UUV (Unmanned Underwater Vehicle) hunting method based on distributed model predictive control

Abstract

The invention relates to the technical field of automatic control, in particular to a UUV (Unmanned Underwater Vehicle) hunting method based on distributed model predictive control, which comprises the following steps: establishing a distributed control system structure, acquiring a predictive state sequence of a current UUV and an optimal predictive state sequence of a neighborhood UUV through an outer-layer controller, and acquiring the optimal predictive state sequence of the current UUV and the optimal predictive state sequence of the neighborhood UUV; and carrying out optimization solution to respectively obtain an optimal prediction state sequence and an optimal control sequence of the current UUV, and adding the current UUV control compensation amount and the optimal prediction state sequence and the optimal control sequence of the current UUV to obtain a final control amount of the current UUV. According to the invention, the outer-layer controller is responsible for communication interaction and calculation of a detection target and an optimal control sequence, and the inner-layer controller is responsible for compensating an uncertain part of a system model and resisting external interference.

Description

Technical Field

[0001] The present invention relates to the field of automatic control technology, and in particular to a UUV encirclement method and system based on distributed model predictive control. Background Art

[0002] An unmanned underwater vehicle (UUV) is an underwater vehicle that uses various sensors and other mission modules to perform tasks such as minesweeping, reconnaissance, surveying and intelligence gathering, and uses its own energy equipment, navigation equipment and actuators to perform autonomous navigation, autonomous decision-making and autonomous control.

[0003] With the increasing demand for ocean missions, single vehicles face limitations such as poor redundancy, small mission scope and low work efficiency. Using multiple UUV systems to perform common tasks can overcome the above problems. Multi-UUV collaboration can increase the redundancy and robustness of the system, improve mission execution efficiency and reduce costs. Therefore, the multi-UUV collaboration system has received more and more attention from scientific research institutions and scholars. The multi-UUV system has the characteristics of spatial distribution, time distribution, information distribution, resource distribution and function distribution.

[0004] As a new type of collaborative platform, the multi-UUV collaborative formation system not only enriches and expands the maritime unmanned intelligent technology system, but is also an important member of the "sea, land and air" multi-body collaborative operation system. The multi-UUV collaborative formation system breaks through the limitations of a single UUV in performing operational tasks, providing more possibilities for promoting marine development and exploration. At the same time, the development and mass equipment of the marine multi-UUV system are of great significance for conducting far-reaching marine scientific research, promoting the construction of the "transparent ocean" project, and maintaining marine safety.

[0005] Traditional multi-agent control methods mostly use centralized model control algorithms, whose controller design is relatively convenient and the internal communication scheme of the system is relatively simple. However, the application of centralized MPC controllers in large-scale intelligent agent groups is prone to problems such as large computational complexity, large communication load, low fault tolerance and reliability. In view of the complexity and difficulty of non-global communication in multi-UUV systems, improvements have been made to local communication methods. Distributed model prediction can solve the above problems well and has better processing effects on multi-input and multi-output systems. It improves the robustness of the system and significantly reduces the computational complexity and communication volume. Summary of the invention

[0006] In order to solve the above technical problems, the present invention mainly aims at the complexity and difficulty of non-global communication of multi-UUV system, improves the local communication mode, and proposes a UUV roundup method based on distributed model predictive control. This method designs a two-layer MPC control framework for multi-UUV roundup control tasks, in which the outer controller is responsible for communication interaction, detection target and calculation of optimal control sequence, and the inner controller is responsible for compensating the uncertain part of the system model and resisting external interference. The communication mode in the group adopts the local communication mode, and the two-dimensional grid coding method is used to restrict the individual neighborhood, simplifying the communication neighbors of each UUV to two, which is conducive to reducing the communication pressure and computational complexity in large multi-UUV systems.

