process, and to provide an AUV formation deadlock-free collision avoidance method based on an artificial potential field method under communication delay.
[0006] A deadlock-free collision avoidance method for AUV formation based on artificial potential field method under communication delay includes the following steps:
[0007] Step 1: For an AUV formation consisting of a leader and n followers, the AUV formation has the same time delay τ; obtain the position information, attitude information, speed information and angular velocity information of each AUV in the formation;
[0008] Step 2: Establish a nonlinear model of the AUV and perform feedback linearization on it;
[0009] Ignoring the rolling motion, a three-dimensional mathematical model of the fully driven AUV is established in the hull coordinate system;
[0010]
[0011] in, Represents the pose vector of AUV in the earth coordinate system; represents the velocity vector of the AUV in the hull coordinate system; M is the inertia matrix, including the additional mass; J(η) is the transformation matrix; C(μ) is the centripetal force and Coriolis force matrix, including the centripetal force and Coriolis force generated by the additional mass; D(μ) is the hydrodynamic drag and lift moment; g(η) is the restoring force and moment vector; is the controller input vector in the ship coordinate system; is the parameter matrix of the actuator, ignoring the effects of high-order nonlinear hydrodynamic damping terms and rolling motion on the AUV;
[0012] The three items C(μ)μ, D(μ)μ, and g(η) are combined into a column vector N(η,μ), and the AUV nonlinear model is transformed into:
[0013]
[0014] In order to facilitate the linearization of the model, the above formula is written as follows:
[0015]
[0016] The output of the nonlinear system is the pose vector, so the nonlinear model of AUV is:
[0017]
[0018] Among them, h(ξ)=(x,y,z,θ,ψ) T ; g(ξ)=M1g′;
[0019] According to Lie derivative:
[0020]
[0021] The first-order Lie derivative L can be obtained f h(ξ),L g h(ξ) has the following form:
[0022]
[0023] According to the definition of the second-order Lie derivative, we know that:
[0024]
[0025] Second-order Lie derivative L g L f h(ξ) has the following form:
[0026]
[0027] The sum of the relative orders of the model is: ρ1+ρ2+ρ3+ρ4+ρ5=10, where ρ1=ρ2=ρ3=ρ4=ρ5=2, that is, the sum of the relative orders is equal to the order of the system, 10; it can be seen that the AUV model can be accurately feedback linearized and has a solution; the coordinate changes are selected as follows:
[0028]
[0029] Substituting the Lie derivative into the equation yields:
[0030]
[0031] make Then the actual control input under the new coordinate transformation is:
[0032]
[0033] The AUV linearization mathematical model is obtained as follows:
[0034]
[0035] All AUVs in the formation have the same nonlinear model, so other AUVs in the formation can perform the same feedback linearization. The linearized mathematical model of the i-th AUV after conversion is:
[0036]
[0037] Where, 0≤i≤n, i=0 represents the leader AUV, i=1,2,...,n represents the i-th follower AUV;
[0038]
[0039] Step 3: Design a formation consistency controller for multiple AUVs under time delay;
[0040] Step 3.1: The AUV formation has the same delay τ. After considering the formation delay, the multi-AUV formation formation consistency controller is expressed as:
[0041] Among them, α represents the position gain; β is the speed gain; σ l (t) and v l(t) are the position and speed of the leader in the formation; a ij (t) represents the communication relationship between the i-th follower AUV and the j-th follower AUV; c il (t) represents the communication relationship between the i-th follower AUV and the leader AUV; N i is the neighborhood of the i-th AUV;
[0042]
[0043] Step 3.2: Solve the collision problem during the formation transformation process and design the potential field function related to collision avoidance in the controller;
[0044] Assume Δ ij represents the relative distance between the two AUVs, r represents the safe distance of the AUV, R is the radius of the repulsive area, and AUV i =(x i (t),y i (t),z i (t)) represents the coordinates of the i-th AUV, AUV j Represents the coordinate information of the jth AUV, and the specific potential energy function is:
[0045]
[0046] Among them, k r is the control gain of the potential energy function, and by adjusting the control gain k r Make the AUV control smoother, so that the collision avoidance effect is more ideal; for the potential energy function U ij The gradient function is calculated to obtain the potential field force of the i-th AUV on the j-th AUV:
[0047]
