Swarm robot deployment method and device in dynamic diffusion field, server and medium

Abstract

The application discloses a kind of dynamic diffusion field under group robot deployment method, device, server and medium, belong to robot deployment technical field.It includes: dynamic diffusion field is divided into sub-region, sensor robot is arranged in sub-region;Parabolic pollution diffusion field model is constructed, global field density observer is constructed;The projection operator of diffusion coefficient is constructed, the adaptive update law of diffusion coefficient is constructed;The estimated value of diffusion coefficient is obtained, the estimated value of global field density is obtained;Dynamic diffusion field is divided into several sub-regions, and Voronoi cell is obtained;Estimated centroid is calculated;Sliding mode variable and auxiliary intermediate variable are constructed, and distributed guidance law is constructed;Drive each execution robot to move to the target deployment position, to realize deployment.By constructing global field density observer and speedless distributed guidance law, the deployment of group robot in dynamic diffusion field is realized under the condition that diffusion coefficient is unknown, speed is not measurable and resource is limited.

Description

Technical Field

[0001] This invention relates to the field of robot deployment technology, and in particular to a method, apparatus, server, and medium for deploying swarm robots in a dynamic diffusion field. Background Technology

[0002] With the continued acceleration of global industrialization, the frequency and scale of major environmental accidents such as oil spills and chemical spills are increasing year by year, posing a serious threat to natural ecosystems, human life and health, and socio-economic stability. In emergency response to such high-risk environments, traditional response models rely heavily on manual operation, which is not only inefficient and has limited coverage, but also exposes rescue personnel directly to high concentrations of hazardous substances, posing extremely high safety risks.

[0003] In recent years, with the rapid development of robotics and artificial intelligence, swarm intelligent robot systems have shown great potential in key tasks such as pollution assessment and source location due to their distributed collaborative advantages. They can significantly improve the accuracy and efficiency of emergency response while reducing the risk of casualties.

[0004] In related research, coverage control of static spatial fields (pollution independent of time) has received widespread attention. However, pollution diffusion in real-world scenarios often manifests as dynamic spatial fields (pollution is time-dependent), making coverage control more challenging. Especially in typical scenarios such as oil spills and toxic gas diffusion, the diffusion coefficient of pollutants is often unknown. This parameter-unknown characteristic makes traditional collaborative deployment methods relying on precise models or global information difficult to apply directly. How to achieve precise deployment of robots in highly perceptible areas within dynamic and parameter-unknown diffusion fields has become a pressing technical challenge. Summary of the Invention

[0005] This invention provides a method, apparatus, server, and medium for deploying swarm robots in a dynamic diffusion field, to solve the problem that existing technologies struggle to achieve precise robot deployment in dynamic pollution diffusion fields with unknown diffusion coefficients.

[0006] In a first aspect, embodiments of the present invention provide a method for deploying swarm robots in a dynamic diffusion field, comprising: The dynamic diffusion field is divided into multiple non-overlapping sub-regions, and sensing robots are deployed in the sub-regions to collect the local field density of their respective sub-regions. A parabolic pollution diffusion field model with an unknown diffusion coefficient is constructed, and a global field density observer is constructed based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot. Construct a projection operator for the diffusion coefficient, and construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient; The estimated value of the diffusion coefficient is obtained according to the adaptive update law of the diffusion coefficient, and the estimated value of the diffusion coefficient is input into the global field density observer to obtain the estimated value of the global field density. Based on the real-time position of the execution robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology to obtain the Voronoi cavity corresponding to each execution robot. Based on the estimated value of the global field density, the estimated centroid of the Voronoi cell corresponding to each execution robot is calculated, which is used as the target deployment position of the execution robot; Based on the estimated centroid, sliding mode variables and auxiliary intermediate variables are constructed, and a distributed guidance law is constructed based on the sliding mode variables and auxiliary intermediate variables. According to the distributed guidance law, each execution robot is driven to move towards the target deployment location to achieve deployment.

[0007] Secondly, embodiments of the present invention also provide a swarm robot deployment device in a dynamic diffusion field, comprising: The first partitioning module is used to divide the dynamic diffusion field into multiple non-overlapping sub-regions, and to deploy sensing robots in the sub-regions. The sensing robots are used to collect the local field density of their respective sub-regions. The observer construction module is used to construct a parabolic pollution diffusion field model with an unknown diffusion coefficient, and to construct a global field density observer based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot. An adaptive update law construction module is used to construct a projection operator for the diffusion coefficient, and to construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient. The acquisition module is used to obtain an estimated value of the diffusion coefficient according to the adaptive update law of the diffusion coefficient, and input the estimated value of the diffusion coefficient into the global field density observer to obtain an estimated value of the global field density. The second partitioning module is used to divide the dynamic diffusion field into several sub-regions based on the real-time position of the execution robot using Voronoi partitioning technology, thereby obtaining the Voronoi cell cavity corresponding to each execution robot. The centroid estimation module is used to calculate the estimated centroid of the Voronoi cell corresponding to each execution robot based on the estimated value of the global field density, and use it as the target deployment position of the execution robot; A distributed guidance law construction module is used to construct sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and to construct a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables; The drive module, according to the distributed guidance law, drives each execution robot to move towards the target deployment location to achieve deployment.

[0008] Thirdly, embodiments of the present invention also provide a server, comprising: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the swarm robot deployment method in a dynamic diffusion field as provided in the above embodiments.

[0009] Fourthly, embodiments of the present invention also provide a medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the swarm robot deployment method in a dynamic diffusion field as provided in the above embodiments.