[0007] The first object of the present invention is to provide a UUV capture method based on distributed model predictive control, which is used to control multiple UUVs to capture a moving target, comprising: Establish a distributed control system structure, including inner and outer controllers corresponding to each UUV; The predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV are obtained through the outer controller, and the optimal predicted state sequence and the optimal control sequence of the current UUV are obtained through optimization. The predicted state sequence of the current UUV is subtracted from the state quantity of the actual measurement system through the inner controller to obtain the error, and then the error model is defined by the error, and the error model is solved to obtain the control compensation of the current UUV; The current UUV control compensation amount is added to the current UUV optimal prediction state sequence and optimal control sequence to obtain the current UUV final control amount; After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again to carry out the encirclement control of the current UUV.

[0008] Preferably, the outer controller corresponding to each UUV is responsible for target detection, information interaction between neighbors, and solving the optimal prediction state and the optimal control sequence; the inner controller corresponding to each UUV uses the optimal prediction state output by the outer controller as a reference state to perform error compensation on the current UUV to obtain the control compensation amount.

[0009] Preferably, it also includes: Establish a group local communication network for local information exchange between UUVs; The optimal predicted state sequence and optimal control sequence of the current UUV are obtained through the outer controller and uploaded to the local communication network for the outer controller of the neighboring UUV to solve.

[0010] Preferably, the neighborhood UUV set corresponding to each UUV is obtained by applying a two-dimensional grid coding process in a local communication network to obtain the neighborhood UUV set of the current UUV.

[0011] Preferably, the predicted state sequence of the current UUV is obtained according to the following steps: Construct nonlinear UUV dynamic model equations; The current state of the UUV is predicted according to the nonlinear UUV dynamics model equation to obtain the predicted state sequence of the current UUV.

[0012] Preferably, during the optimization process, each UUV maintains the following three control objectives: (1) With the moving target as the center, the capture radius converges to a fixed expected value; (2) The UUV forms a surround around the target at the desired angular velocity; (3) The angular spacing between two UUVs distributed around the moving target is the same; Among them, moving targets are encirclement targets.

[0013] Preferably, after a control cycle, the current UUV obtains the final control amount of the next cycle through the outer controller and the inner controller again, including: After a control cycle, the current UUV contains control vectors with N steps. In each control cycle, only the optimal control vector of the first step is selected. After the controller outputs to the actuator, after a control cycle, the final control quantity of the next cycle is obtained again through the outer controller and the inner controller.

[0014] Preferably, solving the error model to obtain the current UUV control compensation includes: The feedback gain K of the UUV system in a stable state is selected as the solution of the linear quadratic regulator of the error model and as the control compensation of the current UUV.

[0015] Preferably, the outer controller is a DMPC-based trapping controller; the inner controller is a Tube-based MPC controller.

[0016] The second object of the present invention is to provide a UUV capture system based on distributed model predictive control, comprising: The control system building module is used to establish a distributed control system structure, including an inner controller and an outer controller corresponding to each UUV; The control sequence calculation module is used to obtain the predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV through the outer controller, and perform optimization to obtain the optimal predicted state sequence and optimal control sequence of the current UUV respectively; The control compensation module is used to obtain the error by subtracting the predicted state sequence of the current UUV from the state quantity of the actual measurement system through the inner controller, and then define the error model through the error, and solve the error model to obtain the control compensation of the current UUV; The control module is used to add the current UUV control compensation amount to the optimal predicted state sequence and the optimal control sequence of the current UUV to obtain the final control amount of the current UUV; after a control cycle, the current UUV obtains the final control amount of the next cycle through the outer controller and the inner controller again to perform encirclement control on the current UUV.

[0017] The present invention has at least the following beneficial effects: The present invention provides a UUV capture method based on distributed model predictive control. Aiming at the multi-UUV capture control task, the present invention designs a two-layer MPC control framework, in which the outer controller is responsible for communication interaction, detection target and calculation of optimal control sequence, and the inner controller is responsible for compensating the uncertain part of the system model and resisting external interference. The communication mode in the group adopts the local communication mode, and the two-dimensional grid coding method is used to restrict the individual neighborhood, simplifying the communication neighbors of each UUV to two, which is conducive to reducing the communication pressure and calculation complexity in large multi-UUV systems.

