Equipment agent perception behavior analysis method based on D-S evidence theory

Abstract

According to the equipment agent perception behavior analysis method based on the D-S evidence theory provided by the invention, the influence of various factors on the equipment agent perception behavior is analyzed, and a merging method based on the D-S evidence theory is used for reasoning, so that the analysis result of the equipment agent perception behavior is more accurate and more suitable for practical application.

Description

Technical Field

[0001] The present invention mainly relates to the field of simulation technology, and particularly relates to a method for analyzing the perception behavior of equipment agents based on the D-S evidence theory. Background Art

[0002] An agent can be used as a unified model to describe machine intelligence, animal intelligence, and human intelligence, emphasizing the role of rationality. An agent is an entity that perceives its environment through sensors and acts on that environment through actuators, and can be defined as a mapping from a perception sequence to an entity action. Researchers at home and abroad mainly focus on the research of the perception ability, decision-making ability, control ability, etc. of agents. In terms of the perception ability, researchers are committed to improving the agent's ability to obtain, process, and identify environmental information; in terms of the decision-making ability, researchers focus on the agent's autonomous decision-making and planning ability in complex environments; in terms of the control ability, researchers are committed to achieving the precise control and stable operation of agents.

[0003] Agent-based modeling is a bottom-up modeling method that takes agents as the basic abstract units of the system. Generally, the system model of agents can be described from the following three levels: the agent layer, the individual agent layer, and the MAS (Multi-Agent Systems) layer. For agent-based behavior modeling, the main work focuses on the individual agent layer, that is, determining the structure of the agent and establishing the behavior model of the agent. However, the existing technology for modeling and analyzing the perception behavior of agents is relatively intuitive and simple, the analysis results are inaccurate, and it is difficult to apply to actual scenarios. Summary of the Invention

[0004] Based on this, the present application provides a method for analyzing the perception behavior of equipment agents based on the D-S evidence theory. This method analyzes the influence of various factors on the perception behavior of equipment agents and uses a merging method based on the D-S evidence theory for reasoning, making the analysis results of the perception behavior of equipment agents more accurate and more suitable for actual applications.

[0005] A method for analyzing the perception behavior of equipment agents based on the D-S evidence theory, the method comprising:

[0006] Obtaining at least one factor affecting the perception behavior of the equipment;

[0007] Calculating the observation credibility corresponding to at least one attribute of the observation target according to the at least one factor, where the observation credibility is used to represent the observation result of the equipment agent on an attribute of the observation target;

[0008] Based on the D-S evidence theory, the observation credibility corresponding to at least one observation attribute of the observation target is merged to obtain the target credibility, which is used to represent the perception result of the equipment agent on the observation target.

[0009] Among them, the at least one factor includes:

[0010] The distance d between the equipment agent and the target, the performance q of the detection equipment used by the equipment agent, the observation skill s of the equipment agent, and the knowledge level k of the equipment agent;

[0011] Among them, the value of the distance d is between 0 and 1;

[0012] The value of the performance q is between 0 and 1;

[0013] The value of the observation skill s is between 0 and 1;

[0014] The value of the knowledge level k is between 0 and 1.

[0015] Among them, the calculation method of the distance d is as follows:

[0016]

[0017] Among them, d(1) represents the distance between the equipment agent and the observation target when d = 1; d(0) represents the distance between the equipment agent and the observation target when d = 0; d(a) represents the distance between the equipment agent and the observation target when d = a, and a takes values between 0 and 1;

[0018] The calculation method of the performance q is as follows:

[0019]

[0020] Among them, q(1) represents the performance of the detection equipment used by the equipment agent when q = 1; q(0) represents the performance of the detection equipment used by the equipment agent when q = 0; q(a) represents the performance of the detection equipment used by the equipment agent when q = a, and a takes values between 0 and 1;

[0021] The calculation method of the observation skill s is as follows:

[0022]

[0023] Among them, s(1) represents the observation skill of the equipment agent when s = 1; s(0) represents the observation skill of the equipment agent when s = 0; s(a) represents the observation skill of the equipment agent when s = a, and a takes values between 0 and 1;

[0024] The calculation method of the knowledge level k is as follows:

[0025]

[0026] Among them, k(1) represents the knowledge level of the equipped agent when k = 1; k(0) represents the knowledge level of the equipped agent when k = 0; k(a) represents the knowledge level of the equipped agent when k = a, and the value of a ranges between 0 and 1.