[0010] The present invention provides a method, apparatus, server, and medium for deploying swarm robots in a dynamic diffusion field. The method divides the dynamic diffusion field into multiple non-overlapping sub-regions, deploys sensing robots within each sub-region, and the sensing robots collect the local field density of their respective sub-regions. A parabolic pollution diffusion field model with an unknown diffusion coefficient is constructed. A global field density observer is built based on the parabolic pollution diffusion field model, the local field density, and the real-time positions of the sensing robots. A projection operator for the diffusion coefficient is constructed, and an adaptive update law for the diffusion coefficient is built based on the projection operator. An estimated value of the diffusion coefficient is obtained according to the adaptive update law, and the estimated value of the diffusion coefficient is input... The global field density is estimated by inputting into the global field density observer. Based on the real-time position of the execution robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology, resulting in Voronoi cavities for each execution robot. Based on the estimated global field density, the estimated centroid of the Voronoi cavity for each execution robot is calculated, serving as the target deployment position for the execution robot. Sliding mode variables and auxiliary intermediate variables are constructed based on the estimated centroids, and a distributed guidance law is built based on these variables. The distributed guidance law drives each execution robot to move towards the target deployment position to achieve deployment. By constructing a global field density observer that couples an adaptive update law with a projection operator, accurate estimation of the global field density is achieved under conditions where only local measurements are relied upon and the diffusion coefficient is unknown. Furthermore, a distributed guidance law without velocity measurement is designed based on the estimated centroids to drive the execution robots to deploy towards the target position. Attached Figure Description

[0011] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0012] Figure 1 This is a flowchart of the swarm robot deployment method under a dynamic diffusion field provided in Embodiment 1 of the present invention; Figure 2 This is a motion trajectory diagram of the sensing robot in the swarm robot deployment method under dynamic diffusion field provided in Embodiment 1 of the present invention; Figure 3 This is a graph showing the global field density estimation error variation of the swarm robot deployment method under a dynamic diffusion field provided in Embodiment 1 of the present invention. Figure 4 This is a diagram illustrating the collaborative deployment of the execution robots in the swarm robot deployment method under a dynamic diffusion field provided in Embodiment 1 of the present invention. Figure 5 This is a flowchart of the swarm robot deployment method under a dynamic diffusion field provided in Embodiment 2 of the present invention; Figure 6 This is a graph showing the event triggering frequency variation of the distributed guidance law in the swarm robot deployment method under a dynamic diffusion field provided in Embodiment 2 of the present invention. Figure 7 This is a schematic diagram of the swarm robot deployment device under a dynamic diffusion field provided in Embodiment 3 of the present invention; Figure 8 This is a structural diagram of the server provided in Embodiment 4 of the present invention. Detailed Implementation

[0013] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0014] Example 1 Figure 1 This is a flowchart of a swarm robot deployment method in a dynamic diffusion field provided in Embodiment 1 of the present invention. This embodiment is applicable to emergency response to environmental accidents such as oil spills and toxic gas diffusion, enabling the collaborative deployment of swarm robots in a high-perception area under a dynamic pollution diffusion field with an unknown diffusion coefficient. The specific steps include: Step 110: Divide the dynamic diffusion field into multiple non-overlapping sub-regions, and deploy sensing robots in the sub-regions. The sensing robots are used to collect the local field density of their respective sub-regions.

[0015] A dynamic diffusion field refers to the distribution of physical quantities that changes continuously in time and space. Its core characteristic is its dynamic nature; the values ​​in the field not only change with spatial location but also evolve continuously over time until a stable state is reached or the boundary conditions change.

[0016] A sensor robot is an autonomous mobile device equipped with various sensors, whose core function is to sense and collect environmental information. It can carry devices such as gas concentration sensors, temperature sensors, or cameras, moving within a designated area and recording key data about its surrounding environment in real time. The main task of a sensor robot is to act as a mobile monitoring station, converting the collected physical quantities into processable electrical signals or digital information, providing raw data support for subsequent analysis and decision-making.

[0017] Local field density refers to the intensity of a physical quantity at a specific spatial point or within a small region, such as the concentration of pollutants, temperature, or radiation intensity at a location. It reflects the instantaneous state of the environment in that local area and is the basic unit constituting the entire dynamic field. Because sensors can only measure values ​​at their own location, local field density has spatial limitations and is real-time; a complete global field distribution can only be derived from local measurements at multiple different locations.

[0018] For example, the entire dynamic diffusion field area can be... Divided horizontally into The intervals are divided vertically into: The intervals are divided by setting points. interval Divided into The number of sub-intervals, the first The number of sub-intervals is denoted as ,satisfy Setting up points interval Divided into Let the number of subintervals be denoted as i.e., the i-th subinterval. The number of sub-intervals is ,satisfy Based on the above division, we obtain Non-overlapping rectangular sub-regions ,in In each sub-region One sensor robot is deployed inside, totaling The platform is numbered according to its sub-regions, the first... Line 1 The robot number in the column is Its real-time position is recorded as ,in and Representative number is The real-time coordinates of the sensing robots. The sensing robots are used to collect the local field density of their respective sub-regions. When each sensing robot collects the local field density of its sub-region, it uses a spatial distribution function... Filtering is performed, among which ,in, For any point to be measured in the region, The sensor robot's sensing area size (i.e., the surrounding side length is...) (a square) For a side length of The square sensing area, when the point to be measured Located at the center of the current position of the sensing robot, with a side length of Square sensing area When the point is within the sensor's perception region, the spatial distribution function outputs 1, indicating that the local field density at that point can be sensed; otherwise, it outputs 0. The spatial distribution function is used to determine whether any spatial point is within the sensor robot's current perception region, outputting 0 or 1 as a filtering signal, thus incorporating only the robot's actually measurable local field density information into subsequent observer calculations. To avoid collisions between mobile sensor robots and ensure their movement range does not exceed their respective sub-regions, a projection operator is introduced to constrain the robot's speed. Taking direction as an example, the projection operator ,when When this is the case, it indicates that the robot can move freely within a safe area. When When, it indicates that the robot's velocity is smoothly compressed as it attempts to exceed the upper bound. When the robot attempts to exceed the lower bound, its velocity is smoothly compressed, where... For the projection constraint range, Let x be the desired velocity in the x-direction. Similarly, the y-direction projection operator... Defined using the same rules. For example... Figure 2 The diagram shows the motion trajectory of the sensing robots after introducing projection operators in the x and y directions. Green triangles mark the initial positions of the four sensing robots, and blue circles mark their final positions. The trajectory characteristics in the diagram show that the motion range of all sensing robots is strictly limited to their respective sub-regions, with no boundary violations occurring. Finally, the actual motion speed of the sensing robots is set to... and And when the initial position satisfies In this way, the above design can ensure that the robot's position is always strictly constrained within its own sub-region at all times, thereby avoiding cross-region collisions while collecting local field density.

[0019] Step 120: Construct a parabolic pollution diffusion field model with an unknown diffusion coefficient, and construct a global field density observer based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot.