[0018] The UUV cluster capture control method based on distributed MPC of the present invention performs information interaction through local communication, thereby improving the robustness and stability of the system.

[0019] Aiming at the problem of uncertain communication volume in a multi-UUV system and the problem of high communication network throughput easily caused by a large group, the present invention designs a two-dimensional grid coding communication method, which reduces the node calculation complexity and communication pressure.

[0020] In the method provided by the present invention, local information exchange can be performed between UUVs, and the state sequence and predicted control sequence of neighboring UUVs can be obtained through local communication. The local optimization problem of each MPC controller is only related to the local UUV state sequence and control input sequence, which greatly reduces the scale of solving the optimization problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a multi-UUV capture control system; Figure 2 It is a schematic diagram of distributed control structure; Figure 3 It is a schematic diagram of two-dimensional grid mapping in the group; Figure 4 Encircle dynamic target trajectories for multiple UUVs; Figure 5 is the distance error during the round-up process; Figure 6 It is the separation angle of each UUV during the roundup process; Figure 7 To capture the angular velocity change. DETAILED DESCRIPTION

[0022] In order to illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description in conjunction with embodiments.

[0023] The control system designed by the present invention is divided into two parts: Figure 1 As shown in the figure, the outer controller is responsible for target detection, information exchange between neighbors, and solving the optimal prediction state and optimal control sequence. The inner controller uses the optimal prediction state output by the outer controller as the reference state to compensate the error of the current under-actuated UUV. Both controllers are designed based on the model predictive control algorithm.

[0024] See also Figure 1 As shown, the present invention is based on a multi-UUV capture control system. Each UUV applies two-dimensional grid coding processing in a local communication topology to obtain a neighborhood set of the i-th unmanned boat, thereby exchanging target information and state information; in a distributed MPC controller, the capture control rule and the formulated control target are used for optimization calculation to obtain an optimal control sequence and a predicted state sequence; the inner controller subtracts the predicted state sequence from the state quantity of the actual measurement system to obtain an error, and then defines an error model through the error, designs a control law for the error model, and solves the control compensation amount to add to the optimal control sequence obtained by the outer DMPC controller to output the final control amount.

[0025] Distributed controllers are often more convenient for solving control problems of multi-agent systems. Therefore, in order to improve the reliability of the system, distributed model prediction is used for algorithm design.

[0026] In the following embodiment, taking five UUVs and a dynamic target as an example, the controller described above is used to simulate and verify the multi-UUV capture task and display the results, and then the change curves of each state quantity designed by the algorithm during the capture process are displayed and analyzed.

[0027] In the following embodiments, the outer controller is a DMPC-based trap controller; the inner controller is a Tube-based MPC controller.

[0028] The present invention provides a UUV capture method based on distributed model predictive control, which is used to control multiple UUVs to capture a moving target, including: Step 1: Establish a distributed control structure.

[0029] Establish a distributed control system structure, including inner and outer controllers corresponding to each UUV; The outer controller corresponding to each UUV is responsible for target detection, information exchange between neighbors, and solving the optimal prediction state and optimal control sequence; the inner controller corresponding to each UUV uses the optimal prediction state output by the outer controller as the reference state to perform error compensation on the current UUV to obtain the control compensation amount.

[0030] The nonlinear state equation of the entire system can be rewritten as follows: f(x(k),u(k))=[f1f2…f N ] f i (x i (k),u i (k)), (i=1,2,…,N) Among them, x(k) is the state of UUV at time k, and u(k) represents the control input at time k.

[0031] From this, the cost function of the system can be written as follows

[0032] The above formula satisfies the following constraints X i (k)={x i (k | k),x i (k+1 | k),…,x i (k+N-1 | k)} U i (k)={u i (k | k),u i (k+1 | k),…,u i (k+N-1 | k)} {X j≠i (k)}={X j (k) | j≠i} {U j≠i (k)}={U j (k) | j≠i} Among them, X i (k) and U i (k) respectively represent the N-step predicted state and N-step predicted control input sequence of the i-th UUV; {X j≠i (k)} and {U j≠i (k)} respectively represent the N-step predicted state set of other UUVs and the N-step predicted control input set of other UUVs; is the weight constant; x i(k+N-1 | k) and u i (k+N-1 | k) represents the N-step predicted state and the N-step predicted control input sequence of the i-th UUV starting from the current k-th moment.