[0048] Among them, P i =[ΔxΔyΔz] T , Δx=x i (t)-x j (t), Δy=y i (t)-y j (t), Δz=z i (t)-z j (t), and Then the potential force along the x direction is: The potential field forces along the y and z directions can be obtained in the same way, so the total potential field force on the i-th AUV is expressed as follows:
[0049]
[0050] Step 3.3: Solve the problem of adding the collision avoidance component uri (t) The deadlock problem that may be caused after that;
[0051] Due to the addition of the collision avoidance component u ri (t), which may cause the control force in a certain direction of the driving formation and the collision avoidance component u ri (t) are collinear and cancel each other out, so the AUV cannot reach the desired position, resulting in a deadlock problem. To avoid the deadlock problem, a deadlock avoidance component u is added to the formation controller. rj (t) = ρu ri (t), where ρ is a small control coefficient; when the collision avoidance force u ri When (t) is collinear with a certain direction control force driving the formation, a force u perpendicular to the collinear direction will be generated for a period of time. rj (t), so that the two AUVs can bypass each other and avoid the deadlock problem. The total collision avoidance force is
[0052] u rk (t) = u ri (t)+u rj (t)
[0053] The total collision avoidance force u rk (t) is added to the controller designed in step 3.1 to obtain the final controller:
[0054]
[0055] With the help of this controller, the collision and deadlock problems that may occur during the AUV formation control process under the conditions of communication delay and parameter change disturbance can be avoided.
[0056] The beneficial effects of the present invention are:
[0057] Based on the artificial potential field method and the second-order integral model, the present invention provides an AUV formation deadlock-free collision avoidance method based on the artificial potential field method under communication delay, which solves the collision and deadlock problems that may occur in the AUV formation control process, so that the formation can converge stably under the condition of communication delay. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] Figure 1 Schematic diagram of the moving coordinate system and the fixed coordinate system.
[0059] Figure 2 This is a diagram of the formation.
[0060] Figure 3 This is the three-dimensional trajectory diagram of the AUV formation without the collision avoidance component controller.
[0061] Figure 4It is the two-dimensional trajectory diagram of the AUV formation without the collision avoidance component controller.
[0062] Figure 5 Schematic diagram of the AUV formation position information without the collision avoidance component controller.
[0063] Figure 6 Schematic diagram of the AUV formation attitude information without the collision avoidance component controller.
[0064] Figure 7 Schematic diagram of the AUV formation speed information without the collision avoidance component controller.
[0065] Figure 8 Schematic diagram of the angular velocity information of the AUV formation without the collision avoidance component controller.
[0066] Fig. 9 It is the three-dimensional trajectory diagram of the AUV formation under the action of the collision avoidance component controller.
[0067] Fig.10 It is the two-dimensional trajectory diagram of the AUV formation under the action of the collision avoidance component controller.
[0068] Fig.11 Schematic diagram of the AUV formation position information with the collision avoidance component controller.
[0069] Fig.12 Schematic diagram of the AUV formation attitude information with the collision avoidance component controller.
[0070] Fig.13 Schematic diagram of the AUV formation speed information with the collision avoidance component controller.
[0071] Fig.14 Schematic diagram of the angular velocity information of the AUV formation under the action of the collision avoidance component controller. DETAILED DESCRIPTION
[0072] The present invention is further described below in conjunction with the accompanying drawings.
[0073] The purpose of the present invention is to solve the collision and deadlock problems that may occur in the AUV formation control process, and to propose an AUV formation control method based on the combination of artificial potential field method and second-order integral model, which can make the formation converge stably under the condition of communication delay.
[0074] First, establish a leader-follower formation and initialize it;
[0075] Consider an AUV formation with a leader and n followers, where n is a positive integer, the leader's state is marked with 0, and the states of the n followers are marked from 1 to n. Initialize the position information, attitude information, velocity information, and angular velocity information of each AUV in the formation.
