Unmanned ship cluster obstacle avoidance method based on strategy search MPCC

Abstract

The invention particularly relates to an unmanned ship cluster obstacle avoidance method based on strategy search MPCC. The method comprises the following steps: defining a virtual reference point of an unmanned ship cluster; sampling the relative distance between each unmanned ship in the unmanned ship cluster and the virtual reference point, and solving a preset optimization controller according to a plurality of relative distance samples obtained by sampling to obtain an unmanned ship cluster trajectory; according to the unmanned ship cluster trajectory, calculating a corresponding cooperative obstacle avoidance cost for the relative distance samples, and carrying out weight updating on each relative distance sample based on the cooperative obstacle avoidance cost, so as to determine a cooperative obstacle avoidance strategy according to a weight updating result of each relative distance sample; wherein the collaborative obstacle avoidance strategy comprises a target relative distance; and resolving a multi-target cost function by using the collaborative obstacle avoidance strategy to obtain a control input sequence, and executing the control input sequence to control the unmanned ship to avoid obstacles. According to the method, the obstacles on the water can be efficiently avoided.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an unmanned vessel cluster obstacle avoidance method based on strategy search MPCC. Background Art

[0002] Among related technologies, the collaborative control of multi-agent systems has attracted increasing attention. Benefiting from the high flexibility and stability of unmanned vessel swarms, they have broad applications in a variety of fields, including maritime patrols, resource exploration, and rescue operations. In multi-unmanned vessel swarm collaborative tasks, obstacle avoidance path planning, tracking control, and the design of swarm collaborative control algorithms are key issues to ensure the stable operation of the swarm. Traditional unmanned vessel control algorithms lack high accuracy and reliability, and are unable to accurately, quickly, and reliably track the target vessel or desired path in complex and challenging sea conditions. Model Predictive Contouring Control (MPCC), an advanced control method based on online rolling optimization, offers advantages such as high-precision trajectory tracking, multi-objective optimization capabilities, strong constraint handling capabilities, and suitability for complex dynamic environments. In existing collaborative control algorithms for unmanned vessels based on MPCC, the formation of the swarm is typically preset to fixed parameters. This static setting cannot be flexibly adjusted in dynamic environments, making it difficult for the swarm to achieve an effective trade-off between obstacle avoidance and maintaining formation. In complex water environments, unmanned vessel swarms must simultaneously meet the dual requirements of obstacle avoidance and formation maintenance. Traditional methods, lacking flexibility, often struggle to achieve both, leading to obstacle avoidance failures and disorganized formations. Furthermore, traditional methods lack the ability to adapt to dynamic obstacles and environmental uncertainties, making them difficult to cope with complex and ever-changing practical application scenarios. Fixed formations and preset parameters also limit the efficiency of coordinated control of unmanned vessel swarms, making it impossible to optimize swarm behavior in real time based on mission requirements and environmental changes.

[0003] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0004] The present invention provides an unmanned vessel cluster obstacle avoidance method based on strategy search MPCC, which can realize efficient obstacle avoidance and collaborative control of unmanned vessel clusters in dynamic environments, improve the robustness, safety and adaptability of the cluster system, and thus overcome the defects existing in the existing technology to a certain extent.

[0005] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0006] According to a first aspect of the present invention, a method for obstacle avoidance of an unmanned ship cluster based on a strategy-based search MPCC is provided, the method comprising:

[0007] Define a virtual reference point for the unmanned vessel cluster; the virtual reference point is a global reference point dynamically anchored on the reference path by the unmanned vessel cluster;

[0008] The relative distances between each unmanned vessel in the swarm and a virtual reference point are sampled, and a preset optimization controller is solved based on the multiple relative distance samples obtained to obtain the swarm trajectory. The optimization controller is constructed based on the distributed model predictive contour control framework (MPCC).

[0009] According to the trajectory of the unmanned vessel cluster, the corresponding collaborative obstacle avoidance cost is calculated for the relative distance samples, and the weight of each relative distance sample is updated based on the collaborative obstacle avoidance cost, so as to determine the collaborative obstacle avoidance strategy according to the weight update result of each relative distance sample; wherein the collaborative obstacle avoidance strategy includes the target relative distance;

[0010] The collaborative obstacle avoidance strategy is used to solve the multi-objective cost function to obtain a control input sequence, and the control input sequence is executed to control the unmanned boat to avoid obstacles.

[0011] In some exemplary embodiments, the method further comprises:

[0012] The state of the unmanned ship is defined according to the position and yaw angle of the unmanned ship in the inertial coordinate system, and the ship speed and angular velocity in each direction of the unmanned ship in the hull coordinate system;

[0013] Based on the thrust parameters and torque parameters of the unmanned ship's power system and the state of the unmanned ship, the kinematic model of the unmanned ship in the horizontal plane is defined.