[0033] The overall optimization problem of the system can be decomposed into N small-scale local rolling time domain optimization problems. Each optimization problem corresponds to a UUV. For the i-th UUV, its local rolling optimization control model is:

[0034] The following constraints apply to the above formula: X i (k+q+1|k)=f i (x i (k+q|k),u i (k+q|k)), q=0,1,…,N-1;i=1,2,…,N x i (k|k)=x i (k) ≤0 Among them, G()≤0 is the constraint condition of the system, It can be seen from the above analysis that since UUVs can exchange local information, the state sequence and predicted control sequence of neighboring UUVs can be obtained through local communication.

[0035] Step 2: Establish group local communication mode.

[0036] Establish a group local communication network for local information exchange between UUVs; The optimal predicted state sequence and optimal control sequence of the current UUV are obtained through the outer controller and uploaded to the local communication network for the outer controller of the neighboring UUV to solve.

[0037] It should be noted that the multi-UUV cluster capture control system proposed in the present invention is composed of multiple independent UUVs, and each capture UUV is controlled relatively independently, and only interacts with each other through common capture control tasks and upper-level capture path constraints.

[0038] Each UUV exchanges local information and obtains the state sequence and prediction control sequence of the neighboring UUV through local communication. Apply two-dimensional grid coding processing in the local communication topology to obtain the neighborhood set N of the i-th UUV i , thereby exchanging target information and status information.

[0039] Each UUV predicts its own state at time k, obtains the predicted state sequence, and applies two-dimensional grid coding processing in the local communication topology to obtain the neighborhood set N of the i-th UUV. i , thereby exchanging target information and status information.

[0040] The state feature encoding method is introduced to deal with the problem of uncertainty in the number of neighborhood nodes. Since each node in the UUV group knows its own global coordinates and heading angle and other information, each UUV node is mapped to a two-dimensional grid based on the spatial relationship within the communication range. The mapping principle is as follows: Figure 3 shown.

[0041] Figure 3 A communication model based on two-dimensional grid mapping is presented. Figure 3 (a) shows the intuitive mapping of neighborhood relationships in a local group with UUV as the center and communication distance as the radius. In the figure, a certain angle interval is shaded in yellow to correspond to Figure 3 (b) and Figure 3 The mapping relationship in (c) is Figure 3 (b) is the deconstruction description of the mapping relationship. Figure 3 (c) is to map the radial communication area in a two-dimensional grid, divide the communication area according to the distance and orientation of equal intervals, and further information interaction is performed by judging whether there is a neighboring UUV in the interval. The shaded area in the figure can be the coding area designed by the UUV for different tasks. The present invention uses the heading angle concerned by the roundup task as the interval division principle, and determines the angle interval according to the number of individuals in the group, thereby ensuring that the UUV only communicates and interacts with at most two individuals in the neighborhood.

[0042] Step 3: The optimal predicted state sequence and optimal control sequence of the current UUV.

[0043] The predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV are obtained through the outer controller, and the optimal predicted state sequence and the optimal control sequence of the current UUV are obtained through optimization. The initial optimal predicted state sequence of the neighboring UUV is obtained based on the initial state of the neighboring UUV. The neighboring UUV set corresponding to each UUV is obtained by applying two-dimensional grid coding processing in the local communication network to obtain the neighboring UUV set of the current UUV.

[0044] In this example, the outer DMPC-based capture controller obtains the state vector x of the current UUV prediction of the system state i (t) and the optimal predicted state sequence of the neighboring UUV , , and optimize to obtain the optimal prediction state sequence and the optimal control sequence .

[0045] During the optimization process, the roundup control needs to keep the following three control objectives for each UUV: (1) With the moving target as the center, the capture radius converges to a fixed expected value; (2) The UUV forms a surround around the target at the desired angular velocity; (3) The angular spacing between two UUVs distributed around the moving target is the same; Among them, moving targets are encirclement targets.