[0076] Secondly, a nonlinear model of the AUV is established and feedback linearized;
[0077] In order to realize the motion control of AUV, the motion coordinate system and the fixed coordinate system are first established, such as Figure 1 As shown, a three-dimensional mathematical model of the fully driven AUV is established in the hull coordinate system (ignoring the rolling motion):
[0078]
[0079] Where: Represents the pose vector of AUV in the earth coordinate system; represents the velocity vector of the AUV in the hull coordinate system; M is the inertia matrix, including the additional mass; J(η) is the transformation matrix; C(μ) is the centripetal force and Coriolis force matrix, including the centripetal force and Coriolis force generated by the additional mass; D(μ) is the hydrodynamic drag and lift moment; g(η) is the restoring force and moment vector; is the controller input vector in the ship coordinate system. is the parameter matrix of the actuator, ignoring the effects of high-order nonlinear hydrodynamic damping terms and rolling motion on the AUV.
[0080] For the research object of this paper, the three items C(μ)μ, D(μ)μ, and g(η) in the model are combined into a column vector N(η,μ), then the AUV nonlinear model can be transformed into:
[0081]
[0082] In order to facilitate the linearization of the model, the above formula (2-15) is written as follows:
[0083]
[0084] The output of the nonlinear system is the pose vector, so the nonlinear model of AUV is:
[0085]
[0086] And h(ξ)=(x,y,z,θ,ψ) T , g(ξ)=M1g′.
[0087] According to Lie derivative:
[0088]
[0089] The first-order Lie derivative L can be obtained f h(ξ),L g h(ξ) has the following form:
[0090]
[0091] Similarly, according to the definition of the second-order Lie derivative, we know that:
[0092]
[0093] And the second-order Lie derivative L g L f h(ξ) has the following form:
[0094]
[0095] The sum of the relative orders of the model is: ρ1+ρ2+ρ3+ρ4+ρ5=10, where ρ1=ρ2=ρ3=ρ4=ρ5=2, that is, the sum of the relative orders is equal to the order of the system, 10. It can be seen that the AUV model can be accurately feedback linearized and has a solution. The coordinate changes are selected as follows:
[0096]
[0097] Substituting the Lie derivative (2-19) into the equation, we get:
[0098]
[0099] make Then the actual control input under the new coordinate transformation is:
[0100]
[0101] The AUV linearization mathematical model is obtained as follows:
[0102]
[0103] All AUVs in the formation have the same nonlinear model, so other AUVs in the formation can perform the same feedback linearization. The linearized mathematical model of the i-th AUV after conversion is:
[0104]
[0105] Among them, 0≤i≤n,
[0106] Then, a formation consistency controller for multiple AUVs in the presence of time delay is designed.
[0107] Assuming that the AUV formation has the same time delay τ, after considering the formation delay, the multi-AUV formation formation consistency controller is expressed as:
[0108]
[0109] Where α is the position gain, β is the velocity gain, σ l (t) and v l (t) are the position and speed of the leader in the formation, a ij (t) represents the communication relationship between the ith follower AUV and the jth follower AUV, c il (t) represents the communication relationship between the i-th follower AUV and the leader AUV, N i is the neighborhood of the i-th AUV.
[0110] Finally, multi-AUV collision avoidance control based on the artificial potential field method is realized.
[0111] Consider the possible collision and deadlock problems that may occur during formation changes.