[0027] Among them, calculating the observation credibility corresponding to at least one attribute of the observation target according to the at least one factor includes:

[0028] Performing weighted summation according to at least one external factor to obtain an external factor weighted sum;

[0029] Performing weighted summation according to at least one internal factor to obtain an internal factor weighted sum;

[0030] Performing weighted multiplication according to the external factor weighted sum and the internal factor weighted sum to obtain the observation credibility.

[0031] Among them, performing weighted summation according to at least one external factor to obtain an external factor weighted sum includes:

[0032]

[0033] Among them, x1 represents the external factor weighted sum, x 1i represents one of the external factors, and the weight ω 1i satisfies

[0034] Among them, performing weighted summation according to at least one internal factor to obtain an internal factor weighted sum includes:

[0035]

[0036] Among them, x2 represents the internal factor weighted sum, x 2i represents one of the internal factors, and ω 2i satisfies

[0037] Among them, performing weighted multiplication according to the external factor weighted sum and the internal factor weighted sum to obtain the observation credibility; includes:

[0038] Calculating the observation credibility according to the following calculation formula:

[0039]

[0040] Among them, x j represents the external factor weighted sum or the internal factor weighted sum; ω j represents the weight corresponding to x j and

[0041] ω1 + ω2 = 2, ω1, ω2 > 0.

[0042] Among them, the observation target includes N attributes;

[0043] Based on the D-S evidence theory, the observation credibility corresponding to at least one observation attribute of the observation target is merged to obtain the target credibility, including:

[0044]

[0045] In the formula

[0046]

[0047] Among them, m j (A j ) represents the basic probability assignment function of the j-th attribute of the observation target to the observation target A j , j = 1, 2,..., N; M N represents the target credibility.

[0048] The method for analyzing the perception behavior of an equipment agent based on the D-S evidence theory provided by this application analyzes the influence of various factors on the perception behavior of the equipment agent and uses the merging method based on the D-S evidence theory for reasoning, making the analysis result of the perception behavior of the equipment agent more accurate and more in line with actual applications. Brief Description of the Drawings

[0049] Figure 1 is the structural block diagram of the perception behavior analysis framework of the equipment agent in an embodiment of this application;

[0050] Figure 2 is the schematic diagram of the influencing factors of the perception behavior of the equipment agent in an embodiment of this application;

[0051] Figure 3 is the flowchart of the method for analyzing the perception behavior of the equipment agent in an embodiment of this application;

[0052] Figure 4 is the calculation flowchart of the observation credibility of the equipment agent in an embodiment of this application. Detailed Description of the Embodiment

[0053] In order to make the purpose, technical solution and advantages of this application clearer, the following further details this application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0054] An agent can serve as a unified model for describing machine intelligence, animal intelligence, and human intelligence, emphasizing the role of rationality. The main body of an agent is an entity that perceives its environment through sensors and acts on that environment through actuators, and can be defined as a mapping from a sequence of perceptions to the actions of the entity. Researchers at home and abroad mainly focus on the research of the agent's perception ability, decision-making ability, control ability, etc. In terms of perception ability, researchers are committed to improving the agent's ability to obtain, process, and recognize environmental information; in terms of decision-making ability, researchers focus on the agent's autonomous decision-making and planning ability in complex environments; in terms of control ability, researchers are committed to achieving precise control and stable operation of the agent.

[0055] Agent-based modeling is a bottom-up modeling method that takes agents as the basic abstract units of the system. Generally, the system model of an agent can be described from the following three levels: the agent layer, the individual agent layer, and the MAS (Multi-Agent Systems) layer. For agent-based behavior modeling, the main work focuses on the individual agent layer, that is, determining the structure of the agent and establishing the behavior model of the agent.