[0014] In some exemplary embodiments, the method further comprises:

[0015] Considering the kinematic model of the unmanned vessel, a multi-objective cost function is constructed based on the distributed model predictive contour control framework (MPCC). The multi-objective cost function is used to achieve path tracking of the unmanned vessel cluster.

[0016] The distributed MPCC is configured as an optimization controller based on hyperparameters; the hyperparameter is the relative distance between the unmanned ship and the virtual reference point, and the formation of the unmanned ship cluster changes according to the change of the hyperparameter.

[0017] In some exemplary embodiments, defining a virtual reference point of the unmanned vessel cluster includes:

[0018] Real-time calculation is performed based on the path parameter equation and the preset speed to obtain the position information of the virtual reference point on the reference path;

[0019] According to the number of unmanned ships, the formation of the unmanned ship cluster, and the initial relative distance, the relative distance range between each unmanned ship and the virtual reference point is configured.

[0020] In some exemplary embodiments, updating the weight of each relative distance sample based on the collaborative obstacle avoidance cost includes:

[0021]

[0022] Among them, α a , α b , β b , β c are adjustment parameters, N is the prediction time domain, is the number of unmanned ships, is the number of adjacent unmanned ships of the i-th unmanned ship, d i，a is the distance between the i-th unmanned ship and the obstacle, d i，j is the distance between the i-th unmanned ship and the j-th unmanned ship, d s is the safe distance between the unmanned ship and other unmanned ships and obstacles; Δζ(k) is the change in the expected distance, ζ r The preset relative distance.

[0023] In some exemplary embodiments, the method further comprises:

[0024] Each relative distance sample is updated until the number of updates reaches the preset upper limit or the state cost function value is lower than the designed threshold.

[0025] In some exemplary embodiments, determining a collaborative obstacle avoidance strategy based on weight update results of each relative distance sample includes:

[0026] Update the relative distance distribution of the unmanned ship cluster through the weighted updated relative distance samples;

[0027] The mean of the updated relative distance distribution is used as the expected distance for the next N steps.

[0028] In some exemplary embodiments, the method further comprises:

[0029] According to the number of unmanned ships and the formation of the unmanned ship cluster, configure the obstacle avoidance recognition range corresponding to each unmanned ship and / or unmanned ship cluster;

[0030] When an obstacle is identified within the obstacle avoidance recognition range, an obstacle avoidance task for the unmanned ship cluster is triggered.

[0031] According to a second aspect of the present invention, a computer program product is provided, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned unmanned ship cluster obstacle avoidance method based on strategy search MPCC is implemented.

[0032] According to a third aspect of the present invention, there is provided an electronic device, comprising:

[0033] A processor and a memory; wherein the memory is used to store executable instructions of the processor; the processor is configured to implement the above-mentioned unmanned ship cluster obstacle avoidance method based on strategy search MPCC when executing instructions by executing the executable instructions.

[0034] According to a fourth aspect of the present invention, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned unmanned ship cluster obstacle avoidance method based on strategy search MPCC is implemented.

[0035] The unmanned boat cluster obstacle avoidance method based on policy search MPCC provided by the embodiment of the present invention configures a virtual reference point for the unmanned boat cluster. When the unmanned boat encounters an obstacle on the water, the relative distance between the unmanned boat and the virtual reference point is sampled and the trajectory of the unmanned boat cluster is calculated. The unmanned boat cluster trajectory can be used to calculate the collaborative obstacle avoidance cost of the relative distance sample, and the weight of each relative distance sample is updated based on the collaborative obstacle avoidance cost to determine the collaborative obstacle avoidance strategy. The collaborative obstacle avoidance strategy is further used to solve the optimization controller to obtain the control input sequence to achieve obstacle avoidance control of the unmanned boat. The distributed MPCC algorithm based on policy search is used to construct an optimization controller. When encountering an obstacle, the optimal obstacle avoidance formation is calculated by the policy search algorithm, and the optimal expected formation is calculated, thereby controlling the unmanned boat cluster to efficiently avoid obstacles on the water. The obstacle avoidance capability and collaborative control efficiency of the unmanned boat cluster in a dynamic environment are improved, providing reliable technical support for realizing autonomous navigation and collaborative operation of the unmanned boat cluster.