[0112] (1) Solve the collision problem during formation transformation: Design the potential field function related to collision avoidance in the controller. Let Δ ij represents the relative distance between the two AUVs, r represents the safe distance of the AUV, R is the radius of the repulsive area, and AUV i =(x i (t),y i (t),z i (t)) represents the coordinates of the i-th AUV, AUV j Represents the coordinate information of the jth AUV. The specific potential energy function is given below:
[0113]
[0114] Where: k r is the control gain of the potential energy function. By adjusting the gain, the AUV control can be smoother, thus achieving a better collision avoidance effect. ij By finding the gradient function, we can get the potential force of the i-th AUV on the j-th AUV:
[0115]
[0116] Among them, P i =[Δx Δy Δz] T , Δx=x i (t)-x j (t), Δy=y i (t)-y j (t), Δz=z i (t)-z j (t), and Then the potential force along the x direction is: The potential field forces along the y and z directions can be obtained in the same way, so the total potential field force on the i-th AUV is expressed as follows:
[0117]
[0118] The collision avoidance component u ri (t) is added to the controller designed in step 3 to obtain the new controller:
[0119]
[0120] (2) Solve the problem by adding the collision avoidance component u ri (t) may cause deadlock problems
[0121] Due to the addition of the collision avoidance component u ri (t), which may cause the control force in a certain direction of the driving formation and the collision avoidance component u ri (t) are collinear and cancel each other out, so the AUV cannot reach the desired position, resulting in a deadlock problem. To avoid the deadlock problem, a deadlock avoidance component u is added to the formation controller. rj (t) = ρu ri (t), where ρ is a small control coefficient. ri When (t) is collinear with a certain direction control force driving the formation, a force u perpendicular to the collinear direction will be generated for a period of time. rj (t), so that the two AUVs can bypass each other and avoid the deadlock problem. The total collision avoidance force is:
[0122] u rk (t) = u ri (t)+u rj (t) (19) The total collision avoidance force u rk (t) is added to the controller designed in step 3 to obtain the new controller:
[0123]
[0124] Under the action of the controller shown in formula (12), the collision and deadlock problems that may occur during the AUV formation control process under the conditions of communication delay and parameter change disturbance can be avoided.
[0125] The stability of the above design is proved as follows:
[0126] set up is the posture state of the AUV formation, is the speed state of the AUV formation, L1 is the Laplacian matrix of the follower AUV communication topology, and L2 represents the communication state matrix of the leader and follower. The state of the formation can be expressed as:
[0127]
[0128] in, represents the set of potential field forces of each AUV in the formation, 1 n Represents an n-dimensional column vector whose elements are all 1, i.e. [1 1...1] T ∈R n Assume the position state σ of the follower relative to the leader l (t) and velocity state v l The state error of (t) is in:
[0129]
[0130] The following formula can be derived from the Laplacian matrix and the Kronecker product:
[0131]
[0132] According to equations (20) and (22), the state equation of the original AUV formation can be converted into the error equation of the formation system, which is in the following form:
[0133]
[0134] Simplifying formula (23), we get the total system error E(x):
[0135]
[0136] in, If the LMI matrix inequality in formula (3-26) is satisfied, all follower AUVs can track the position and velocity of the leader under the action of the designed coordinated controller, thereby forming the desired formation and maintaining consistency.
[0137]
[0138] Among them, ψ 11 =2PA+2PB+Q+τA T RA-PBM -1 B T P T , M is a 2n-dimensional positive definite matrix.
[0139] Consider the following Lyapunov-Krasovskii functional:
[0140]
[0141] Where the symmetric matrix P≥0,Q≥0,R≥0, let V1(t)=E T(t)PE(t), By taking the derivative of the total energy function V(t), we can obtain:
[0142]
[0143]
[0144]
[0145] In formula (29), f(C) is the correlation function of the collision avoidance term C, which also exists in formula (27). If the collision avoidance condition is met, the potential field force eventually tends to 0. In formula (29), it can be obtained from R positive definite that From the Newton-Leibniz formula we can get:
[0146]
[0147] so:
[0148]
[0149] It can be concluded that if the AUV formation satisfies the LMI matrix inequality of formula (25), it converges. Therefore, the AUV formation based on feedback linearization can stably converge to the desired formation and maintain it under the same communication delay when satisfying the constraints such as collision avoidance and LMI.
[0150] Embodiment 1:
[0151] Step 1: Establish a leader-follower formation and initialize parameters;
[0152] Consider an AUV formation with 1 leader and 2 followers. The leader's status is marked as 0, and the status of the 2 followers are marked as 1 and 2. The formation maintained by the three AUVs is as follows: Figure 2 As shown, R is 10 meters and β is π / 4. The AUV model in the literature (Study on the coordinated control method of underwater submersible formation ocean exploration) is selected as the simulation object, and the detailed model parameters are the same.