[0056] In the field of simulation, to analyze the perception behavior of an agent through a computer, a perception behavior model should be established first. Currently, the main methods for modeling the perception behavior of an agent mainly include the following several types:

[0057] (1) Rule-based reasoning method

[0058] This method describes the behavior and interaction logic of equipment through preset rules. For example, some rules can be set to describe the reactions of equipment agents in different environments and how they interact with other equipment. The rule-based reasoning method has the advantages of simplicity, intuitiveness, and easy implementation, but may lack sufficient flexibility and adaptability in the face of complex environments.

[0059] (2) Bayesian network-based method

[0060] Bayesian Networks (BN) is a graph theory and statistical method for dealing with uncertain knowledge. In the modeling of the perception behavior of equipment agents, Bayesian networks can be used to describe the perception and decision-making processes of equipment in different situations. By defining nodes and probability relationships, this method can achieve probabilistic reasoning and prediction of equipment behavior. This method has advantages in dealing with uncertainty, but has a high computational complexity and requires defining a large number of probability relationships.

[0061] (3) Machine learning-based method

[0062] Machine learning is a method of building models driven by data. In the modeling of the perception behavior of equipment agents, machine learning algorithms can be used to train models so that they can predict and generate the behavior of equipment agents based on input data (such as environmental information, equipment status, etc.). This method has strong adaptability and learning ability and can handle complex and variable environments. However, the performance of machine learning methods depends on the quality and quantity of training data, as well as the applicability of the selected algorithms.

[0063] It can be seen that there are various methods for modeling the perception behavior of equipment agents, and each method has its advantages and limitations. Therefore, it is necessary to select a suitable modeling method according to the actual situation.

[0064] Among them, perception behavior modeling includes equipment modeling, and equipment modeling includes environmental modeling (2D or 3D appearance models and dynamic models of equipment), physical modeling (weather, terrain, etc.), and behavior modeling. In the prior art, perception behavior models can be divided into two categories: descriptive models and prescriptive or computational models. Currently, the vast majority of perception behavior models are descriptive models, which divide perception behavior into three levels: (1) observing the elements in the environment, identifying key elements, and defining the current situation of the environment; (2) understanding the current situation and forming a situation by synthesizing the observation results of step (1); (3) predicting the trend of the situation. Among the few computational models, early research mainly focused on production perception rules, abstracting the perception behavior model into a production rule system and using rules to observe and understand the current situation. This method is relatively intuitive and easy to implement, but it does not consider the perception differences of perceivers, nor does it consider many deficiencies such as the uncertainty of events, situations, and rules. Since the prior art does not involve how to comprehensively analyze the perception behavior of equipment agents with different target attributes under different influencing factors, the analysis results of the perception behavior of equipment agents in the prior art are inaccurate and difficult to apply to actual scenarios.

[0065] Based on this, the present application provides a method for analyzing the perception behavior of equipment agents based on the D-S evidence theory. This method analyzes the influence of various factors on the perception behavior of equipment agents and uses a combination method based on the D-S evidence theory for reasoning, making the analysis results of the perception behavior of equipment agents more accurate and more suitable for actual applications.

[0066] Such as Figure 1As shown in the figure, the embodiment of the present application provides an equipment agent perception behavior analysis framework 100 based on the D-S evidence theory. The analysis framework 100 includes an observation module 110 and a perception behavior analysis module 120. Among them, the observation module 110 is used to obtain the state information of the equipment agent itself (including position, speed, direction, etc.), the environmental information where the equipment agent is located (including terrain, weather, etc. information), and relevant information such as observation targets through sensors. In the embodiment of the present application, the observation module 110 obtains at least one factor affecting the perception behavior of the equipment agent through direct observation and / or indirect observation. These influencing factors can be divided into external factors and internal factors, and these external factors and internal factors will affect the perception behavior of the equipment agent.

[0067] Among them, the external factors may include the distance between the equipment agent and the observation target and the performance of the detection equipment used by the equipment agent. Among them, the value of the distance d ranges from 0 to 1. d = 1 means that the observation target is very close to the equipment agent. In good weather conditions, without or with general equipment, the attribute value of the observation target can be clearly observed. d = 0 means that the observation target is very far from the equipment agent, and the attribute value of the target cannot be judged with any equipment. The value of the performance q is between 0 and 1. When q = 1, it means that the performance of the reconnaissance equipment used by the equipment agent is very poor; when q = 0, it indicates that the performance of the reconnaissance equipment is particularly excellent, which can help the equipment agent obtain the best observation effect.