[0020] The diffusion coefficient is a physical quantity that describes the rate at which a substance diffuses in a medium; it is usually represented by the symbol [symbol missing]. The diffusion coefficient reflects the migration ability of molecules or particles due to thermal motion. A larger diffusion coefficient indicates a wider diffusion range and a faster rate at which the concentration distribution becomes uniform per unit time. Conversely, a smaller diffusion coefficient indicates that the substance tends to remain in its original position, resulting in a longer concentration gradient. Mathematically, the diffusion coefficient is the coefficient multiplied by the second spatial derivative term in the governing equation, directly affecting the rate of concentration distribution evolution over time. In practical engineering scenarios, the diffusion coefficient is often difficult to obtain accurately in advance. For example, in oil spills, the diffusion rate depends on various factors such as oil properties, seawater temperature, and ocean current speed, which are difficult to measure in real time at the time of the incident. In toxic gas leaks, the diffusion coefficient is also affected by complex conditions such as atmospheric stability, wind speed, and terrain, making it impossible to know in advance. Therefore, the diffusion coefficient becomes an unknown parameter.

[0021] The parabolic pollution diffusion field model is a model of the spatiotemporal evolution of pollution concentration described by parabolic partial differential equations. For example, the parabolic pollution diffusion field model is constructed as follows: ,in, Indicates time Location The actual field density at that location, This represents the rate of change of the actual field density over time. For the Lass-Rapp operator, The coefficient of the reaction term, The diffusion coefficient is unknown. The formula for the parabolic pollution diffusion field model mainly consists of two parts: the first part is the diffusion term. The first part reflects the migration process of pollutants from areas of high concentration to areas of low concentration. The Laplace operator characterizes the degree of difference between the concentration and the surrounding space, while the diffusion coefficient controls the rate of this process. The second part is the reaction term. This describes the growth or decay of the pollutant itself. Furthermore, for any location... This model requires initial conditions. Given the initial concentration distribution and boundary conditions Only by ensuring that the concentration at the specified boundary remains at 0 can the evolution of the entire diffusion process be uniquely determined.

[0022] A global field density observer is a mathematical algorithm or system that estimates the overall distribution state online based on local measurement information. In practical applications, it is often impossible to directly obtain the physical quantity values ​​at every point in the entire region; instead, real-time data from local locations can only be collected through a limited number of sensors. The core function of the observer is to use these discrete, local measurements, combined with a theoretical model describing the evolution of the physical quantity, to dynamically calculate the complete distribution of the physical quantity across the entire region.

[0023] For example, a global field density observer can be constructed based on the local field density obtained in step 110 and the constructed parabolic pollution diffusion field model. The global field density observer is constructed as follows: ,in, This represents the estimated global field density at time t. The rate of change of the global field density estimate over time. This is an estimate of the unknown diffusion coefficient. The coefficient of the reaction term, It is a positive gain parameter. It is a spatial distribution function. The real-time position of the sensor robot is measured at time t. Let be the local field density at time t. For feedback correction items, The number of sub-regions into which the dynamic diffusion field region Ω is divided along the x-axis. To dynamically diffuse the field area The number of sub-regions divided along the y-direction. This is the number of the sub-region in the x-direction. This is the number of the sub-region along the y-axis. For dynamic diffusion field Any point within the field. The main body of the observer employs partial differential equations of the same form as the real diffusion field to simulate the evolution of pollution diffusion. Simultaneously, a correction term driven by local measurements from a sensing robot is introduced into the equation. This correction term utilizes the spatial distribution function. Measurement data within the perceptible area surrounding each sensing robot's current position are selected. The difference between the actual local field density and the current estimate is used as feedback, multiplied by the gain and diffusion coefficient estimates, and then superimposed into the observer equation, thereby achieving online correction of the global field density estimate. Simultaneously, the observer's initial and boundary conditions satisfy those of the original diffusion field model, ensuring that the estimation process conforms to actual physical constraints. In this way, the observer can dynamically reconstruct the field density distribution of the entire region relying solely on local measurement information.

[0024] Step 130: Construct a projection operator for the diffusion coefficient, and construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient.

[0025] The projection operator is a mathematical tool used to constrain the range of values ​​for variables. Its core function is to always limit the direction of change of the constrained variable within a predefined feasible region. In parameter estimation problems, when the parameters to be estimated have known physical boundaries, directly using an unconstrained update law may lead to parameter estimates exceeding a reasonable range, resulting in system instability or estimation divergence. The projection operator addresses this problem through a safe interval. When the parameter estimate is within the safe interval, the projection operator allows it to adjust freely. When the estimate attempts to exceed the interval boundary, the projection operator smoothly compresses its update direction. This smoothing avoids the discontinuities caused by abrupt truncation, ensuring the continuity and smoothness of the parameter estimation process, while mathematically guaranteeing that the estimate will never deviate from the predefined feasible region.

[0026] For example, the specific process of constructing the projection operator of the diffusion coefficient is as follows: first, determine the theoretical range of values ​​for the diffusion coefficient. , Given a known lower bound, Given a known upper bound, a projection constraint range is introduced. The compact set of the diffusion coefficient is defined as follows: This serves as the allowable range of motion for the estimated value. Simultaneously, an indicator error driving term reflecting the parameter adjustment requirements is defined. ,in, Let be the partial derivative of the global field density estimate in the x-direction. The partial derivative of the global field density estimate in the y-direction is used. This error-driving term is calculated from the spatial gradient of the global field density estimate. A larger gradient indicates a more drastic change in the current estimated field, and a stronger need for parameter adjustment. Therefore, the projection operator for the diffusion coefficient is set to... ,in, As an error-driven term, It is a positive gain coefficient. Indicates the upper bound compression correction amount and This represents the lower bound compression correction amount, when When, it indicates the use of unconstrained update amounts; when When, it indicates that the update amount is compressed and corrected, so that the update magnitude decreases smoothly as the distance exceeds the limit; when Similarly, this indicates a compression correction of the update amount. This piecewise design always constrains the estimated value within a preset reasonable range, while avoiding discontinuities caused by abrupt truncation through smooth compression. After obtaining the projection operator of the diffusion coefficient, an adaptive update law for the diffusion coefficient is constructed based on the projection operator. This refers to using the output of the projection operator as the rate of change of the diffusion coefficient estimate. This combination of the update law and the projection operator allows the diffusion coefficient estimation process to continuously optimize using real-time observation data while always maintaining the estimate within a compact set. Meanwhile, the continuity of the projection operator ensures smooth changes in the estimated value.

[0027] Step 140: Obtain an estimated value of the diffusion coefficient according to the adaptive update law of the diffusion coefficient, and input the estimated value of the diffusion coefficient into the global field density observer to obtain an estimated value of the global field density.