[0153] The trajectory of the design leader is:
[0154]
[0155] Two followers are randomly placed: x i (0) and y i The initial range of (0) is [0,5]m, z i The initial range of (0) is [-3,0]m, and the initial posture θ i (0) and heading angle ψ i(0) The initial range is [0,2]rad, and the longitudinal velocity is u i (0) The initial range is [0,0.5] m / s, and other speeds are initialized to 0 m / s.
[0156] Step 2: Establish a nonlinear model of the AUV and perform feedback linearization
[0157] In order to realize the motion control of AUV, the motion coordinate system and the fixed coordinate system are first established, such as Figure 1 As shown, a three-dimensional mathematical model of the fully driven AUV is established in the hull coordinate system (ignoring the rolling motion):
[0158]
[0159] Where: Represents the pose vector of AUV in the earth coordinate system; represents the velocity vector of the AUV in the hull coordinate system; M is the inertia matrix, including the additional mass; J(η) is the transformation matrix; C(μ) is the centripetal force and Coriolis force matrix, including the centripetal force and Coriolis force generated by the additional mass; D(μ) is the hydrodynamic drag and lift moment; g(η) is the restoring force and moment vector; is the controller input vector in the ship coordinate system. is the parameter matrix of the actuator, ignoring the effects of high-order nonlinear hydrodynamic damping terms and rolling motion on the AUV.
[0160] For the research object of this paper, the three items C(μ)μ, D(μ)μ, and g(η) in the model are combined into a column vector N(η,μ), then the AUV nonlinear model can be transformed into:
[0161]
[0162] In order to facilitate the linearization of the model, the above formula (2-15) is written as follows:
[0163]
[0164] The output of the nonlinear system is the pose vector, so the nonlinear model of AUV is:
[0165]
[0166] And h(ξ)=(x,y,z,θ,ψ) T , g(ξ)=M1g′.
[0167] According to Lie derivative:
[0168]
[0169] The first-order Lie derivative L can be obtained f h(ξ),L g h(ξ) has the following form:
[0170]
[0171] Similarly, according to the definition of the second-order Lie derivative, we know that:
[0172]
[0173] And the second-order Lie derivative L g L f h(ξ) has the following form:
[0174]
[0175] The sum of the relative orders of the model is: ρ1+ρ2+ρ3+ρ4+ρ5=10, where ρ1=ρ2=ρ3=ρ4=ρ5=2, that is, the sum of the relative orders is equal to the order of the system, 10. It can be seen that the AUV model can be accurately feedback linearized and has a solution. The coordinate changes are selected as follows:
[0176]
[0177] Substituting the Lie derivative (2-19) into the equation, we get:
[0178]
[0179] make Then the actual control input under the new coordinate transformation is:
[0180]
[0181] The AUV linearization mathematical model is obtained as follows:
[0182]
[0183] All AUVs in the formation have the same nonlinear model, so other AUVs in the formation can perform the same feedback linearization. The linearized mathematical model of the i-th AUV after conversion is:
[0184]
[0185] Among them, 0≤i≤n,
[0186] Step 3: Design a formation consistency controller for multiple AUVs;
[0187] Assuming that the AUV formation has the same time delay τ, after considering the formation delay, the multi-AUV formation formation consistency controller is expressed as:
[0188]
[0189] Where α is the position gain, β is the velocity gain, σ l (t) and v l (t) are the position and speed of the leader in the formation, a ij (t) represents the communication relationship between the ith follower AUV and the jth follower AUV, c il (t) represents the communication relationship between the i-th follower AUV and the leader AUV, N i is the neighborhood of the i-th AUV.
[0190] Step 4: Multi-AUV collision avoidance control based on artificial potential field method
[0191] Consider the possible collision and deadlock problems that may occur during formation changes.