[0036] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings are incorporated into and constitute a part of this specification, illustrate embodiments consistent with the present invention, and together with the description, serve to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and it is clear that those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0038] Figure 1A schematic diagram schematically illustrates an exemplary embodiment of the present invention, a method for unmanned ship cluster obstacle avoidance based on strategic search MPCC;

[0039] Figure 2 A schematic diagram schematically illustrates an unmanned ship trajectory in an inertial system and a hull coordinate system according to an exemplary embodiment of the present invention;

[0040] Figure 3 A schematic diagram schematically illustrates a process flow of an exemplary embodiment of the present invention for an unmanned vessel cluster to telescope and transform its formation to avoid obstacles;

[0041] Figure 4 A schematic diagram schematically illustrates an unmanned ship formation according to an exemplary embodiment of the present invention;

[0042] Figure 5 A schematic diagram schematically illustrates a comparison result of the cooperative obstacle avoidance cost of a traditional MPCC and a strategic MPCC in a simulation according to an exemplary embodiment of the present invention;

[0043] Figure 6 A schematic diagram schematically illustrates an electronic device according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0044] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0045] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0046] In related technologies, in traditional model predictive contour control (MPCC) algorithms, the formation of unmanned ship clusters is usually preset as fixed parameters. This static setting cannot be flexibly adjusted in a dynamic environment, resulting in the unmanned ship cluster having difficulty in achieving an effective trade-off between avoiding obstacles and maintaining formation. In complex water environments, unmanned ship clusters need to meet the dual requirements of obstacle avoidance and formation maintenance at the same time. However, due to the lack of flexibility, traditional methods often find it difficult to take both into account, and are prone to obstacle avoidance failures or disorganized formations. In addition, traditional methods lack the ability to adapt to dynamic obstacles and environmental uncertainties, making it difficult to cope with complex and changing practical application scenarios. Fixed formations and preset parameters also limit the efficiency of collaborative control of unmanned ship clusters, making it impossible to optimize cluster behavior in real time according to mission requirements and environmental changes.

[0047] In view of the shortcomings and deficiencies of the existing technology, this example embodiment provides an unmanned ship cluster obstacle avoidance method based on strategy search MPCC. Figure 1 As shown, the method may include the following steps:

[0048] Step S11, defining a virtual reference point of the unmanned vessel cluster; wherein the virtual reference point is a global reference point dynamically anchored on the reference path by the unmanned vessel cluster;

[0049] Step S12: sampling the relative distance between each unmanned vessel in the unmanned vessel cluster and the virtual reference point, and solving a preset optimization controller based on the multiple relative distance samples obtained to obtain the trajectory of the unmanned vessel cluster; wherein the optimization controller is constructed based on the distributed model predictive contour control framework MPCC;

[0050] Step S13, calculating the corresponding collaborative obstacle avoidance cost for the relative distance samples according to the unmanned vessel cluster trajectory, and updating the weight of each relative distance sample based on the collaborative obstacle avoidance cost, so as to determine the collaborative obstacle avoidance strategy according to the weight update result of each relative distance sample; wherein the collaborative obstacle avoidance strategy includes the target relative distance;

[0051] Step S14: using the collaborative obstacle avoidance strategy to solve the optimization controller to obtain a control input sequence, and executing the control input sequence to control the unmanned boat to avoid obstacles.

[0052] The present invention uses swarm formation as a controller hyperparameter and combines it with a strategy search algorithm to dynamically optimize the target formation. This method adaptively adjusts the geometric parameters of the swarm formation (such as relative distance and angle) based on the distribution and motion of environmental obstacles. This method achieves efficient obstacle avoidance while maintaining the swarm's coordination and stability, significantly improving the overall performance and adaptability of unmanned vessel swarms in complex environments.

[0053] Below, the various steps of the unmanned ship cluster obstacle avoidance method based on strategy search MPCC in this example implementation will be described in more detail with reference to the accompanying drawings and embodiments.

[0054] In this example implementation, the unmanned ship cluster obstacle avoidance method based on strategy search MPCC can be executed by an intelligent terminal device installed on an unmanned ship in the unmanned ship cluster. For example, the master unmanned ship in the unmanned ship cluster performs the calculation of the obstacle avoidance strategy, updates the control input sequence data, and pushes the updated control input sequence data to other slave unmanned ships in the cluster. Alternatively, each unmanned ship can perform the calculation of the obstacle avoidance strategy and control input sequence data. Alternatively, it can be executed collaboratively between the intelligent terminal device installed on the unmanned ship and the server side. For example, the server side performs the calculation of the obstacle avoidance strategy and control input sequence data, and pushes the data update results to each unmanned ship in the unmanned ship cluster.