[0028] For example, after obtaining the adaptive update law of the diffusion coefficient in step 130, the output of the projection operator is used as the rate of change of the diffusion coefficient estimate. The diffusion coefficient estimate at each time step is obtained by integrating this rate of change over time. In this way, the diffusion coefficient estimate is continuously adjusted over time, providing a crucial input parameter for the global field density observer. It should be noted that the solutions for the diffusion coefficient estimate and the global field density estimate are coupled and iteratively performed. Initially, there are initial estimates for both the diffusion coefficient and the global field density. Subsequently, the error driving term is first calculated based on the current global field density estimate. The diffusion coefficient estimate for the next time step is obtained through an adaptive update law. Simultaneously, the updated diffusion coefficient estimate is substituted into the global field density observer, combined with the local field density collected by the sensing robot, to calculate the global field density estimate for the next time step. This process is repeated iteratively, with the two estimates being updated and corrected alternately in each step, ultimately achieving common convergence. It should be noted that this process's ability to achieve common convergence relies on a series of strict conditions: the original diffusion field must satisfy the conditions that the diffusion coefficient is bounded and positive definite, and the region partitioning rule of step 110 must be met; the global field density observer and the adaptive update law must be used in coupled fashion; and the motion speed of the sensing robot must satisfy the projection operator constraints obtained in step 110. Furthermore, the parameters in the global field density observer and the adaptive update law must satisfy the following inequality constraints: , , , ,in, For parameter combination terms; The coefficient of the reaction term, It is a positive real number. Only when all these conditions are met simultaneously, such as... Figure 3 As shown, only by ensuring that the global field density estimation error approaches zero over time can the global field density estimate eventually converge to the true field distribution.

[0029] Step 150: Based on the real-time position of the execution robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology to obtain the Voronoi cell cavity corresponding to each execution robot.

[0030] Voronoi partitioning is a geometric method for spatially dividing a planar region according to the nearest neighbor principle. The basic idea is to divide the entire plane into multiple regions given several discrete seed points, such that the distance from any point within each region to a seed point in that region is less than the distance to any other seed point. The result of Voronoi partitioning is unique, and the partitioned regions do not overlap, completely cover the entire plane, and their boundaries are formed by the perpendicular bisectors of the lines connecting the seed points.

[0031] A Voronoi cavity is an independent subregion obtained after Voronoi partitioning. Each cavity corresponds to a seed point, and its geometry depends on the distribution of the seed points and the distance metric. When the seed points are evenly distributed, the cavity is hexagonal or approximately hexagonal. When the seed points are unevenly distributed, the size and shape of the cavity also change. All points within a cavity have the same nearest neighbor relationship, meaning that any position within the cavity is closest to its corresponding seed point. The cavity boundary is formed by the perpendicular bisector of the line connecting adjacent seed points, and points on the boundary are equidistant from two or more seed points. Physically, a Voronoi cavity can be understood as the influence region of each seed point.

[0032] For example, based on the real-time position of the executing robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology. The core idea is to dynamically segment the space based on the principle of closest proximity. Specifically, for any executing robot... The corresponding Voronoi cavities Defined as any spatial point within the region that satisfies the condition that the distance to the robot is no greater than the distance to any other robot. The set, i.e. ,in, and To execute the robot number, and For the number The real-time location and number of the execution robot are as follows: The real-time position of the execution robot, where q is the dynamic diffusion field. Any point within the range. This division rule depends on the real-time positions of all executing robots. As the robots move, the boundaries of each cavity adjust in real time, ensuring that each robot is always responsible for the area closest to it. Ultimately, the entire search area is divided into... The union of non-overlapping Voronoi cavities, i.e. Each cell cavity uniquely corresponds to an execution robot, which serves as the responsible area for subsequent execution of covered tasks.

[0033] Step 160: Based on the estimated value of the global field density, calculate the estimated centroid of the Voronoi cell corresponding to each execution robot, and use it as the target deployment position of the execution robot.

[0034] Estimated mass refers to the total estimated value of a continuously distributed physical quantity within a specified region, obtained by integrating the integral. It reflects the total content or intensity of that physical quantity within the region; a larger integral value indicates a greater cumulative amount of that physical quantity within the region. Mathematically, estimated mass is obtained by summing (integrating) the estimated values ​​at every tiny unit within the region, and it is a fundamental statistical measure describing the overall characteristics of the region. Estimated centroid refers to the weighted average position of the region calculated using the estimated values ​​of a physical quantity as weights. It is equivalent to the center of gravity of the physical quantity's distribution within the region. If the physical quantity is unevenly distributed, the centroid will naturally shift towards areas with higher estimated values. If the distribution is uniform, the centroid will be located at the geometric center of the region.

[0035] For example, the estimated centroid of the Voronoi cell corresponding to each executing robot can be calculated based on the estimated global field density obtained in step 140, and its mathematical expression is as follows: ,in, Let be the estimated mass of the Voronoi cavity corresponding to the i-th executing robot. Let be the estimated centroid of the Voronoi cavity corresponding to the i-th executing robot. The estimated centroid is equivalent to the centroid of the contamination distribution within that cavity. If the contamination distribution is uneven, the centroid will naturally shift towards the region with higher estimated field density. If the contamination is uniformly distributed, the centroid will be located at the geometric center of the cavity. Since the actual field density is unknown, the actual centroid cannot be directly calculated; therefore, the estimated centroid is used as the target deployment position of the executing robot.

[0036] Step 170: Construct sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and construct a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables.

[0037] Sliding mode variables are a core concept in sliding mode control theory. They are typically defined as a comprehensive function of the system's state deviation. When the actual state of the system deviates from the desired state, the sliding mode variables quantify the degree of this deviation and serve as a key indicator guiding the system's control behavior. The basic idea of ​​sliding mode control is to design a control law that forces the system's state of motion to reach and remain on a specific manifold where the sliding mode variables are zero within a finite time. The dynamic characteristics of the system thereafter are determined by this sliding surface, thus achieving robust control against system uncertainties and external disturbances.

[0038] Auxiliary intermediate variables are virtual state variables artificially introduced to address certain difficulties encountered in controller design. In control systems, sometimes some states are difficult to measure directly, or the controller expression is too complex. In such cases, one or more auxiliary variables can be constructed to indirectly reflect the internal dynamics of the system or compensate for unknown information by establishing differential or algebraic relationships between them and measurable states. Auxiliary intermediate variables themselves may not have explicit physical meaning, but their introduction simplifies the controller structure, makes the design of control laws more flexible, and ultimately achieves the expected control objectives. Distributed guidance laws refer to a class of cooperative control methods that utilize only the information of a single agent and its local neighbors to calculate control commands. In multi-agent systems, each agent does not rely on global information or a central controller, but independently generates the next motion command based on its current state and its relative relationship with neighboring agents.