[0192] (1) Solve the collision problem during formation transformation: Design the potential field function related to collision avoidance in the controller. Let Δ ij represents the relative distance between the two AUVs, r represents the safe distance of the AUV, R is the radius of the repulsive area, and AUV i =(x i (t),y i (t),z i (t)) represents the coordinates of the i-th AUV, AUV j Represents the coordinate information of the jth AUV. The specific potential energy function is given below:
[0193]
[0194] Where: k r is the control gain of the potential energy function. By adjusting the gain, the AUV control can be smoother, thus achieving a better collision avoidance effect. ij By finding the gradient function, we can get the potential force of the i-th AUV on the j-th AUV:
[0195]
[0196] Among them, P i =[Δx Δy Δz] T , Δx=x i (t)-x j (t), Δy=y i (t)-y j (t), Δz=z i (t)-zj (t), and Then the potential force along the x direction is: The potential field forces along the y and z directions can be obtained in the same way, so the total potential field force on the i-th AUV is expressed as follows:
[0197]
[0198] The collision avoidance component u ri (t) is added to the controller designed in step 3 to obtain the new controller:
[0199]
[0200] (2) Solve the problem by adding the collision avoidance component u ri (t) may cause deadlock problems
[0201] Due to the addition of the collision avoidance component u ri (t), which may cause the control force in a certain direction of the driving formation and the collision avoidance component u ri (t) are collinear and cancel each other out, so the AUV cannot reach the desired position, resulting in a deadlock problem. To avoid the deadlock problem, a deadlock avoidance component u is added to the formation controller. rj (t) = ρu ri (t), where ρ is a small control coefficient. ri When (t) is collinear with a certain direction control force driving the formation, a force u perpendicular to the collinear direction will be generated for a period of time. rj (t), so that the two AUVs can bypass each other and avoid the deadlock problem. The total collision avoidance force is:
[0202] u rk (t) = u ri (t)+u rj (t)(52) The total collision avoidance force u rk (t) is added to the controller designed in step 3 to obtain the new controller:
[0203]
[0204] Under the action of the controller shown in formula (53), the collision and deadlock problems that may occur during the AUV formation control process under the conditions of communication delay and parameter change disturbance can be avoided.
[0205] The design control parameters are α = 0.06, β = 0.7, and the delay time τ = 1s. Since the two follower AUVs receive the position and posture information of the leader respectively, there is no communication between the two follower AUVs. The coefficient k in the collision avoidance potential energy function rSet to 0.5, the collision avoidance radius R of each AUV in the formation is 8.5 meters, and the safety distance r is 5 meters.
[0206] According to formula (54) and formula (55), the collision avoidance component u is calculated respectively: rj (t) and the controller with collision avoidance component u rj Controller of (t):
[0207]
[0208]
[0209] Then, the controller is transformed into the motion coordinate system according to formula (56):
[0210]
[0211] Apply τ to the AUV nonlinear model. Assume that at 1600 seconds, a change in formation occurs. The expected paths of AUV1 and AUV2 are swapped. rj (t) controller, at this time what happens Figure 3 and Figure 4 The deadlock phenomenon in the formation seriously affects the stability of formation control. Figure 5 and Figure 6 is the position information of the AUV formation, Figure 7 and Figure 8 The speed and angular velocity information of the AUV formation; rj (t) controller, such as Fig. 9 and Fig.10 As shown in the figure, the two AUVs staggered in a bypassing manner instead of being in a deadlock state for a long time, and the formation converged stably. Fig.11 and Fig.12 is the position information of the AUV formation, Fig.13 and Fig.14 It is the speed and angular velocity information of the AUV formation.