[0055] Exemplarily, the method further includes:

[0056] The state of the unmanned ship is defined according to the position and yaw angle of the unmanned ship in the inertial coordinate system, as well as the ship speed and angular velocity in each direction of the unmanned ship in the hull coordinate system;

[0057] Based on the thrust parameters and torque parameters of the unmanned ship's power system and the state of the unmanned ship, the kinematic model of the unmanned ship in the horizontal plane is defined.

[0058] Specifically, a dynamic model of the unmanned ship can be established in advance. Figure 2 As shown in FIG, a group of unmanned ship formations consisting of 3 unmanned ships is used as an example to illustrate the present method. Of course, in other exemplary embodiments of the invention, the unmanned ship cluster may also include other numbers of unmanned ships, such as 4, 5, 6, 10, 20, etc. A group of multiple unmanned ship formations consisting of 3 ships, where the i-th unmanned ship is denoted as V i , i∈{1,...,3}, and establish I system and B i There are two coordinate systems to describe their motion. System I is the inertial system, and system X is the inertial system. I 、Y I , Z I The axes point to the east, north, and up respectively; B i Fastened to the ship's V i superior, The axes point to the right, forward, and up of the ship respectively.

[0059] Assuming that the unmanned ship has good symmetry and moves at low speed on calm water, the non-diagonal elements of the inertia and damping matrices and the nonlinear terms of the damping matrix are ignored. In addition, considering that the collaborative path tracking control problem is carried out on the horizontal plane, the present invention decouples the ship's motion equation into a horizontal model, which includes longitudinal, lateral and yaw motions. Then, the unmanned ship V i The state is defined as where x i and y i is the coordinate of the unmanned ship in the inertial system; represents the yaw angle; and Respectively represent the ship speed in the y-axis and x-axis directions of the hull coordinate system; ω i The system input is the thrust generated by the thruster. and torque Based on the above, the horizontal dynamic equation of the unmanned ship can be described as:

[0060]

[0061]

[0062] in, and are the diagonal elements of the inertia matrix and the linear part of the damping matrix, respectively.

[0063] Exemplarily, the method further includes:

[0064] Considering the kinematic model of the unmanned vessel, a multi-objective cost function is constructed based on the distributed model predictive contour control framework MPCC; wherein the multi-objective cost function is used to achieve path tracking of the unmanned vessel cluster;

[0065] The distributed MPCC is configured as an optimization controller based on hyperparameters; the hyperparameter is the relative distance between the unmanned ship and the virtual reference point, and the formation of the unmanned ship cluster changes according to the change of the hyperparameter.

[0066] Specifically, based on the unmanned vessel dynamics model, a multi-objective cost function is constructed based on the distributed model predictive contour control framework to achieve path tracking of multiple unmanned vessels. The optimization problem of distributed MPCC is defined as follows:

[0067]

[0068] in, is the contour error term, is the lagged error term, is the profile parameter deviation term, represents the deviation of the profile parameter of the i-th ship from the average value of the profile parameters of its neighboring ships, is an artificial potential field term used to avoid collisions between unmanned ships.

[0069] also, Indicates tracking speed; and Indicates the change in control input.

[0070] Through the objective function of the above optimization problem The system can optimize based on the current state and target of the vessel, ensuring efficient path tracking, collision avoidance, and adjusting control inputs to minimize errors during collaborative navigation. The current state of the unmanned vessel is represented by the unmanned vessel state parameters.

[0071] The model predictive contour control (MPCC) is considered as a i (i.e. the relative distance between the unmanned ship and the virtual reference point) is determined by the optimization controller. The distance between the unmanned ship cluster and the virtual tracking point is set to The model prediction profile controller of each unmanned ship is calculated by the hyperparameter D i The real-time distance between the own ship and the virtual tracking point is elastically constrained. Through strategy search, the formation of the unmanned ship swarm can be adjusted based on the formation parameter ζ. When the swarm encounters an obstacle, it can iteratively determine the optimal formation strategy to avoid the obstacle and maintain the formation. The formation parameter can be the aforementioned hyperparameter, namely the relative distance between the unmanned ship and the virtual reference point.

[0072] In step S11, a virtual reference point of the unmanned ship cluster is defined; wherein the virtual reference point is a global reference point where the unmanned ship cluster is dynamically anchored on the reference path.

[0073] Exemplarily, the definition of a virtual reference point of the unmanned vessel cluster includes:

[0074] Real-time calculation is performed based on the path parameter equation and the preset speed to obtain the position information of the virtual reference point on the reference path;

[0075] According to the number of unmanned ships, the formation of the unmanned ship cluster, and the initial relative distance, the relative distance range between each unmanned ship and the virtual reference point is configured.