[0068] The internal factors may include the observation skills of the equipment agent and the knowledge level of the equipment agent. Among them, the observation skill s is used to represent the physiological and psychological characteristics of the equipment agent, such as all physiological and psychological elements affecting its observation, such as vision, hearing, alertness, intelligence acquisition skills, etc. The value of the observation skill s is between 0 and 1. When s = 1, it means that the observation and analysis skills of the equipment agent are poor and the observation result is not feasible; when s = 0, it means that the equipment agent has very high observation and analysis skills and the observation result is very credible. The knowledge level k is used to represent the knowledge level of the equipment agent, that is, the knowledge of the equipment agent about the characteristics of the target object, the relationship between various attributes, etc. The value of the knowledge level k is between 0 and 1. The value of k can be determined according to factors such as the work experience and education level of the equipment agent. k = 1 indicates that the equipment agent knows nothing about the observation target. At this time, the equipment agent cannot combine the observation results of the various attributes of the observation target for comprehensive reasoning; k = 0 indicates that the equipment agent is very familiar with the observation target.

[0069] Further, the analysis framework 100 of the embodiments of the present application can analyze based on these influencing factors to obtain the observation credibility corresponding to at least one observation attribute of the observation target, that is, by analyzing the above influencing factors, the observation result of the observation target can be obtained. The perception behavior analysis module 120 is used to perform perception behavior analysis based on the multiple influencing factors obtained by the above observation module, and obtain the observation credibility corresponding to the perception behavior, so as to obtain the perception result corresponding to an observation behavior (i.e., an attribute) of the current observation target, such as Figure 2 as shown

[0070] Furthermore, since an observation target has multiple attributes and there is a certain degree of correlation between each attribute, the analysis framework of the embodiments of the present application can utilize this characteristic to realize the mutual reasoning and mutual verification between target attributes, thereby improving the credibility of the observation results of the equipment agent. Specifically, the perception behavior analysis module 120 is further used to analyze and obtain the target perception result of the observation target based on the D-S evidence theory and according to the perception results corresponding to different observation attributes of the observation target.

[0071] The analysis framework of the embodiments of the present application can obtain the influence of various factors on the perception behavior of the equipment agent, and perform comprehensive reasoning by using the merging method based on the D-S evidence theory, so that the analysis result of the perception behavior of the equipment agent is more accurate and more suitable for practical applications.

[0072] Such as Figure 3 as shown, the embodiments of the present application also provide a method for analyzing the perception behavior of an equipment agent based on the D-S evidence theory. This analysis method can be applied to the above analysis framework, and this analysis method can include the following steps:

[0073] S310. Obtain at least one factor affecting the perception behavior of the equipment;

[0074] In the real environment, there are many factors affecting the perception behavior of the equipment. The embodiments of the present application can select at least one key influencing factor to analyze the perception behavior of the equipment agent by using these key influencing factors. These key influencing factors can be divided into external factors affecting the perception behavior of the equipment agent and internal factors affecting the perception behavior of the equipment agent. This can enable the equipment agent to comprehensively obtain various dimensions of influencing factors, which is beneficial to the accuracy and practicality of perception behavior analysis.

[0075] Optionally, the at least one factor includes: the distance d between the equipment agent and the target, the performance q of the detection device used by the equipment agent, the observation skill s of the equipment agent, and the knowledge level k of the equipment agent; wherein, the distance d and the performance q are external factors affecting the perception behavior of the equipment agent, and the observation skill s and the knowledge level k are internal factors affecting the perception behavior of the equipment agent.

