A situation fusion and task allocation method for distributed unmanned aerial vehicles

Abstract

The application discloses a distributed unmanned aerial vehicle situation fusion and task allocation method, which is used for enhancing the cooperative operation ability and task execution efficiency of the unmanned aerial vehicle group. The unmanned aerial vehicle collects environmental data through an airborne sensor, adopts a Gaussian filter for denoising processing, and extracts feature information. By using a distributed situation fusion algorithm, the unmanned aerial vehicle group synchronizes and fuses information to form a global situation map. Through a Bayesian network, optimal task allocation is carried out to realize the optimized utilization of resources. The adaptability and decision-making ability of the unmanned aerial vehicle group in a complex environment can be effectively improved, and the method is suitable for military and civilian fields.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) technology, specifically relating to a method for situational awareness fusion and task allocation for distributed UAVs. Background Technology

[0002] With the continuous advancement of technology, unmanned aerial vehicle (UAV) technology has been widely applied in various fields such as military reconnaissance, surveillance, cargo transportation, environmental monitoring, and agricultural spraying. Especially in the military field, the collaborative combat capability of UAV swarms has become a crucial force in modern warfare. However, existing UAV systems typically face problems such as insufficient information sharing, low collaborative efficiency, and limited situational awareness. In complex battlefield environments, UAVs need to acquire and process large amounts of data accurately and in real time to achieve effective situational awareness and mission execution.

[0003] Most existing unmanned aerial vehicle (UAV) systems employ centralized control, which not only limits system scalability but also increases reliance on a central control unit. If the control unit is damaged, the entire system may be paralyzed. Furthermore, the perception range and processing capabilities of a single UAV are limited, making it difficult to achieve comprehensive awareness of a large-scale battlefield environment. To overcome these limitations, a distributed UAV situational awareness fusion approach is needed, enabling information sharing and collaborative operations among UAVs to improve the overall combat effectiveness of UAV swarms.

[0004] In distributed unmanned aerial vehicle (UAV) systems, each UAV acts as an independent node, autonomously collecting environmental information and sharing this information with other UAVs. Through situational awareness fusion, the UAV swarm can construct a global battlefield situational map, supporting decision-making. However, existing situational awareness fusion methods often rely on complex data processing algorithms and substantial computing resources, making them difficult to implement on resource-constrained UAV platforms. Furthermore, how to maintain UAV swarm coordination while achieving dynamic task allocation and optimization remains a pressing issue. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a distributed UAV situational awareness fusion and task allocation method to enhance the collaborative operation capability and task execution efficiency of UAV swarms. UAVs collect environmental data through onboard sensors, perform noise reduction using Gaussian filters, and extract feature information. A distributed situational awareness fusion algorithm is used to synchronize and fuse information within the UAV swarm, forming a global situational awareness map. Optimal task allocation is achieved through a Bayesian network, optimizing resource utilization. This method effectively improves the adaptability and decision-making capabilities of UAV swarms in complex environments and is applicable to both military and civilian fields.

[0006] The technical solution adopted by this invention to solve its technical problem is as follows:

[0007] Step 1: UAV situational awareness;

[0008] Step 1-1: The UAV uses its onboard sensors to collect surrounding environmental data I. The environmental data I is preprocessed using signal processing operations. The preprocessed environmental data information is I. f :

[0009] I f =G*I+λI (1)

[0010] Among them: I f G represents the filtered environmental data, where G is the Gaussian kernel, I is the raw environmental data collected by the UAV's onboard sensors, and λ is the smoothing parameter.

[0011] Step 1-2: Assume the UAV senses the filtered environmental data information I. f (x, y), where x and y are the horizontal and vertical coordinates of the environmental information perceived by the UAV, respectively. I is obtained through calculation. f The gradient D of (x,y) in the horizontal and vertical directions x With D y :

[0012]

[0013] in: To determine the sign of the partial derivative;

[0014] Calculate I f The product of the gradients in the (x, y) directions, i.e., the product along the X-axis. Product in the Y-axis direction

[0015]

[0016] Use Gaussian function and D x D y Gaussian weighting is performed to generate intermediate computational matrix elements M. A M B and M C :

[0017]