[0046] The predicted state sequence of the current UUV is obtained according to the following steps: Construct nonlinear UUV dynamic model equations; The current state of the UUV is predicted according to the nonlinear UUV dynamics model equation to obtain the predicted state sequence of the current UUV.

[0047] At each time of loop solving, each UUV will use the future prediction control input sequence to predict the state of the next M steps. According to the system dynamics model:

[0048] The prediction of the system state can be expressed as

[0049] in, represents the predicted state of the system at the nth sampling moment in the future, represents the predicted control vector at the future nth sampling moment, F is the above nonlinear UUV dynamics model equation, is the sampling time.

[0050] The present invention designs a distributed MPC controller specifically for the capture task. The three control objectives of the capture task are all related to the position relationship and angle relationship between the capture UUV and the moving target. To capture the distance between the UUV and the moving target, the specific calculation is as follows:

[0051] The first-order derivative of the envelope angle during the capture process and the velocity vector of the captured UUV are introduced as the control targets in the capture mission. The first-order derivative of the envelope angle is calculated as follows:

[0052] Define a global cost function that describes the interaction between all UUVs in the group and the capture strategy between them and the moving target. The angular velocity of the ith UUV orbiting around the moving target is denoted as (t), and the angle difference between the i-th and j-th adjacent UUVs relative to the target is recorded as 0,2π], according to the description of the local communication structure in step 2, the UUV that is ahead of the i-th UUV in the capture process is labeled k, and the UUV that is behind the i-th UUV is labeled j. A positive constant γ∈[0,1] is introduced as a gain, which can adjust the angle difference between the captured UUVs. The convergence speed of the i-th UUV around the moving target before forming an encirclement formation is related to the angle difference between the UUVs in front and behind it, and is calculated as follows:

[0053] After being constrained by the above expressions, the angular velocity of each UUV surrounding the moving target will eventually converge to the desired angular velocity For the multi-UUV cluster capture control task, the cost function in the distributed MPC algorithm is defined as follows:

[0054] When the interval between two adjacent UUVs is larger than the expected interval, the speed needs to be controlled to achieve a uniform encirclement state. This situation may be contrary to the control instructions given by the second principle of encirclement control. Therefore, the angle error of the formation during the encirclement process is introduced for adaptive optimization:

[0055] Among them, the expected value of the angle difference between adjacent individuals is ∆Φ D =2π / N, N is the number of captured UUVs; ∆Φ ik (t) is the angle difference between the i-th UUV and the leading UUV (labeled as k) relative to the capture target, ∆Φ N1 (t) is the angle difference between the Nth UUV and the leading UUV (the first one) relative to the encirclement target.

[0056] The alternative Lyapunov function is defined as follows:

[0057] make:

[0058] When the parameter γ is positive, it can be proved that the above Lyapunov function is positive definite.

[0059] It can be concluded that the expected capture angular velocity of the i-th UUV is

[0060] After introducing this adaptive expected angular velocity, the expected capture angular velocity of the i-th UUV is and the final desired angular velocity The relationship is shown in the above formula. The cost function in the model control system eventually becomes

[0061] Among them, σ represents the sampling time in the model prediction process, η i,t (t+mσ) represents the predicted distance between the current UUV and the target at the next m steps, then represents the predicted angular velocity of the current UUV rotating around the target at the next m-th step. sd and sp are constants, which are used to de-dimensionalize the errors in the cost function, and sd = χ is the expected capture distance and sp = is the expected capture angular velocity; α, β are both positive constants, representing the weights of various control objectives in the cost function. Set N i represents the set of UUVs in front and behind the current i-th UUV neighborhood, subscript j represents the UUV behind the current one, and subscript k represents the UUV ahead of the current one. The constant α can be regarded as the weight coefficient for controlling the capture distance, and the constant β can be regarded as the weight coefficient for the convergence speed of the separation angle during the capture process.