[0039] For example, after obtaining the estimated centroid of the Voronoi cavity corresponding to each executing robot in step 160, sliding mode variables and auxiliary intermediate variables are constructed based on this as the core reference. First, the sliding mode variables are constructed as follows: ,in, This represents the approach gain coefficient. Indicates the damping gain coefficient. Represents the robustness gain coefficient. Let be the sliding mode variable of the i-th robot at time t. This represents the real-time position of the i-th executing robot at time t. Let be the estimated centroid of the i-th Voronoi cell at time t. Let be the rate of change of the sliding mode variable of the i-th robot over time. This formula, by taking the current sliding mode variable, the robot's position, and the estimated centroid as inputs, calculates the sliding mode variable for the next time step after differentiation. Its function is to map the deviation between the robot's position and the target centroid into sliding mode dynamics, providing a robust basis for subsequent control. Secondly, the intermediate amplitude variable is constructed as... ,in, This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. Let be the auxiliary intermediate variable for the i-th robot at time t. Let t be the real-time position of the i-th executing robot. Let represent the rate of change of the auxiliary intermediate variable of the i-th robot over time. This variable depends only on the robot's own position information and is used to compensate for velocity information that cannot be directly measured in the controller. By constructing the above auxiliary intermediate variable and sliding mode variable, the velocity term in the guidance law is directly eliminated, thus designing a truly velocity-free distributed guidance law, completely eliminating the dependence on velocity measurement. The final constructed distributed guidance law is as follows: ,in, This represents the sliding mode feedback gain coefficient. This represents the total gain coefficient of the auxiliary channel. This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. For a moment The sliding mode variable of the i-th executing robot. For a moment The auxiliary intermediate variable for the i-th executing robot, For a moment The real-time position of the i-th executing robot. Let be the guidance law for the i-th executing robot. This guidance law only requires the sliding mode variable, auxiliary intermediate variable, and robot position as inputs, and obtains the control command through weighted calculation. .

[0040] Step 180: According to the distributed guidance law, drive each execution robot to move towards the target deployment location to achieve deployment.

[0041] For example, according to the distributed guidance law, each execution robot is driven to move towards the target deployment location to achieve deployment. The core lies in the control commands obtained in step 170. Substituting these equations into the dual integrator dynamics model of the robot, the actual motion trajectory is generated through integration. The dual integrator dynamics model is described by the following set of differential equations: ,in, Let be the derivative of the position coordinates of the i-th executing robot. Let be the speed of the i-th executing robot. Let be the derivative of the velocity of the i-th executing robot. Let be the guidance law for the i-th executing robot. This model characterizes the fundamental physical laws governing robot motion, and the control input... Acting directly on acceleration, velocity is obtained through a single integration. The position information is obtained by integrating again. In actual deployment, the updated guidance law is input as an acceleration command into the model, driving the robot to adjust its speed and position in real time, gradually moving towards the target deployment position, i.e., the estimated centroid of the Voronoi cavity. Since the guidance law design completely eliminates the velocity term, relying only on sliding mode variables, auxiliary intermediate variables, and robot position information, even at high speeds... Under conditions where direct measurement is not possible, the integral action of this dynamic model, such as... Figure 4 As shown, the robots can achieve precise position tracking and eventually converge to their estimated centroids. Over time, the estimated centroids asymptotically match the true centroids, thus completing the collaborative deployment task of the swarm of robots in a dynamic diffusion field.

[0042] This embodiment divides the dynamic diffusion field into multiple non-overlapping sub-regions, deploys sensing robots within these sub-regions, and uses these robots to collect local field densities within their respective sub-regions. A parabolic pollution diffusion field model with an unknown diffusion coefficient is constructed. Based on this model, the local field density, and the real-time positions of the sensing robots, a global field density observer is built. A projection operator for the diffusion coefficient is constructed, and an adaptive update law for the diffusion coefficient is built based on this operator. An estimated value of the diffusion coefficient is obtained according to the adaptive update law, and this estimated value is input into the global field density observer to obtain the global field density. The process involves estimating the global field density; dividing the dynamic diffusion field into several sub-regions using Voronoi partitioning based on the real-time position of the executing robot, obtaining the Voronoi cavity corresponding to each executing robot; calculating the estimated centroid of the Voronoi cavity corresponding to each executing robot based on the estimated global field density, which serves as the target deployment position of the executing robot; constructing sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and then constructing a distributed guidance law based on these variables; and finally, driving each executing robot to move towards the target deployment position according to the distributed guidance law. By constructing a global field density observer that couples an adaptive update law with a projection operator, accurate estimation of the global field density is achieved under the condition of relying only on local measurements and having an unknown diffusion coefficient. Furthermore, a distributed guidance law without velocity measurement is designed based on the estimated centroid to drive the executing robot to deploy to the target position.

[0043] Example 2 Figure 5This is a flowchart of a swarm robot deployment method under a dynamic diffusion field provided in Embodiment 2 of the present invention. This embodiment is based on the above embodiment and optimized. After the step of constructing sliding mode variables and auxiliary intermediate variables according to the estimated centroid, and constructing a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables, it further includes: constructing an event triggering error function, establishing an event triggering mechanism based on the event triggering error function, and outputting a new triggering time when the triggering condition is met, which is used to update the distributed guidance law.

[0044] See Figure 5 The cooperative control method for the multiphase indirect matrix converter includes: Step 210: Divide the dynamic diffusion field into multiple non-overlapping sub-regions, and deploy sensing robots in the sub-regions. The sensing robots are used to collect the local field density of their respective sub-regions.

[0045] Step 220: Construct a parabolic pollution diffusion field model with an unknown diffusion coefficient, and construct a global field density observer based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot.

[0046] Step 230: Construct a projection operator for the diffusion coefficient, and construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient.

[0047] Step 240: Obtain an estimated value of the diffusion coefficient according to the adaptive update law of the diffusion coefficient, and input the estimated value of the diffusion coefficient into the global field density observer to obtain an estimated value of the global field density.

[0048] Step 250: Based on the real-time position of the execution robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology to obtain the Voronoi cell cavity corresponding to each execution robot.