[0018] Where: ω x ω y and ω xy These are the Gaussian weighted values, The cross product symbol;

[0019] Steps 1-3: Calculate the data I sensed by the drone. fHarris response value R per pixel map :

[0020] R map ={det(M A M B M C )-α(trace(M A M B M C )) 2 <T r} (6)

[0021] Where: det(M A M B M C ) represents the determinant of the matrix, trace(M) A M B M C The locus is the matrix determinant, α represents the corner response parameter, and T... r Indicates the threshold for judgment;

[0022] This yields the feature information T of the data sensed by the UAV. For any sensed pixel (x,y)∈T, the pixel always satisfies the Harris response value R. map ,Right now:

[0023]

[0024] Step 2: Cloud-based situational awareness integration;

[0025] Step 2-1: For feature information T i and T j Construct feature windows Win=(w x ,h y Then, the two-dimensional windows in the feature window are converted into feature vectors n according to the arrangement of a one-dimensional matrix, i.e., Win x,y =n 1,xy Construct feature window vectors

[0026] For eigenvectors The similarity constraint problem involves minimizing the similarity of categories to identify target categories with similar features, and then performing similarity matching, as shown in the following formula:

[0027]

[0028] Let matrix S V The singular values ​​are:

[0029] S V =PΣP T(9)

[0030] Where: P is the transformation matrix, and Σ is the N×N singular value matrix;

[0031] Step 2-2: Let Σ K Let matrix S V The diagonal matrix formed by the first K singular values, P ·K For Σ K The corresponding left singular value vector then has A definite solution is:

[0032]

[0033] The original objective function problem can be rewritten as:

[0034]

[0035] Steps 2-3: Place U K ={u1,u2,u3,...,u N H is the input layer of the network, and H is the output layer. The objective function of the problem is solved using a deep belief network model. The input layer u of each layer is used as the network input layer. i ∈U K As a visible variable, h j Treating ∈H as a hidden variable, we obtain the definition of the energy function:

[0036]

[0037] Where θ = {T, d, c} are model parameters, and D and M represent the number of visible and hidden units in the network, respectively; W ij d represents the feature window value based on parameters i and j. i c represents the corresponding weight of the visible variable. j This represents the corresponding weight of the hidden variable;

[0038] Step 2-4: Define the joint distribution of {u,h} as:

[0039]

[0040] Where Z(θ)=∑ u ∑ h exp(-E(u,h;θ)) is the allocation function that guarantees probability normalization;

[0041] By defining the range of values ​​for the model similarity constraint parameter θ, the similarity of data feature value classification can be further altered, thus achieving the desired similarity for different feature information T. i and T j The matching and fusion problem.

[0042] Step 3: Optimal task allocation;

[0043] Step 3-1: For a multi-UAV target assignment task, its Bayesian network representation is as follows:

[0044] B =<G,P> (14)

[0045] Where: G =<U,S,A> It is a directed acyclic graph, U = {u1, u2, u3, ..., u} N} represents the drone crew members participating in the mission, S = {S T1 ,S T2 ,...,S Tm To identify the true target to be attacked, A is the set of arcs in graph G, and P is the probabilistic annotation of graph G; for any drone member u performing the mission k In P, each element represents the conditional probability density of the target node, which, according to the probability density rule, is:

[0046]

[0047] Step 3-2: For any target task node S in the Bayesian network Tm It is possible to find a match with S Tm The smallest subset where none of the conditions are independent Make:

[0048] P(S Tm |S T1 ,S T2 ,...,S Tm-1 )=P(S Tm |S u (16)

[0049] Wherein: S u For node S Tm In graph G =<U,S,A> The set of parent nodes in;

[0050] This allows for the unique determination of task node S. Tm Assigned to drone u k probability distribution:

[0051]

[0052] For drone member u l have:

[0053]

[0054] This allows for the fusion of UAV situational awareness and task allocation methods, enabling UAVs to fly safely in the environment while distributively collecting environmental information and fusing all the information. Based on this, the perceived task objectives are uniformly allocated to achieve global optimization.

[0055] Preferably, the preprocessing is denoising using a Gaussian filter.

[0056] Preferably, the Gaussian kernel G is

[0057] A computer program that causes a computer to perform the above-described situation fusion and task allocation method.