[0076] Among them, the value of the distance d ranges from 0 to 1. d = 1 means that the observed target is very close to the equipment intelligent agent. In good weather conditions, without or with general equipment, the attribute value of the observed target can be clearly observed. At this time, the distance between the equipment intelligent agent and the observed target is denoted as d(1). d = 0 means that the observed target is very far from the equipment intelligent agent, and the attribute value of the target cannot be judged with any equipment. At this time, the distance between the equipment intelligent agent and the observed target is denoted as d(0). Specifically, the calculation method of d is as follows:

[0077]

[0078] Among them, d(1) represents the distance between the equipment intelligent agent and the observed target when d = 1; d(0) represents the distance between the equipment intelligent agent and the observed target when d = 0; d(a) represents the distance between the equipment intelligent agent and the observed target when d = a, and the value of a ranges from 0 to 1.

[0079] Among them, the value of the performance q is between 0 and 1. When q = 1, it means that the performance of the reconnaissance equipment used by the equipment intelligent agent is very poor; when q = 0, it indicates that the performance of the reconnaissance equipment is particularly excellent, which can help the equipment intelligent agent obtain the maximum observation effect. Specifically, the calculation method of the performance q is as follows:

[0080]

[0081] Among them, q(1) represents the performance of the detection equipment used by the equipment intelligent agent when q = 1; q(0) represents the performance of the detection equipment used by the equipment intelligent agent when q = 0; q(a) represents the performance of the detection equipment used by the equipment intelligent agent when q = a, and the value of a ranges from 0 to 1.

[0082] The observation skill s is used to represent the physiological and psychological characteristics of the equipment intelligent agent, such as all physiological and psychological elements that affect its observation, such as eyesight, hearing, alertness, intelligence acquisition skills, etc. The value of the observation skill s is between 0 and 1. When s = 1, it means that the observation and analysis skill of the equipment intelligent agent is poor, and the observation result is not feasible; when s = 0, it means that the equipment intelligent agent has a very high observation and analysis skill, and the observation result is very credible. The calculation method of the observation skill s is as follows:

[0083]

[0084] Among them, s(1) represents the observation skill of the equipment intelligent agent when s = 1; s(0) represents the observation skill of the equipment intelligent agent when s = 0; s(a) represents the observation skill of the equipment intelligent agent when s = a, and the value of a ranges from 0 to 1.

[0085] The knowledge level k is used to represent the knowledge level of the equipped intelligent agent, that is, the knowledge of the equipped intelligent agent about the characteristics of the target object, the relationships between various attributes, etc. The value of k can be determined according to factors such as the work experience and educational level of the equipped intelligent agent. The value of the knowledge level k is between 0 and 1. k = 1 indicates that the equipped intelligent agent knows nothing about the observed target. At this time, the equipped intelligent agent cannot associate the observation results of the various attributes of the observed target for comprehensive reasoning; k = 0 indicates that the equipped intelligent agent is very familiar with the observed target. The calculation method of the knowledge level k is as follows:

[0086]

[0087] Among them, k(1) represents the knowledge level of the equipped intelligent agent when k = 1; k(0) represents the knowledge level of the equipped intelligent agent when k = 0; k(a) represents the knowledge level of the equipped intelligent agent when k = a, and the value of a is between 0 and 1.

[0088] S320. Calculate the observation credibility corresponding to at least one attribute of the observed target according to the at least one factor, where the observation credibility is used to represent the observation result of the equipped intelligent agent on an attribute of the observed target.

[0089] Among them, the attributes (i.e., the observation content) of the observed target can include more than one. For each attribute of the observed target in the method of this embodiment of the application, its corresponding observation credibility can be calculated. For each observation credibility, the method of this embodiment of the application can calculate its corresponding observation credibility by using the weighted summation method based on the above influencing factors, and store the attribute of the observed target and its corresponding observation credibility in one-to-one correspondence.

[0090] For example, the attributes of the observed target and their corresponding observation credibilities are stored in one-to-one correspondence in the form of a binary array. Let Φ(t) represent the set of all observation values at time t. Φ(t) is a binary tuple, and this binary tuple includes: <observation content F, observation credibility m>, that is

[0091] Φ(t) = <F, m>

[0092] Among them, F is the observation content, that is, F is an attribute of the observed target, such as "the camp to which the observed target belongs = the other party"; m represents the observation credibility of F, and m is a value between 0 and 1, such as m = 0.6; m indicates that there is a 60% possibility that the observed target belongs to the other party.