[0049] Step 260: Based on the estimated value of the global field density, calculate the estimated centroid of the Voronoi cell corresponding to each execution robot, and use it as the target deployment position of the execution robot.

[0050] Step 270: Construct sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and construct a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables.

[0051] Step 280: Construct an event triggering error function, establish an event triggering mechanism based on the event triggering error function, and output a new triggering time when the triggering conditions are met, which is used to update the distributed guidance law.

[0052] For example, a time-triggered mechanism can be introduced based on the distributed guidance law. First, an event-triggered error function is constructed. ,in, This represents the sliding mode feedback gain coefficient. This represents the total gain coefficient of the auxiliary channel. This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. Let be the sliding mode variable of the i-th executing robot at time t. Let be the auxiliary intermediate variable for the i-th executing robot at time t. Let t be the real-time position of the i-th executing robot. Let's define the guidance law for the i-th robot at time t, under the introduction of an event-triggered error function. The event-triggered error function measures the deviation between the control input if it were updated immediately and the currently executed control input. When the robot's actual state deviates only slightly from the ideal control requirements, the old control instructions can be used temporarily without triggering an update. Based on this error function, the event-triggered mechanism is set as follows: That is, whenever the square of the error Exceeding the dynamic threshold At that time, a new trigger moment is generated. Used to update distributed guidance laws Among them, dynamic threshold The evolutionary law is derived from differential equations Description, in which For threshold coefficient, Let be the error feedback coefficient. This equation enables the threshold to adaptively adjust according to the error magnitude; the larger the error, the faster the threshold decreases, and the more sensitive the triggering. When the error is small, the threshold decays slowly, reducing unnecessary triggering. The final distributed guidance law can be expressed as: ,in, This represents the sliding mode feedback gain coefficient. This represents the total gain coefficient of the auxiliary channel. This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. For a moment The sliding mode variable of the i-th executing robot. For a moment The auxiliary intermediate variable for the i-th executing robot, For a moment The real-time position of the i-th executing robot. Let be the guidance law for the i-th executing robot at time t, given the introduction of an event-triggered error function. Under the event-triggered mechanism, this guidance law applies to adjacent trigger times. The internal remains constant, such as Figure 6As shown, the value is updated only when the triggering condition is met. While ensuring the system's convergence performance, this effectively reduces the triggering frequency of the guidance law, thereby significantly reducing the computation and communication overhead of the control command.

[0053] Step 290: According to the distributed guidance law, drive each execution robot to move towards the target deployment location to achieve deployment.

[0054] This embodiment divides the dynamic diffusion field into multiple non-overlapping sub-regions, deploys sensing robots within these sub-regions, and uses these robots to collect local field densities within their respective sub-regions. A parabolic pollution diffusion field model with unknown diffusion coefficients is constructed. Based on this model, the local field density, and the real-time positions of the sensing robots, a global field density observer is built. A projection operator for the diffusion coefficients is constructed, and an adaptive update law for the diffusion coefficients is built based on this operator. An estimated value of the diffusion coefficients is obtained according to the adaptive update law, and this estimated value is input into the global field density observer to obtain an estimated value of the global field density. Based on the real-time positions of the robots, Voronoi... A partitioning technique is used to divide the dynamic diffusion field into several sub-regions, obtaining the Voronoi cavity corresponding to each execution robot. Based on the estimated global field density, the estimated centroid of the Voronoi cavity corresponding to each execution robot is calculated, serving as the target deployment position of the execution robot. Sliding mode variables and auxiliary intermediate variables are constructed based on the estimated centroids, and a distributed guidance law is constructed based on these variables. An event triggering error function is constructed, and an event triggering mechanism is established based on this function. When the triggering condition is met, a new triggering time is output to update the distributed guidance law. According to the distributed guidance law, each execution robot is driven to move towards the target deployment position to achieve deployment. By constructing a global field density observer that couples an adaptive update law with a projection operator, accurate estimation of the global field density is achieved under the condition of relying only on local measurements and having an unknown diffusion coefficient. Based on the estimated centroid, a distributed guidance law without velocity measurement is designed to drive the execution robot to deploy to the target position. By constructing a global field density observer, accurate estimation of global field density is achieved under the condition of unknown diffusion coefficient. An event-triggered mechanism is introduced to update the distributed guidance law on demand, thereby reducing resource consumption while ensuring robot convergence performance. Ultimately, this enables precise and low-consumption deployment of swarm robots in dynamic diffusion fields.

[0055] Example 3 Figure 7 This is a schematic diagram of the swarm robot deployment device under a dynamic diffusion field provided in Embodiment 3 of the present invention, as shown below. Figure 7 As shown, the device includes: The first partitioning module 310 is used to divide the dynamic diffusion field into multiple non-overlapping sub-regions, and to deploy sensing robots in the sub-regions. The sensing robots are used to collect the local field density of their respective sub-regions. The observer construction module 320 is used to construct a parabolic pollution diffusion field model with an unknown diffusion coefficient, and to construct a global field density observer based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot. The adaptive update law construction module 330 is used to construct a projection operator for the diffusion coefficient and construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient. The acquisition module 340 is used to obtain an estimated value of the diffusion coefficient according to the adaptive update law of the diffusion coefficient, and input the estimated value of the diffusion coefficient into the global field density observer to obtain an estimated value of the global field density. The second partitioning module 350 is used to divide the dynamic diffusion field into several sub-regions based on the real-time position of the execution robot using Voronoi partitioning technology, thereby obtaining the Voronoi cavity corresponding to each execution robot. The centroid estimation module 360 ​​is used to calculate the estimated centroid of the Voronoi cell corresponding to each execution robot based on the estimated value of the global field density, and use it as the target deployment position of the execution robot; The distributed guidance law construction module 370 is used to construct sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and to construct a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables; Drive module 380, according to the distributed guidance law, drives each execution robot to move towards the target deployment position to achieve deployment. The swarm robot deployment device under a dynamic diffusion field provided in this embodiment divides the dynamic diffusion field into multiple non-overlapping sub-regions, deploys sensing robots within these sub-regions, and the sensing robots are used to collect the local field density of their respective sub-regions. A parabolic pollution diffusion field model with an unknown diffusion coefficient is constructed. Based on the parabolic pollution diffusion field model, the local field density, and the real-time positions of the sensing robots, a global field density observer is constructed. A projection operator for the diffusion coefficient is constructed, and an adaptive update law for the diffusion coefficient is constructed based on the projection operator. An estimated value of the diffusion coefficient is obtained according to the adaptive update law, and the estimated value of the diffusion coefficient is input into the global field density observer. The global field density is estimated using a field density observer. Based on the real-time position of the executing robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology, resulting in Voronoi cavities for each executing robot. Based on the estimated global field density, the estimated centroid of the Voronoi cavity for each executing robot is calculated, serving as the target deployment position for that robot. Sliding mode variables and auxiliary intermediate variables are constructed based on the estimated centroids, and a distributed guidance law is built based on these variables. The distributed guidance law drives each executing robot to move towards the target deployment position to achieve deployment. By constructing a global field density observer that couples an adaptive update law with a projection operator, accurate estimation of the global field density is achieved under conditions where only local measurements are relied upon and the diffusion coefficient is unknown. Furthermore, a distributed guidance law without velocity measurement is designed based on the estimated centroids to drive the executing robots to deploy towards the target position.