[0058] An electronic device includes: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to enable the electronic device to perform the above-described situational fusion and task allocation method.

[0059] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described situation fusion and task allocation method.

[0060] A chip includes a processor for retrieving and running a computer program from a memory, causing a device equipped with the chip to perform the aforementioned situational fusion and task allocation method.

[0061] A computer program product includes a computer storage medium storing a computer program, the computer program including instructions executable by at least one processor, which, when executed by the at least one processor, implement the aforementioned situational fusion and task allocation method.

[0062] The beneficial effects of this invention are as follows:

[0063] This invention not only improves the situational awareness and information sharing capabilities of UAV swarms, but also enhances their adaptability and decision-making capabilities in complex environments through intelligent task allocation strategies, thus possessing significant military and civilian application value. Attached Figure Description

[0064] Figure 1 This is a flowchart of the steps of the present invention.

[0065] Figure 2 This is a schematic diagram of the situation fusion method of the present invention.

[0066] Figure 3This is a schematic diagram illustrating the classification of different features according to an embodiment of the present invention. Objects of different shapes are placed on an open ground, and a drone is used to take aerial photos of the test area and conduct the test.

[0067] Figure 4 This is a distributed fusion example of an embodiment of the present invention. The left side shows the local information perceived by the distributed UAV, and the right side shows the overall situational information after information fusion.

[0068] Figure 5 A flowchart illustrating the optimal allocation of drone missions. Detailed Implementation

[0069] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0070] like Figure 1 and Figure 2 As shown, this invention provides a method for situational awareness fusion and task allocation of distributed unmanned aerial vehicles (UAVs). This method achieves efficient collaborative operation of UAV swarms through three steps: UAV situational awareness, cloud-based situational awareness fusion, and optimal task allocation.

[0071] In the UAV situational awareness step, each UAV uses its onboard sensors to collect environmental data and preprocesses it to reduce noise and interference. A Gaussian filter is used to denoise the raw data, and feature information is extracted by calculating gradients and Harris response values, thereby obtaining the feature information of the data perceived by the UAV.

[0072] In the cloud-based situational awareness fusion step, the UAV swarm uses a distributed situational awareness fusion algorithm to synchronize and fuse the local information perceived by each UAV to form a global situational awareness map. This algorithm converts feature information into feature vectors and uses a deep belief network model for similarity matching and fusion, thereby achieving accurate fusion of information perceived by different UAVs.

[0073] In the optimal task allocation step, based on the global situation map, the task allocation module of the UAV swarm can dynamically allocate tasks to each UAV to optimize resource utilization and improve task execution efficiency. A Bayesian network is introduced for task allocation modeling, and prediction and causal analysis are performed using a directed acyclic graph with probabilistic annotations to achieve dynamic adjustment of the real-time task allocation strategy.

[0074] The distributed UAV situational awareness and task allocation method of the present invention not only improves the situational awareness and information sharing capabilities of UAV swarms, but also enhances the adaptability and decision-making capabilities of UAV swarms in complex environments through intelligent task allocation strategies, and has important military and civilian application value.

[0075] Example:

[0076] The situation fusion and task allocation method for distributed UAVs mainly includes three steps: UAV situation awareness, cloud situation fusion, and optimal task allocation.

[0077] Step 1: UAV situational awareness.

[0078] Each UAV is equipped with a situational awareness module to collect its own status information and information about the surrounding environment. Assuming the UAV uses onboard sensors (typically an onboard visible light payload) to collect environmental data I, this data undergoes preprocessing through signal processing to reduce noise and interference. Here, a Gaussian filter is used for noise reduction. The processed environmental data is I. f :

[0079] I f =G*I+λI (19)

[0080] Among them: I f The filtered environmental data information, where G is the Gaussian kernel, typically taken as... I represents the raw environmental data collected by the UAV's onboard sensors, and λ is the smoothing parameter.

[0081] After obtaining I f Subsequently, in order to reduce the amount of data transmitted and transmit more effective information within the carrying capacity of the effective communication network, it is necessary to process the data sensed by the UAV. f Feature extraction is performed.