[0076] Specifically, refer to Figure 3 、 Figure 4 As shown in the figure, the virtual reference point is the global reference point where the formation is dynamically anchored on the reference path. It can be generated by real-time solution of the path parameter equation and the preset speed to determine the overall propulsion direction of the cluster.

[0077] The distance between the unmanned ship cluster and the virtual tracking point can be set to The model prediction profile controller of each unmanned ship is calculated by the hyperparameter D i The real-time distance between the own ship and the virtual tracking point is elastically constrained. Through strategy search, the formation of the unmanned ship cluster can be flexibly changed according to the formation parameter ζ. When the cluster encounters an obstacle, it can iteratively determine the optimal formation strategy to avoid the obstacle and maintain the formation.

[0078] In step S12, the relative distance between each unmanned ship in the unmanned ship cluster and the virtual reference point is sampled, and the preset optimization controller is solved according to the multiple relative distance samples obtained by sampling to obtain the trajectory of the unmanned ship cluster; wherein, the optimization controller is constructed based on the distributed model predictive contour control framework MPCC.

[0079] For example, the formation parameter (ie, relative distance) ζ of the unmanned ship cluster is set to Gaussian distribution, ζ [k] ~π θ (μ, ∑), k = 1...N. From the distribution π θ (μ, ∑) sampling M S The MPCC problem is solved based on the relative distance to obtain the unmanned ship cluster trajectory.

[0080] The random values ​​generated by this Gaussian distribution may exceed the reasonable range in practical applications, so clipping is required to limit them to a preset reasonable range. The mathematical expression of the clipping action is as follows:

[0081] ζ [k] =max(min(ζ [k] , clip_high), clip_low) (4)

[0082] In step S13, according to the trajectory of the unmanned ship cluster, the corresponding collaborative obstacle avoidance cost is calculated for the relative distance samples, and the weight of each relative distance sample is updated based on the collaborative obstacle avoidance cost, so as to determine the collaborative obstacle avoidance strategy according to the weight update result of each relative distance sample; wherein, the collaborative obstacle avoidance strategy includes the target relative distance.

[0083] For example, based on the cluster trajectory, calculate M S The collaborative obstacle avoidance cost of the relative distance samples is then calculated for each sampled relative distance ζ [k] The weight is given by the collaborative obstacle avoidance cost, and the mean and variance of the strategy distribution are updated accordingly. s After times, the optimal collaborative obstacle avoidance strategy is obtained.

[0084] Specifically, to calculate the optimal obstacle avoidance formation, the present invention designs a reward event, where reward event R = 1 is defined as an observation variable. Let p(R|u) denote the probability of this reward event, which is inversely proportional to the cost of collaborative obstacle avoidance. The maximum likelihood problem formula is:

[0085]

[0086] Where γ = [μ, ∑] is the distribution of the optimal relative distance, μ and ∑ represent the mean and variance of the relative distance decision, respectively.

[0087] In order to find the optimal decision distribution, the present invention uses the maximum expectation algorithm based on Monte Carlo to transform the maximum likelihood problem into:

[0088] logp γ (R=1)=Γ γ (q(u))+D L (q(u), p γ (u|R)) (6)

[0089] Among them, q(u) is the distribution of the relative distance of the unmanned ship, D L () is the relative distance trajectory distribution p between q(u) and reward weight γ Kullback-Leibler (KL) divergence of (u|E). Since KL divergence is always greater than 0, Γ γ (q(u)) is the maximum likelihood problem logp γ (R=1) lower bound.

[0090] In the expectation maximization algorithm, the optimal control strategy is calculated by alternately executing the expectation step and the maximization step. In the expectation step, in order to minimize the KL divergence, the present invention designs the relative distance distribution of the unmanned ship as q(u)=p γ (u|E)∝p(E|u)p γ (u).

[0091] In the maximization step, the policy parameters are updated by maximizing the following likelihood function:

[0092]

[0093] Among them, M s is the number of samples.

[0094] After sampling to a desired distance, we solve the following MPCC optimization problem to obtain the trajectory of the unmanned ship:

[0095]

[0096] The following constraints are met:

[0097]

[0098] Exemplarily, determining a collaborative obstacle avoidance strategy based on the weight update results of each relative distance sample includes:

[0099] Update the relative distance distribution of the unmanned ship cluster through the weighted updated relative distance samples;

[0100] The mean of the updated relative distance distribution is used as the expected distance for the next N steps.

[0101] Exemplarily, the method further includes: performing updating processing on each relative distance sample until the number of updates reaches a preset upper limit or the state cost function value is lower than a designed threshold.