[0212] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A deadlock-free collision avoidance method for AUV formation based on artificial potential field method under communication delay, characterized in that: The following steps are involved: Step 1: For an AUV formation consisting of a leader and n followers, the AUV formation has the same time delay τ; obtain the position information, attitude information, speed information and angular velocity information of each AUV in the formation; Step 2: Establish a nonlinear model of the AUV and perform feedback linearization on it; Ignoring the rolling motion, a three-dimensional mathematical model of the fully driven AUV is established in the hull coordinate system; in, Represents the pose vector of AUV in the earth coordinate system; represents the velocity vector of the AUV in the hull coordinate system; M is the inertia matrix, including the additional mass; J(η) is the transformation matrix; C(μ) is the centripetal force and Coriolis force matrix, including the centripetal force and Coriolis force generated by the additional mass; D(μ) is the hydrodynamic drag and lift moment; g(η) is the restoring force and moment vector; is the controller input vector in the ship coordinate system; is the parameter matrix of the actuator, ignoring the effects of high-order nonlinear hydrodynamic damping terms and rolling motion on the AUV; The three items C(μ)μ, D(μ)μ, and g(η) are combined into a column vector N(η,μ), and the AUV nonlinear model is transformed into: In order to facilitate the linearization of the model, the above formula is written as follows: The output of the nonlinear system is the pose vector, so the nonlinear model of AUV is: Where, h(ξ)=(x,y,z,θ,ψ) T ;g(ξ)=M1g′; According to Lie derivative: The first-order Lie derivative L can be obtained f h(ξ),L g h(ξ) has the following form: According to the definition of the second-order Lie derivative, we know that: Second-order Lie derivative L g L f h(ξ) has the following form: The sum of the relative orders of the model is: ρ1+ρ2+ρ3+ρ4+ρ5=10, where ρ1=ρ2=ρ3=ρ4=ρ5=2, that is, the sum of the relative orders is equal to the order of the system, 10; it can be seen that the AUV model can be accurately feedback linearized and has a solution; the coordinate changes are selected as follows: Substituting the Lie derivative into the equation yields: make Then the actual control input under the new coordinate transformation is: The AUV linearization mathematical model is obtained as follows: All AUVs in the formation have the same nonlinear model, so other AUVs in the formation can perform the same feedback linearization. The linearized mathematical model of the i-th AUV after conversion is: Where, 0≤i≤n, i=0 represents the leader AUV, i=1,2,...,n represents the i-th follower AUV; Step 3: Design a formation consistency controller for multiple AUVs under time delay; Step 3.1: The AUV formation has the same delay τ. After considering the formation delay, the multi-AUV formation formation consistency controller is expressed as: Among them, α represents the position gain; β is the speed gain; σ l (t) and v l (t) are the position and speed of the leader in the formation; a ij (t) represents the communication relationship between the i-th follower AUV and the j-th follower AUV; c il (t) represents the communication relationship between the i-th follower AUV and the leader AUV; N i is the neighborhood of the i-th AUV; Step 3.2: Solve the collision problem during the formation transformation process and design the potential field function related to collision avoidance in the controller; Assume Δ ij represents the relative distance between the two AUVs, r represents the safe distance of the AUV, R is the radius of the repulsive area, and AUV i =(x i (t),y i (t),z i (t)) represents the coordinates of the i-th AUV, AUV j Represents the coordinate information of the jth AUV, and the specific potential energy function is: Among them, k r is the control gain of the potential energy function, and by adjusting the control gain k r Make the AUV control smoother, so that the collision avoidance effect is more ideal; for the potential energy function U ij The gradient function is calculated to obtain the potential field force of the i-th AUV on the j-th AUV: Among them, P i =[Δx Δy Δz] T , Δx=x i (t)-x j (t), Δy=y i (t)-y j (t), Δz=z i (t)-z j (t), and Then the potential force along the x direction is: The potential field forces along the y and z directions can be obtained in the same way, so the total potential field force on the i-th AUV is expressed as follows: Step 3.3: Solve the problem of adding the collision avoidance component u ri (t) The deadlock problem that may be caused after that; Due to the addition of the collision avoidance component u ri (t), which may cause the control force in a certain direction of the driving formation and the collision avoidance component u ri (t) are collinear and cancel each other out, so the AUV cannot reach the desired position, resulting in a deadlock problem. To avoid the deadlock problem, a deadlock avoidance component u is added to the formation controller. rj (t)=ρu ri (t), where ρ is a small control coefficient; when the collision avoidance force u ri When (t) is collinear with a certain direction control force driving the formation, a force u perpendicular to the collinear direction will be generated for a period of time. rj (t), so that the two AUVs can bypass each other and avoid the deadlock problem. The total collision avoidance force is u rk (t)=u ri (t)+u rj (t) The total collision avoidance force u rk (t) is added to the controller designed in step 3.1 to obtain the final controller: With the help of this controller, the collision and deadlock problems that may occur during the AUV formation control process under the conditions of communication delay and parameter change disturbance can be avoided.