[0082] Assume the drone senses the filtered environmental data I. f Let (x, y) be the abscissa and ordinate of the environmental information perceived by the UAV, respectively. Then, I can be calculated. f The gradients Dx and Dy of (x,y) in the horizontal and vertical directions:

[0083]

[0084] in: To determine the sign of the partial derivative.

[0085] Calculate I f The product of the gradients in the (x, y) directions, i.e., the product along the X-axis. Product in the Y-axis direction

[0086]

[0087] Use Gaussian function and D x D yGaussian weighting can generate intermediate computational matrix elements M. A M B and M C :

[0088]

[0089] Where: ω x ω y and ω xy These are the Gaussian weighted values, This is the cross product symbol.

[0090] Calculate the data sensed by the drone I f Harris response value R per pixel map :

[0091] R map ={det(M A M B M C )-α(trace(M A M B M C )) 2 <T r} (twenty four)

[0092] Where: det(M A M B M C ) represents the determinant of the matrix, and trace(M) A M B M C The locus is the matrix determinant, α represents the corner response parameter, and T... r This represents the threshold for judgment. If the value is greater than this threshold, the point is considered a feature point.

[0093] From this, we can obtain the feature information T of the data sensed by the UAV. For any sensed pixel (x,y)∈T, we can obtain that the pixel always satisfies the Harris response value R. map ,Right now:

[0094]

[0095] Step 2: Cloud-based situational awareness integration.

[0096] The drone swarm employs a distributed situational awareness fusion algorithm, which synchronizes and fuses situational information among drones to form a global situational awareness map. This algorithm considers two feature information T... i and T j The matching and fusion problem can be transformed into, for feature information T i and T jwhere the feature window Win = (w x , h y ). The two-dimensional window in the feature window is converted into a feature vector n according to the arrangement of a one-dimensional matrix, that is, Win x,y = n 1,xy . Construct a feature window vector

[0097] Regarding the similarity constraint problem of the feature vector , constraining the similarity of the same category to the minimum value is the target category with similar features, and similarity matching can be performed, as shown in Equation (12):

[0098]

[0099] Since Equation (8) is an optimization problem of a matrix, to find its optimal solution, it is first necessary to transform it into the form of a singular value matrix. Let the singular values of matrix S V be respectively:[[]]

[0100] S V = PΣP T (27)[[]]

[0101] where: P is a transformation matrix, and Σ is a singular value matrix of N×N.[[]]

[0102] Assume that Σ K is a diagonal matrix composed of the first K singular values of matrix S V , and P ·K is the left singular value vector corresponding to Σ K , then there is K a definite solution of is:[[]]

[0103]

[0104] For any orthogonal matrix Τ, it is easy to verify that Ι K = U K <T is still a solution to the problem. Therefore, the problem of the original objective function can be rewritten as:[[]]

[0105]

[0106] Taking U K = {u1, u2, u3,..., u N} as the input layer of the network and H as the output layer of the network, then the objective function of this problem can be used to solve the deep belief network model. Similar to the energy function of the deep belief network, if the network input layer u i ∈U K of each layer is used as the visible variable, while h iIf ∈H is considered a hidden variable, then the definition of the energy function can be obtained:

[0107]

[0108] Where θ = {T, d, c} are model parameters, and D and M represent the number of visible and hidden units in the network, respectively. The joint distribution of {u, h} is defined as:

[0109]

[0110] Where Z(θ)=∑ u ∑ h exp(-E(u,h;θ)) is the allocation function that guarantees probability normalization. By defining the range of values ​​for the model similarity constraint parameter θ, the similarity of data feature values ​​can be changed. This achieves the desired similarity for different feature information T. i and T j The matching and fusion problem.

[0111] By fusing the individual environmental information sensed by multiple UAVs distributed around the battlefield, an overall information on the battlefield situation can be formed. Based on this, collaborative planning of multiple UAVs can be carried out, thereby better executing missions and ensuring the flight safety of UAVs.

[0112] Step 3: Optimal task allocation.

[0113] Based on the global situational awareness map, the task allocation module of the UAV swarm can dynamically assign tasks to each UAV to optimize resource utilization and improve task execution efficiency. By introducing Bayesian networks into UAV target allocation task modeling, the dynamic adjustment of real-time allocation strategies in task allocation can be addressed to select the optimal allocation.