[0102] Specifically, in order to find the obstacle avoidance strategy that minimizes the collaborative obstacle avoidance cost, the present invention designs a weight function to increase the weight of samples with low collaborative obstacle avoidance cost, and then calculates the relative distance ζ of each sample. [k] Weighted by the following collaborative obstacle avoidance costs:

[0103]

[0104] Among them, α a , α b , β a and β c They are adjustment parameters, N is the prediction time domain, is the number of unmanned ships, is the number of adjacent unmanned ships of the i-th unmanned ship, d i，a is the distance between the i-th unmanned ship and the obstacle, d i，j is the distance between the i-th unmanned ship and the j-th unmanned ship, d s is the safe distance between the unmanned ship and other unmanned ships and obstacles; Δζ(k) is the change in the expected distance, ζ r It is a preset relative distance, which aims to keep the search results within a preset reasonable range.

[0105] After calculating the collaborative obstacle avoidance cost, the weight function of each sample is:

[0106]

[0107] Finally, the mean and covariance of the expected distances are updated according to the following equations:

[0108]

[0109] Repeat the above update steps until the number of updates reaches the preset upper limit or the state cost function value is lower than the design threshold, and use the mean of the updated policy distribution as the actual control input.

[0110] In step S14, the coordinated obstacle avoidance strategy is used to solve the optimization controller to obtain a control input sequence, and the control input sequence is executed to control the unmanned vessel to avoid obstacles.

[0111] For example, the mean of the strategy distribution can be used as the optimal relative distance to substitute into the distributed MPCC optimization problem, thereby calculating the control input sequence of the unmanned ship and updating the control input sequence. Specifically, the control input sequence includes the torque parameters and thrust parameters of the unmanned ship's power system.

[0112] In addition, a strategy search can be performed every N time steps to control the swarm of unmanned boats to maintain the swarm formation while avoiding water obstacles. Where N is a positive integer.

[0113] For example, the unmanned vessel cluster obstacle avoidance algorithm based on the strategy search MPCC may include the following calculation process:

[0114] Initialization: initial relative distance distribution π θ (μ, ∑), distributed controller parameters, UAV cluster status;

[0115]

[0116]

[0117] Exemplarily, the method further includes:

[0118] Step S21, configuring the obstacle avoidance recognition range corresponding to each unmanned ship and / or unmanned ship cluster according to the number of unmanned ships and the formation of the unmanned ship cluster;

[0119] Step S22: When an obstacle is identified within the obstacle avoidance recognition range, an obstacle avoidance task for the unmanned vessel cluster is triggered.

[0120] Specifically, obstacle avoidance detection ranges can be pre-configured for each unmanned vessel and / or swarm of unmanned vessels. As the unmanned vessels follow their planned trajectories, they can use radar, laser, sonar, and other equipment to determine in real time whether there are surface and / or underwater obstacles within the obstacle avoidance range. Upon detecting an obstacle within the obstacle avoidance range, the current obstacle avoidance task can be triggered. For example, each unmanned vessel can be configured with a corresponding obstacle avoidance detection range, with each vessel independently detecting obstacles. For example, the obstacle avoidance detection range can be configured based on the size and speed of the unmanned vessel, such as 100 meters, 500 meters, or 1 kilometer. Alternatively, the obstacle avoidance detection range can be configured for the entire swarm of unmanned vessels. For example, the obstacle avoidance detection range can be defined as a circle with a radius of 500 meters, 800 meters, or 1 kilometer, centered around a virtual reference point or the center point of the swarm's formation, based on the number of unmanned vessels, the formation of the swarm, and the speed of the vessels.

[0121] For example, simulation experiments are used to demonstrate the collaborative control effect of the strategy search MPCC. The preset formation parameter ζ of the unmanned ship is set to ζ = 3, and the relative distance between each unmanned ship and the virtual tracking point is set to The trajectory of the unmanned ship formation tracking is set to x = 5cost, y = 5 + 5sint, t∈[-π,π]. The control parameters of MPCC are as follows: the control interval is δ = 0.2s, and the prediction time domain is N = 20. In addition, the cost function of all unmanned ships is the same. At the same time, there is a safety distance d at [4, 9] in the simulation environment. s = 0.5m circular obstacle. For the strategy search algorithm, the maximum number of updates is 2, the number of sampling samples is 20, and the mean and variance of the initial Gaussian strategy distribution are 3 and 2 respectively.