[0056] The swarm robot deployment device under dynamic diffusion field provided in the embodiments of the present invention can execute the swarm robot deployment method under dynamic diffusion field provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0057] Example 4 Figure 8 This is a schematic diagram of the structure of a server provided in Embodiment 4 of the present invention. Figure 8 A block diagram is shown of an exemplary server 12 suitable for implementing embodiments of the present invention. Figure 8 The server 12 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0058] like Figure 8 As shown, server 12 is presented in the form of a general-purpose computing server. The components of server 12 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0059] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0060] Server 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by server 12, including volatile and non-volatile media, removable and non-removable media.

[0061] System memory 28 may include computer system readable media in the form of volatile memory, such as RAM 30 and / or cache 32. Server 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 8 Not shown; usually referred to as a "hard drive"). Although Figure 8 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0062] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0063] Server 12 can also communicate with one or more external devices 14 (e.g., keyboard, pointing server, display 24, etc.), and with one or more servers that enable users to interact with server 12, and / or with any server (e.g., network card, modem, etc.) that enables server 12 to communicate with one or more other computing servers. This communication can be performed through I / O interface 22. Furthermore, server 12 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. Figure 8 As shown, network adapter 20 communicates with other modules of server 12 via bus 18. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with server 12, including but not limited to: microcode, server drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0064] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, such as implementing the swarm robot deployment method in a dynamic diffusion field provided in the embodiments of the present invention.

[0065] Example 5 Embodiment 5 of the present invention also provides a medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a swarm robot deployment method in a dynamic diffusion field as described in any of the above embodiments.

[0066] The computer medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, 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, apparatus, or device.

[0067] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying 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. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0068] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0069] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0070] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for deploying swarm robots in a dynamic diffusion field, characterized in that, include: The dynamic diffusion field is divided into multiple non-overlapping sub-regions, and sensing robots are deployed in the sub-regions to collect the local field density of their respective sub-regions. A parabolic pollution diffusion field model with an unknown diffusion coefficient is constructed. The formula for the parabolic pollution diffusion field model is as follows: ； in, Indicates time Location The actual field density at that location, This represents the rate of change of the actual field density over time. For the Lass-Rapp operator, The coefficient of the reaction term, The diffusion coefficient is unknown. The real-time position of the sensing robot is obtained by using the robot's own sensors; A global field density observer is constructed based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot. The formula for the global field density observer is as follows: ； in, This represents the estimated global field density at time t. The rate of change of the global field density estimate over time. This is an estimate of the unknown diffusion coefficient. The coefficient of the reaction term, It is a positive gain parameter. It is a spatial distribution function. The real-time position of the sensor robot is measured at time t. Let be the local field density at time t. For feedback correction items, The number of sub-regions into which the dynamic diffusion field region Ω is divided along the x-axis. To dynamically diffuse the field area The number of sub-regions divided along the y-direction. This is the number of the sub-region in the x-direction. This is the number of the sub-region along the y-axis. For dynamic diffusion field Any point in; Construct a projection operator for the diffusion coefficient, and construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient; The estimated value of the diffusion coefficient is obtained according to the adaptive update law of the diffusion coefficient, and the estimated value of the diffusion coefficient is input into the global field density observer to obtain the estimated value of the global field density. Based on the real-time position of the execution robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology to obtain the Voronoi cavity corresponding to each execution robot. Based on the estimated value of the global field density, the estimated centroid of the Voronoi cell corresponding to each execution robot is calculated, which is used as the target deployment position of the execution robot; Based on the estimated centroid, sliding mode variables and auxiliary intermediate variables are constructed, and a distributed guidance law is constructed based on the sliding mode variables and auxiliary intermediate variables. According to the distributed guidance law, each execution robot is driven to move towards the target deployment location to achieve deployment.

2. The method for deploying swarm robots in a dynamic diffusion field according to claim 1, characterized in that, The projection operator for constructing the diffusion coefficient, and the adaptive update law for constructing the diffusion coefficient based on the projection operator for the diffusion coefficient, include: Define a compact set of diffusion coefficients, which is expressed as: ； in, Given a known lower bound, Given an upper bound, Given the projection constraint range, This is an estimate of the unknown diffusion coefficient; The projection operator for the diffusion coefficient is constructed based on the compact set, and the formula for the projection operator for the diffusion coefficient is as follows: ； ； in, Indicates the upper bound compression correction amount and This represents the lower bound compression correction amount. As an error-driven term, Let be the partial derivative of the global field density estimate in the x-direction. Let be the partial derivative of the global field density estimate in the y-direction. It is a positive gain coefficient; An adaptive update law for the diffusion coefficient is constructed based on the projection operator of the diffusion coefficient. The formula for the adaptive update law of the diffusion coefficient is as follows: ； in, For the adaptive update law of the diffusion coefficient, The projection operator is the diffusion coefficient.

3. The method for deploying swarm robots in a dynamic diffusion field according to claim 2, characterized in that, The dynamic diffusion field is divided into several sub-regions based on the real-time position of the executing robot using Voronoi partitioning technology, resulting in a Voronoi cavity corresponding to each executing robot, including: The robot's real-time position is obtained using its own sensors. Based on the real-time position of the executing robot, the dynamic diffusion field is divided into several sub-regions using Voronoi partitioning technology, resulting in a Voronoi cavity corresponding to each executing robot. The Voronoi cavity corresponding to each executing robot is represented as follows: ； in, and To execute the robot number, and For the number The real-time location and number of the execution robot are as follows: The real-time position of the execution robot, where q is the dynamic diffusion field. Any point in the array.