[0114] A Bayesian network is a directed acyclic graph with probabilistic annotations, which can be used to reveal learning and statistical inference functions for prediction, causal analysis, etc. For a multi-UAV target assignment task, its Bayesian network can be represented as:

[0115] B =<G,P> (32)

[0116] Where: G =<U,S,A> It is a directed acyclic graph, U = {u1, u2, u3, ..., u} N} represents the drone crew members participating in the mission, S = {S T1 ,S T2 ,...,S Tm To identify the true target to be attacked, let A be the set of arcs in graph G, and P be the probabilistic annotation of graph G. For any drone crew member u performing the mission... kIn P, each element represents the conditional probability density of the target node, which, according to the probability density rule, is:

[0117]

[0118] Obviously, calculating the probability P(S) requires giving 2 m-1 The computational complexity of calculating multiple probability values ​​is enormous. Therefore, introducing the variable independence assumption into Bayesian networks significantly reduces the amount of prior probability definition required. For the constructed probability density rule, for any target task node S in the network structure... Tm It is certain that a match with S can be found. Tm The smallest subset where none of the conditions are independent Make:

[0119] P(S Tm |S T1 ,S T2 ,...,S Tm-1 )=P(S Tm |S u (34)

[0120] Wherein: S u For node S Tm In graph G =<U,S,A> The set of parent nodes in the [database]. This uniquely identifies the task node S. Tm Assigned to drone u k probability distribution:

[0121]

[0122] The same principle applies to the drones that need to be assigned to attack other real targets, such as drone crew members u. l have:

[0123]

[0124] This allows for the fusion of situational awareness and task allocation methods for UAVs, enabling them to fly safely in the environment while distributively collecting environmental information and fusing all the information to form a holistic understanding of the environment. Based on this understanding, the perceived task objectives can be uniformly allocated to achieve global optimization.

[0125] Figure 3 This demonstrates feature classification for different characteristics. Objects of different shapes (squares, triangles, and circles) were placed on an open area, and a drone was used to photograph the test area from above. The results show that regardless of how the objects are placed, the system can accurately identify and classify the object types after translation, rotation, and scaling.

[0126] Figure 4 This example illustrates distributed fusion. The left side shows localized information perceived by the distributed drones. It can be seen that each drone senses a portion of the enemy air defense identification zone (ADIZ) in the environment, providing data support for subsequent information fusion among multiple drones. The right side shows the overall situational awareness after information fusion. The onboard sensors of the distributed drones can identify three enemy ADIZs in the environment (four were actually set in the environment). One ADIZ was not perceived by any drone and therefore could not be represented in the overall situational awareness after fusion. This aligns with actual combat results; the overall battlefield situational awareness constructed in an unknown environment can only be a reproduction of the already sensed information.

[0127] Figure 5 The flowchart shows the optimal allocation of drone missions.