[0122] The unmanned boat cluster obstacle avoidance algorithm based on strategy search MPCC is compared with the traditional MPCC. The traditional MPCC also has the same cost function, but its expected formation is fixed. The expected distance of the traditional MPCC is fixed at 3m, while the expected distance of the strategy search MPCC of the present invention is calculated in real time using the strategy search scheme. Comparing the collaborative obstacle avoidance costs of the above two algorithms to avoid the same water obstacle under the same conditions, Figure 5 The red curve in the figure is the collaborative obstacle avoidance cost of the traditional MPCC, and the blue curve in the figure is the collaborative obstacle avoidance cost of the strategy search MPCC. By comparing the collaborative obstacle avoidance costs of the above two algorithms, it is proved that the strategy search MPCC proposed in this paper has better control performance, ensuring that the unmanned ship cluster successfully avoids obstacles while maintaining the desired formation as much as possible.

[0123] The method provided by the present invention may specifically include the following steps:

[0124] Step 1: Build dynamic and kinematic models for the UAVs and construct a multi-objective cost function based on the distributed model predictive contour control framework to achieve path tracking for multiple UAVs.

[0125] Step 2: The model predictive contour control (MPCC) is considered as an optimization controller determined by a hyperparameter (the relative distance between the unmanned ship and the virtual tracking point). The distance between each unmanned ship in the unmanned ship cluster and the virtual reference point is set based on the formation parameter ζ.

[0126] Step 3: Set the formation parameter ζ of the unmanned ship cluster to Gaussian distribution, ζ [k] ~π θ (μ, ∑), k = 1...N. From the distribution π θ (μ, ∑) sampling M S Relative distance samples ζ are used to solve the MPCC problem based on the relative distance to obtain the unmanned ship cluster trajectory;

[0127] Step 4: Calculate M based on cluster trajectory S The collaborative obstacle avoidance cost of samples, and then the relative distance ζ of each sample [k] The weight is given by the collaborative obstacle avoidance cost, and the mean and variance of the strategy distribution are updated accordingly. s After times, the optimal collaborative obstacle avoidance strategy is obtained;

[0128] Step 5: Substitute the mean of the strategy distribution as the optimal relative distance into the distributed MPCC to calculate the unmanned ship control input sequence, and use it to control the unmanned ship cluster to avoid water obstacles while maintaining the cluster formation.

[0129] The present invention provides an unmanned boat cluster obstacle avoidance algorithm based on policy search MPCC, which obtains the optimal unmanned boat cluster formation in real time through policy search, thereby achieving the purpose of efficiently avoiding obstacles on the water. The main contributions of the present invention include: (1) a distributed MPCC algorithm based on policy search is designed. When encountering obstacles on the water, the optimal obstacle avoidance formation is calculated through the policy search algorithm, thereby controlling the unmanned boat cluster to efficiently avoid obstacles on the water; (2) a policy search algorithm based on Monte Carlo maximum expectation is designed. By maximizing the expected return of the distributed cost and the obstacle avoidance cost, the optimal expected formation is calculated in real time, thereby achieving the purpose of balancing the two indicators of unmanned boat cluster collaboration and obstacle avoidance. Through the above innovative design, the present invention significantly improves the obstacle avoidance capability and collaborative control efficiency of the unmanned boat cluster in a dynamic environment, and provides reliable technical support for realizing autonomous navigation and collaborative operation of the unmanned boat cluster.

[0130] It should be noted that the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention and are not intended to be limiting. It is readily understood that the processes illustrated in the above figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0131] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0132] Figure 6 A schematic diagram of an electronic device suitable for implementing an embodiment of the present invention is shown.

[0133] It should be noted that Figure 6 The electronic device 1000 shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0134] like Figure 6 As shown, electronic device 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in read-only memory (ROM) 1002 or the program loaded from storage portion 1008 into random access memory (RAM) 1003. Various programs and data required for system operation are also stored in RAM 1003. CPU 1001, ROM 1002 and RAM 1003 are connected to each other via bus 1004. Input / output (I / O) interface 1005 is also connected to bus 1004.

[0135] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, and the like; an output section 1007 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 1008 including a hard disk and the like; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 1010 as needed, so that computer programs read therefrom can be installed into the storage section 1008 as needed.

[0136] In particular, according to an embodiment of the present invention, the process described below with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product that includes a computer program carried on a storage medium, the computer program containing program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1009 and / or installed from a removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, the various functions defined in the system of the present application are performed.

[0137] Specifically, the above-mentioned electronic device can be a computer, a tablet computer or a server device.

[0138] It should be noted that the storage medium shown in the embodiments of the present invention can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any storage medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code contained on the storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0139] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0140] The units involved in the embodiments of the present invention may be implemented in software or hardware, and the units described may also be provided in a processor. In some cases, the names of these units do not limit the units themselves.

[0141] It should be noted that, as another aspect, the present application also provides a storage medium, which can be included in an electronic device; or it can exist independently without being installed in the electronic device. The above storage medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the following embodiments. For example, the electronic device can implement the following Figure 1 The individual steps of the method are shown.