[0062] Define the prediction step size of the controller in each subsystem as M, then the local MPC control strategy can be expressed by the following formula.

[0063]

[0064]

[0065] Where M represents the prediction step of the controller, σ represents the sampling time in the model prediction process, and the prediction control vector of M steps in an optimization process is recorded as , (iσ;t) represents the predicted system state at the i-th sampling moment in an optimization step, (iσ;t) represents the system input at the i-th sampling time in an optimization step. By optimizing and solving within M prediction steps, the optimal cost function can be obtained. , so that the optimal prediction state sequence can be obtained respectively and the optimal predictive control sequence .

[0066] Step 4: Final control quantity of current UUV.

[0067] The predicted state sequence of the current UUV is subtracted from the state quantity of the actual measurement system through the inner controller to obtain the error, and then the error model is defined by the error, and the error model is solved to obtain the control compensation of the current UUV; The current UUV control compensation amount is added to the current UUV optimal prediction state sequence and optimal control sequence to obtain the current UUV final control amount; In this embodiment, the current UUV will solve to obtain the optimal predicted state sequence Uploaded to the local communication network for use by the DMPC-based controller of the neighboring UUV. and the optimal control sequence The error model is passed to the inner Tube-based MPC controller. The inner controller takes the difference between the predicted state sequence and the actual state of the measured system as the system error and inputs it into the error model. The control law design of the error model can be summed with the control quantity obtained by the distributed controller to obtain the final control output, which is then added to the optimal control sequence obtained by the outer DMPC controller to output the final control quantity.

[0068] Get the current UUV control compensation, including: A new state vector is defined to describe the motion state of UUV, and the model is linearized and approximated to obtain a new motion model of UUV; the error is obtained by subtracting the predicted state quantity of the nominal system from the state quantity of the actual measurement system, and the error is used as the input of this model and defined as the error model.

[0069] The feedback gain K of the UUV system in a stable state is selected as the solution of the linear quadratic regulator of the error model. , as the control compensation of the current UUV.

[0070] In order to compensate for the inaccuracy of the system and resist the interference of wind and waves on the water surface, the present invention optimizes the design of the control system, introduces the Tube concept, regards the uncertainty part of the system and the surface wind and wave part as the system error, and in the control process, takes the difference between the predicted state calculated by the distributed controller and the actual measured state as the system error.

[0071] First, the feedforward control input in the roundup process is calculated for the system. The present invention calls the system with feedforward control input and zero external disturbance a "nominal system"; then, the present invention designs a state feedback controller to suppress the error between the actual state quantity of the system and the predicted state quantity generated by the nominal system. In addition, the above two state quantities are combined to obtain the final control quantity of the inner control system.

[0072] Define a new state vector ξ = ∈ The motion state of the UUV is described, where x and y of the ground coordinate system point to the east and north respectively, Ψ represents the heading angle of the UUV, u is the longitudinal velocity, v is the lateral velocity, and r represents the linear velocity of the UUV rotating around the particle. The nonlinear model is linearized and approximated, and the motion model of a single UUV is expressed as follows:

[0073] Among them, the system matrix, nonlinear Lipschitz function and control matrix are as follows

[0074]

[0075] Among them, the mass of UUV is recorded as m, , , All are additional mass, is the moment of inertia in the vertical direction, , , , , , are the first-order fluid coefficient and the second-order fluid coefficient, respectively. , , , , , .

[0076] For the reformulated system, the control quantity of the present invention can be divided into two parts. First, the feedforward control input in the roundup process is calculated for the system, and the system with feedforward control input and zero external disturbance is called the "nominal system"; then a state feedback controller is designed to suppress the error between the actual state quantity of the system and the predicted state quantity generated by the nominal system, and the above two state quantities are combined to obtain the final control quantity of the inner control system.