Claims

1. A method for situation fusion and task allocation for distributed unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: Step 1: UAV situational awareness; Step 1-1: The UAV uses its onboard sensors to collect surrounding environmental data I. The environmental data I is preprocessed using signal processing operations. The preprocessed environmental data information is I. f : I f =G*I+λI (1) Among them: I f G represents the filtered environmental data, where G is the Gaussian kernel, I is the raw environmental data collected by the UAV's onboard sensors, and λ is the smoothing parameter. Step 1-2: Assume the UAV senses the filtered environmental data information I. f (x, y), where x and y are the horizontal and vertical coordinates of the environmental information perceived by the UAV, respectively. I is obtained through calculation. f The gradient D of (x,y) in the horizontal and vertical directions x With D y : in: To determine the sign of the partial derivative; Calculate I f The product of the gradients in the (x, y) directions, i.e., the product along the X-axis. Product in the Y-axis direction Use Gaussian function and D x D y Gaussian weighting is performed to generate intermediate computational matrix elements M. A M B and M C : Where: ω x ω y and ω xy These are the Gaussian weighted values, The cross product symbol; Steps 1-3: Calculate the data I sensed by the drone. f Harris response value R per pixel map : R map ={det(M A ,M B ,M C )-α(trace(M A ,M B ,M C )) 2 <T r } (6) Where: det(M A M B M C ) represents the determinant of the matrix, trace(M) A M B M C The locus is the matrix determinant, α represents the corner response parameter, and T... r Indicates the threshold for judgment; This yields the feature information T of the data sensed by the UAV. For any sensed pixel (x,y)∈T, the pixel always satisfies the Harris response value R. map ,Right now: Step 2: Cloud-based situational awareness integration; Step 2-1: For feature information T i and T j Construct feature windows Win=(w x ,h y Then, the two-dimensional windows in the feature window are converted into feature vectors n according to the arrangement of a one-dimensional matrix, i.e., Win x,y =n 1,xy Construct feature window vectors For eigenvectors The similarity constraint problem involves minimizing the similarity of categories to identify target categories with similar features, and then performing similarity matching, as shown in the following formula: Let matrix S V The singular values ​​are: S V =PΣP T (9) Where: P is the transformation matrix, and Σ is the N×N singular value matrix; Step 2-2: Let Σ K Let S be a matrix V The diagonal matrix formed by the first K singular values, P ·K For Σ K The corresponding left singular value vector is then: A definite solution is: The original objective function problem can be rewritten as: Steps 2-3: Place U K ={u1,u2,u3,...,u N H is the input layer of the network, and H is the output layer. The objective function of the problem is solved using a deep belief network model. The input layer u of each layer is used as the network input layer. i ∈U K As a visible variable, h j Treating ∈H as a hidden variable, we obtain the definition of the energy function: Where θ = {T, d, c} are model parameters, and D and M represent the number of visible and hidden units in the network, respectively; W ij d represents the feature window value based on parameters i and j. i c represents the corresponding weight of the visible variable. j This represents the corresponding weight of the hidden variable; Step 2-4: Define the joint distribution of {u,h} as: Where Z(θ)=∑ u ∑ h exp(-E(u,h;θ)) is the allocation function that guarantees probability normalization; By defining the range of values ​​for the model similarity constraint parameter θ, the similarity of data feature value classification can be further altered, thus achieving the desired similarity for different feature information T. i and T j The matching and fusion problem; Step 3: Optimal task allocation; Step 3-1: For a multi-UAV target assignment task, its Bayesian network representation is as follows: B =<G,P> (14) Where: G =<U,S,A> It is a directed acyclic graph, U = {u1, u2, u3, ..., u} N } represents the drone crew members participating in the mission, S = {S T1 ,S T2 ,...,S Tm To identify the true target to be attacked, A is the set of arcs in graph G, and P is the probabilistic annotation of graph G; for any drone member u performing the mission k In P, each element represents the conditional probability density of the target node, which, according to the probability density rule, is: Step 3-2: For any target task node S in the Bayesian network Tm It is possible to find a match with S Tm The smallest subset where none of the conditions are independent Make: P(S Tm |S T1 ,S T2 ,...,S Tm-1 )=P(S Tm |S u ) (16) Wherein: S u For node S Tm In graph G =<U,S,A> The set of parent nodes in; This allows for the unique determination of task node S. Tm Assigned to drone u k probability distribution: For drone member u l have: This allows for the fusion of UAV situational awareness and task allocation methods, enabling UAVs to fly safely in the environment while distributively collecting environmental information and fusing all the information. Based on this, the perceived task objectives are uniformly allocated to achieve global optimization.

2. The situation fusion and task allocation method for a distributed unmanned aerial vehicle (UAV) according to claim 1, characterized in that, The preprocessing involves using a Gaussian filter for noise reduction.

3. The situation fusion and task allocation method for a distributed unmanned aerial vehicle (UAV) according to claim 1, characterized in that, The Gaussian kernel G is 4. An electronic device, characterized in that, include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to cause the electronic device to perform the method as described in any one of claims 1 to 3.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 3.

6. A chip, characterized in that, include: A processor for retrieving and running a computer program from memory, causing a device on which the chip is mounted to perform the method as described in any one of claims 1 to 3.

7. A computer program product, characterized in that, The computer program product includes a computer storage medium storing a computer program, the computer program including instructions executable by at least one processor, which, when executed by the at least one processor, implement the method as described in any one of claims 1 to 3.

Images