[0142] In one embodiment, the present application provides a computer program product, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0143] Furthermore, the above-described figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention and are not intended to be limiting. It is readily understood that the processes illustrated in the above-described figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0144] Other embodiments of the present invention will readily occur to those skilled in the art after considering the specification and practicing the invention herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the claims.

[0145] It should be understood that the present invention is not limited to the exact construction described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof, which is limited only by the appended claims.

Claims

1. A method for unmanned ship cluster obstacle avoidance based on strategy search MPCC, characterized in that: The method comprises: Define a virtual reference point for the unmanned vessel cluster; the virtual reference point is a global reference point dynamically anchored on the reference path by the unmanned vessel cluster; The relative distances between each unmanned vessel in the swarm and a virtual reference point are sampled, and a preset optimization controller is solved based on the multiple relative distance samples obtained to obtain the swarm trajectory. The optimization controller is constructed based on the distributed model predictive contour control framework (MPCC). According to the trajectory of the unmanned vessel cluster, the corresponding collaborative obstacle avoidance cost is calculated for the relative distance samples, and the weight of each relative distance sample is updated based on the collaborative obstacle avoidance cost, so as to determine the collaborative obstacle avoidance strategy according to the weight update result of each relative distance sample; wherein the collaborative obstacle avoidance strategy includes the target relative distance; The coordinated obstacle avoidance strategy is used to solve the optimization controller to obtain a control input sequence, and the control input sequence is executed to control the unmanned boat to avoid obstacles.

2. The method according to claim 1, characterized in that The method further comprises: The state of the unmanned ship is defined according to the position and yaw angle of the unmanned ship in the inertial coordinate system, as well as the ship speed and angular velocity in each direction of the unmanned ship in the hull coordinate system; Based on the thrust parameters and torque parameters of the unmanned ship's power system and the state of the unmanned ship, the kinematic model of the unmanned ship in the horizontal plane is defined.

3. The method according to claim 2, characterized in that The method further comprises: Considering the kinematic model of the unmanned vessel, a multi-objective cost function is constructed based on the distributed model predictive contour control framework MPCC; wherein the multi-objective cost function is used to achieve path tracking of the unmanned vessel cluster; The distributed MPCC is configured as an optimization controller based on hyperparameters; the hyperparameter is the relative distance between the unmanned ship and the virtual reference point, and the formation of the unmanned ship cluster changes according to the change of the hyperparameter.

4. The method according to claim 1, wherein The virtual reference point of the unmanned vessel cluster is defined as follows: Real-time calculation is performed based on the path parameter equation and the preset speed to obtain the position information of the virtual reference point on the reference path; According to the number of unmanned ships, the formation of the unmanned ship cluster, and the initial relative distance, the relative distance range between each unmanned ship and the virtual reference point is configured.

5. The method according to claim 1, wherein The weight of each relative distance sample is updated based on the collaborative obstacle avoidance cost, including: Among them, α a , α b , β b , β c are adjustment parameters, N is the prediction time domain, is the number of unmanned ships, is the number of adjacent unmanned ships of the i-th unmanned ship, d i，a is the distance between the i-th unmanned ship and the obstacle, d i，j is the distance between the i-th unmanned ship and the j-th unmanned ship, d s is the safe distance between the unmanned ship and other unmanned ships and obstacles; Δζ(k) is the change in the expected distance, ζ r The preset relative distance.

6. The method according to claim 5, characterized in that The method further comprises: Each relative distance sample is updated until the number of updates reaches the preset upper limit or the state cost function value is lower than the designed threshold.

7. The method according to claim 5, characterized in that The collaborative obstacle avoidance strategy is determined based on the weight update results of each relative distance sample, including: Update the relative distance distribution of the unmanned ship cluster through the weighted updated relative distance samples; The mean of the updated relative distance distribution is used as the expected distance for the next N steps.

8. The method according to claim 1, characterized in that The method further comprises: According to the number of unmanned ships and the formation of the unmanned ship cluster, configure the obstacle avoidance recognition range corresponding to each unmanned ship and / or unmanned ship cluster; When an obstacle is identified within the obstacle avoidance recognition range, an obstacle avoidance task for the unmanned ship cluster is triggered.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the unmanned ship cluster obstacle avoidance method based on strategy search MPCC according to any one of claims 1 to 9 is implemented.

10. An electronic device, characterized in that: include: processor; Memory; wherein, the memory is used to store executable instructions of the processor; the processor is configured to implement the unmanned ship cluster obstacle avoidance method based on strategy search MPCC as described in any one of claims 1 to 9 when executing instructions by executing the executable instructions.

Images