[0077] The N-step control sequence and N-step disturbance sequence of the system are represented as τ={ , ,…, } and ω={ , ,…, }, the discrete-time linearized nominal model of the system can be expressed as follows:

[0078] The predicted state quantity of the nominal system is subtracted from the state quantity of the actual measured system and expressed as , and divide the control input into the following two parts according to the previous description ,in, : R n →R m , the error model is defined as follows:

[0079] The feedback gain K of the UUV system in a stable state is selected as the solution of the linear quadratic regulator of the above error system. , the controller K is used to calculate the robust invariant set, and the robust invariant set is calculated offline. Therefore, for a single UUV, the overall control output becomes:

[0080] Step 5: Update of optimal control.

[0081] After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again to carry out the encirclement control of the current UUV.

[0082] After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again, including: After a control cycle, the current UUV contains control vectors with N steps. In each control cycle, only the optimal control vector of the first step is selected. After the controller outputs to the actuator, after a control cycle, the final control quantity of the next cycle is obtained again through the outer controller and the inner controller.

[0083] In this example, after the actuator control obtains the optimal control sequence in the previous calculation, which contains a control vector of N steps, the first item in the sequence is applied to the system as the solution, and at the next sampling time, it goes to step 3 and uses the updated state quantity as the initial condition and performs another optimization solution.

[0084] It should be noted that the model predictive control adopts a rolling finite-time domain optimization strategy. At each sampling moment, the UUV solves the optimal control rate of the finite time domain starting from that moment based on the optimization performance index at that moment. After obtaining the optimal control input sequence, the first item in the sequence is used as the solution to act on the system. At the next sampling moment, the updated state quantity is used as the initial condition and another optimization solution is performed. In this way, the actual optimal control input sequence can be obtained by continuously performing online loop solutions.

[0085] In order to further illustrate the UUV encirclement method based on distributed model predictive control provided by the present invention, simulation verification shows that the proposed two-layer control framework can better reduce the control error, improve the robustness of the system, and the designed distributed encirclement control algorithm can effectively and stably form an encirclement situation for irregularly moving targets.

[0086] In the simulation verification, the capture control algorithm was verified by the irregular motion of the moving target. First, it was assumed that the moving target was at the initial position =[-39,42], the initial position of the UUV group is also randomly generated, the control period σ=0.1s, the controller prediction step M=50, the simulation results are as follows Figure 4 , 5 , 6, and 7.

[0087] Depend on Figure 4 , Figure 5 It can be seen that when the moving target makes irregular speed-changing motion, the distributed MPC capture control algorithm proposed in the present invention can also achieve uniform encirclement of the moving target when the target position is known. Figure 5 It can be seen that the multi-UUV group also converges the capture radius error within ±0.2m within 40s. The uniformly distributed formation formed by the collaborative capture is maintained until the end of the mission time, and the distance error changes periodically.

[0088] Figure 6 and Figure 7 The two control targets in the roundup mission were demonstrated separately. Figure 6 The angular interval between each UUV and its subsequent neighbors during the capture process is shown. After a uniformly distributed configuration is formed on the capture circle, the angular interval between adjacent UUVs is expressed as =360 / N=72°, and the angle intervals converge to within 72°±2°. Figure 7 It shows the angular velocity change of each UUV in the enveloping motion clockwise around the moving target in the capture mission. The simulation sets the expected angular velocity to 0.09 rad / s. Figure 7 It can be seen that the angular velocity error changes periodically within ±0.03rad / s.

[0089] The present invention provides a UUV capture system based on distributed model predictive control, comprising: The control system building module is used to establish a distributed control system structure, including an inner controller and an outer controller corresponding to each UUV; The control sequence calculation module is used to obtain the predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV through the outer controller, and perform optimization to obtain the optimal predicted state sequence and optimal control sequence of the current UUV respectively; The control compensation module is used to obtain the error by subtracting the predicted state sequence of the current UUV from the state quantity of the actual measurement system through the inner controller, and then define the error model through the error, and solve the error model to obtain the control compensation of the current UUV; The control module is used to add the current UUV control compensation amount to the optimal predicted state sequence and the optimal control sequence of the current UUV to obtain the final control amount of the current UUV; after a control cycle, the current UUV obtains the final control amount of the next cycle through the outer controller and the inner controller again to perform encirclement control on the current UUV.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A UUV capture method based on distributed model predictive control, characterized in that: Used to control multiple UUVs to capture a moving target, including: Establish a distributed control system structure, including inner and outer controllers corresponding to each UUV; The predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV are obtained through the outer controller, and the optimal predicted state sequence and the optimal control sequence of the current UUV are obtained through optimization. The predicted state sequence of the current UUV is subtracted from the state quantity of the actual measurement system through the inner controller to obtain the error, and then the error model is defined by the error, and the error model is solved to obtain the control compensation of the current UUV; The current UUV control compensation amount is added to the current UUV optimal prediction state sequence and optimal control sequence to obtain the current UUV final control amount; After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again to carry out the encirclement control of the current UUV.