4. The method for deploying swarm robots in a dynamic diffusion field according to claim 3, characterized in that, The calculation of the estimated centroid of the Voronoi cavity for each executing robot, based on the estimated global field density, includes: Based on the estimated global field density, the estimated centroid of the Voronoi cell corresponding to each executing robot is calculated. The centroid calculation formula is as follows: ； in, Let be the estimated mass of the Voronoi cavity corresponding to the i-th executing robot. Let be the estimated centroid of the Voronoi cell corresponding to the i-th executing robot.

5. The method for deploying swarm robots in a dynamic diffusion field according to claim 4, characterized in that, The step of constructing sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and constructing a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables, includes: Based on the estimated centroid, sliding mode variables and auxiliary intermediate variables are constructed. The formula for the sliding mode variables is as follows: ； in, This represents the approach gain coefficient. Indicates the damping gain coefficient. Represents the robustness gain coefficient. Let be the sliding mode variable of the i-th robot at time t. This represents the real-time position of the i-th executing robot at time t. Let be the estimated centroid of the i-th Voronoi cell at time t. Let be the rate of change of the sliding mode variable of the i-th robot over time; The formula for the auxiliary intermediate variable is as follows: ； in, This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. Let be the auxiliary intermediate variable for the i-th robot at time t. Let t be the real-time position of the i-th executing robot. This represents the rate of change of the auxiliary intermediate variable of the i-th robot over time. Based on the sliding mode variable and the auxiliary intermediate variable, a distributed guidance law is constructed, and the formula for the distributed guidance law is as follows: ； in, This represents the sliding mode feedback gain coefficient. This represents the total gain coefficient of the auxiliary channel. This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. For a moment The sliding mode variable of the i-th executing robot, For a moment The auxiliary intermediate variable for the i-th executing robot, For a moment The real-time position of the i-th executing robot. Let be the guidance law for the i-th executing robot; The step of driving each execution robot to move towards the target deployment location according to the distributed guidance law to achieve deployment includes: According to the distributed guidance law, each execution robot is driven to move towards the target deployment position based on the dual integrator dynamics model of each execution robot to achieve deployment. The formula of the dual integrator dynamics model is as follows: ； in, Let be the derivative of the position coordinates of the i-th executing robot. Let be the speed of the i-th executing robot. Let be the derivative of the velocity of the i-th executing robot. Let be the guidance law for the i-th executing robot.

6. The method for deploying swarm robots in a dynamic diffusion field according to claim 1, characterized in that, After constructing sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and constructing the distributed guidance law based on the sliding mode variables and auxiliary intermediate variables, the method further includes: Construct an event-triggered error function, the formula of which is as follows: ； in, This represents the sliding mode feedback gain coefficient. This represents the total gain coefficient of the auxiliary channel. This represents the attenuation gain coefficient of the auxiliary variable. This represents the position feedback gain coefficient. Let be the sliding mode variable of the i-th executing robot at time t. Let be the auxiliary intermediate variable for the i-th executing robot at time t. Let t be the real-time position of the i-th executing robot. To define the guidance law for the i-th executing robot at time t, given the introduction of an event-triggered error function. , For a moment The sliding mode variable of the i-th executing robot, For a moment The auxiliary intermediate variable for the i-th executing robot, For a moment The real-time position of the i-th executing robot; An event triggering mechanism is established based on the event triggering error function. When the triggering condition is met, a new triggering time is output to update the distributed guidance law. The formula for the event triggering mechanism is as follows: ； in, For the (k+1)th event triggered by the i-th robot, Let k be the time when the event is triggered by the i-th robot. For time variables, The time-triggered threshold at time t. , This represents the rate of change of the dynamic trigger threshold of the i-th robot over time. For threshold coefficient, This is the error feedback coefficient.

7. A swarm robot deployment device in a dynamic diffusion field, characterized in that, include: The first partitioning module is used to divide the dynamic diffusion field into multiple non-overlapping sub-regions, and to deploy sensing robots in the sub-regions. The sensing robots are used to collect the local field density of their respective sub-regions. The observer construction module is used to construct a parabolic pollution diffusion field model with an unknown diffusion coefficient. The formula for the parabolic pollution diffusion field model is as follows: ;in, Indicates time Location The actual field density at that location, This represents the rate of change of the actual field density over time. For the Lass-Rapp operator, The coefficient of the reaction term, The diffusion coefficient is unknown; the real-time position of the sensing robot is obtained using its own sensors; a global field density observer is constructed based on the parabolic pollution diffusion field model, the local field density, and the real-time position of the sensing robot. The formula for the global field density observer is as follows: ;in, This represents the estimated global field density at time t. The rate of change of the global field density estimate over time. This is an estimate of the unknown diffusion coefficient. The coefficient of the reaction term, It is a positive gain parameter. It is a spatial distribution function. The real-time position of the sensor robot is measured at time t. Let be the local field density at time t. For feedback correction items, The number of sub-regions into which the dynamic diffusion field region Ω is divided along the x-axis. To dynamically diffuse the field area The number of sub-regions divided along the y-direction. This is the number of the sub-region in the x-direction. This is the number of the sub-region along the y-axis. For dynamic diffusion field Any point in; An adaptive update law construction module is used to construct a projection operator for the diffusion coefficient, and to construct an adaptive update law for the diffusion coefficient based on the projection operator for the diffusion coefficient. The acquisition module is used to obtain an estimated value of the diffusion coefficient according to the adaptive update law of the diffusion coefficient, and input the estimated value of the diffusion coefficient into the global field density observer to obtain an estimated value of the global field density. The second partitioning module is used to divide the dynamic diffusion field into several sub-regions based on the real-time position of the execution robot using Voronoi partitioning technology, thereby obtaining the Voronoi cell cavity corresponding to each execution robot. The centroid estimation module is used to calculate the estimated centroid of the Voronoi cell corresponding to each execution robot based on the estimated value of the global field density, and use it as the target deployment position of the execution robot; A distributed guidance law construction module is used to construct sliding mode variables and auxiliary intermediate variables based on the estimated centroid, and to construct a distributed guidance law based on the sliding mode variables and auxiliary intermediate variables; The drive module, according to the distributed guidance law, drives each execution robot to move towards the target deployment location to achieve deployment.

8. A server, characterized in that, The server includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the swarm robot deployment method in a dynamic diffusion field as described in any one of claims 1-6.

9. A medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the swarm robot deployment method in a dynamic diffusion field as described in any one of claims 1-6.

Images