[0093] The embodiment of the application can input the real information of the real situation (for example, the moving target 7 km ahead is a truck of the other party (60% possibility)) to the observation module of the analysis framework to improve the effectiveness and reliability of the perception behavior analysis.

[0094] S330. Based on the D-S evidence theory, merge the observation credibility corresponding to at least one observation attribute of the observation target to obtain the target credibility, which is used to represent the perception result of the equipment agent on the observation target.

[0095] Among them, the D-S evidence theory (Dempster-Shafer Theory) is a mathematical theory for dealing with uncertainty and incomplete information. It is an extension of probability theory and can handle uncertainty and ambiguity. It is widely used in fields such as information fusion, decision support, and pattern recognition. In the Dempster-Shafer Theory, the basic probability assignment function (BPA) is used to represent the uncertainty or degree of trust in a certain proposition. The method of this embodiment of the present application can merge at least one observation credibility by adopting the basic probability assignment function of the D-S evidence theory to obtain the cumulative information of the perception behavior, so as to determine the target credibility. Optionally, the target credibility can be calculated according to the following formula:

[0096]

[0097] In the formula

[0098]

[0099] where m j (A j ) represents the basic probability assignment function of the j-th attribute of the observation target for the observation target A j , j = 1, 2,..., N; M N represents the target credibility.

[0100] Since an observation target includes multiple attributes and there is a certain degree of correlation between each attribute, the method of this embodiment of the present application can realize the mutual reasoning and mutual verification between the attributes of the observation target by using this characteristic, so as to improve the credibility of the observation result of the equipment agent.

[0101] In one embodiment, as Figure 4 shown, the above step S320 may include:

[0102] S410. Perform weighted summation according to at least one external factor to obtain the external factor weighted sum; among them,

[0103]

[0104] where x1 represents the external factor weighted sum, x 1i represents one of the external factors; w 1iRepresents the weight corresponding to one of the external factors; weight ω 1i Satisfy

[0105] For example, when i = 1, x 1i Represents the distance d, w 1i Represents the weight corresponding to the distance d; when i = 2, x 1i Represents the performance q, w 1i Represents the weight corresponding to the performance q, and the sum of the weight corresponding to the distance d and the weight corresponding to the performance q is equal to 1.

[0106] S420. Perform weighted summation according to at least one internal factor to obtain the internal factor weighted sum; wherein,

[0107]

[0108] Wherein, x2 represents the internal factor weighted sum, x 2i Represents one of the internal factors, w 2i Represents the weight corresponding to one of the internal factors, ω 2i Satisfy

[0109] For example, when i = 1, x 2i Represents the skill s, w 2i Represents the weight corresponding to the skill s; when i = 2, x 2i Represents the knowledge level k, w 2i Represents the weight corresponding to the knowledge level k; and the sum of the weight corresponding to the skill s and the weight corresponding to the knowledge level k is equal to 1.

[0110] S430. Perform weighted multiplication according to the external factor weighted sum and the internal factor weighted sum to obtain the observation credibility. Wherein, the observation credibility is calculated according to the following calculation formula:

[0111]

[0112] Wherein, x j Represents the external factor weighted sum, or the internal factor weighted sum; for example, when j = 1, x j Represents the external factor weighted sum; when j = 2, x j Represents the internal factor weighted sum.

[0113] ω j Is the weight corresponding to x j For example, when j = 1, ω1 represents the weight corresponding to the external factor weighted sum; when j = 2, ω2 represents the weight corresponding to the internal factor weighted sum. In the embodiments of the present application, ω1 and ω2 satisfy the following conditions:

[0114] ω1 + ω2 = 2, ω1, ω2 > 0.

[0115] The present application proposes a method for analyzing the perception behavior of equipment agents based on the D-S evidence theory. This method establishes a computable empirical model that can quantitatively evaluate the perception ability of equipment agents, mainly analyzes the influence of various factors on the equipment perception behavior, considers that equipment agents often face multiple conflicting evidences, and uses a merging method based on the D-S evidence theory to reason about these evidences, providing support for the analysis of the perception behavior of each subject in equipment simulation. The present application uses a comprehensive analysis method to make the analysis results of equipment perception behavior more accurate and closer to reality, providing technical support for subsequent applications in an autonomous simulation environment.