2. The UUV capture method based on distributed model predictive control according to claim 1 is characterized in that: The outer controller corresponding to each UUV is responsible for target detection, information exchange between neighbors, and solving the optimal prediction state and optimal control sequence; The inner controller corresponding to each UUV uses the optimal predicted state output by the outer controller as the reference state to perform error compensation on the current UUV to obtain the control compensation amount.

3. The UUV capture method based on distributed model predictive control according to claim 1 is characterized in that: Also includes: Establish a group local communication network for local information exchange between UUVs; The optimal predicted state sequence and optimal control sequence of the current UUV are obtained through the outer controller and uploaded to the local communication network for the outer controller of the neighboring UUV to solve.

4. The UUV capture method based on distributed model predictive control according to claim 3 is characterized in that: The neighboring UUV set corresponding to each UUV is obtained by applying two-dimensional grid coding processing in the local communication network to obtain the neighboring UUV set of the current UUV.

5. The UUV capture method based on distributed model predictive control according to claim 1, characterized in that: The predicted state sequence of the current UUV is obtained according to the following steps: Construct nonlinear UUV dynamic model equations; The current state of the UUV is predicted according to the nonlinear UUV dynamics model equation to obtain the predicted state sequence of the current UUV.

6. The UUV capture method based on distributed model predictive control according to claim 1, characterized in that: During the optimization process, each UUV maintains the following three control objectives: (1) With the moving target as the center, the capture radius converges to a fixed expected value; (2) The UUV forms a surround around the target at the desired angular velocity; (3) The angular spacing between two UUVs distributed around the moving target is the same; Among them, moving targets are encirclement targets.

7. The UUV capture method based on distributed model predictive control according to claim 1, characterized in that: After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again, including: After a control cycle, the current UUV contains control vectors with N steps. In each control cycle, only the optimal control vector of the first step is selected. After the controller outputs to the actuator, after a control cycle, the final control quantity of the next cycle is obtained again through the outer controller and the inner controller.

8. The UUV capture method based on distributed model predictive control according to claim 1, characterized in that: Solve the error model to obtain the current UUV control compensation, including: The feedback gain K of the UUV system in a stable state is selected as the solution of the linear quadratic regulator of the error model and as the control compensation of the current UUV.

9. The UUV capture method based on distributed model predictive control according to claim 1, characterized in that: The outer controller is a DMPC-based trap controller; the inner controller is a Tube-based MPC controller.

10. A UUV capture system based on distributed model predictive control, characterized in that: include: The control system building module is used to establish a distributed control system structure, including an inner controller and an outer controller corresponding to each UUV; The control sequence calculation module is used to obtain the predicted state sequence of the current UUV and the optimal predicted state sequence of the neighboring UUV through the outer controller, and perform optimization to obtain the optimal predicted state sequence and optimal control sequence of the current UUV respectively; The control compensation module is used to obtain the error by subtracting the predicted state sequence of the current UUV from the state quantity of the actual measurement system through the inner controller, and then define the error model through the error, and solve the error model to obtain the control compensation of the current UUV; A control module, used for adding the current UUV control compensation amount to the current UUV's optimal predicted state sequence and optimal control sequence to obtain the current UUV's final control amount; After a control cycle, the current UUV obtains the final control quantity of the next cycle through the outer controller and the inner controller again to carry out the encirclement control of the current UUV.

Images