[0116] The above-described embodiments merely represent several implementation manners of the present application, and their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. An analysis method for the perception behavior of equipment agents based on the D-S evidence theory, characterized in that, The method includes: Obtaining at least one factor that affects the perception behavior of the equipment; Calculating the observation credibility corresponding to at least one attribute of the observation target according to the at least one factor, where the observation credibility is used to represent the observation result of the equipment agent on an attribute of the observation target; Based on the D-S evidence theory, merging the observation credibilities corresponding to at least one observed attribute of the observation target to obtain the target credibility, where the target credibility is used to represent the perception result of the equipment agent on the observation target.

2. The method according to claim 1, characterized in that, The at least one factor includes: The distance d between the equipment agent and the target, the performance q of the detection equipment used by the equipment agent, the observation skill s of the equipment agent, and the knowledge level k of the equipment agent; Wherein, the value of the distance d is between 0 and 1; The value of the performance q is between 0 and 1; The value of the observation skill s is between 0 and 1; The value of the knowledge level k is between 0 and 1.

3. The method according to claim 2, wherein The calculation method of the distance d is as follows: Wherein, d(1) represents the distance between the equipment agent and the observation target when d = 1; d(0) represents the distance between the equipment agent and the observation target when d = 0; d(a) represents the distance between the equipment agent and the observation target when d = a, and a takes a value between 0 and 1; The calculation method of the performance q is as follows: Wherein, q(1) represents the performance of the detection equipment used by the equipment agent when q = 1; q(0) represents the performance of the detection equipment used by the equipment agent when q = 0; q(a) represents the performance of the detection equipment used by the equipment agent when q = a, and a takes a value between 0 and 1; The calculation method of the observation skill s is as follows: Wherein, s(1) represents the observation skill of the equipment agent when s = 1; s(0) represents the observation skill of the equipment agent when s = 0; s(a) represents the observation skill of the equipment agent when s = a, and a takes a value between 0 and 1; The calculation method of the knowledge level k is as follows: Wherein, k(1) represents the knowledge level of the equipment agent when k = 1; k(0) represents the knowledge level of the equipment agent when k = 0; k(a) represents the knowledge level of the equipment agent when k = a, and a takes a value between 0 and 1.

4. The method according to claim 3, wherein Calculating the observation credibility corresponding to at least one attribute of the observation target according to the at least one factor includes: Performing weighted summation according to at least one external factor to obtain the external factor weighted sum; Performing weighted summation according to at least one internal factor to obtain the internal factor weighted sum; Performing weighted multiplication according to the external factor weighted sum and the internal factor weighted sum to obtain the observation credibility.

5. According to the method described in claim 4, performing weighted summation according to at least one external factor to obtain the external factor weighted sum, including: Among them, x1 represents the weighted sum of external factors, and x 1i represents one of the external factors, and the weight ω 1i satisfies 6. According to the method described in claim 4, performing weighted summation according to at least one internal factor to obtain the internal factor weighted sum, including: Among them, x2 represents the weighted sum of internal factors, and x 2i represents one of the internal factors, and ω 2i satisfies 7. According to the method described in claim 4, performing weighted multiplication according to the external factor weighted sum and the internal factor weighted sum to obtain the observation credibility; including: Calculating the observation credibility according to the following calculation formula: where x j represents the weighted sum of external factors or the weighted sum of internal factors; ω j represents the weight corresponding to x j and ω1 + ω2 = 2, ω1, ω2 > 0。 8. The method according to any one of claims 1-7, characterized in that, The observed target includes N attributes; Based on the D-S evidence theory, the observed credibility corresponding to at least one observed attribute of the observed target is merged to obtain the target credibility; including: wherein where m j (A j ) represents the basic probability assignment function of the j-th attribute of the observation target with respect to the observation target A j , j = 1, 2, …, N; M N represents the target credibility.

Images