A multi-radar multi-task adaptive scheduling method based on tracking advantage

Abstract

The present invention relates to a multi-radar multi-task adaptive scheduling method based on tracking advantage, and belongs to the field of radar detection. In view of the fact that the traditional intelligent optimization algorithm does not consider the quality and advantage of the radar in completing the task when allocating the tracking task when solving the multi-target task allocation problem, resulting in a single task scheduling model problem, the present invention first constructs the task tracking advantage, establishes an objective function together with the advantage and benefits, handover times, resource load and other factors, and solves the objective function using the Deep Q Network (DQN) algorithm, thereby improving the advantage of multi-function radar task allocation and effectively reducing the complexity of moving platform radar task allocation. The present invention has a certain reference value for moving platform multi-radar task allocation.

Description

Technical Field

[0001] The invention belongs to the technical field of radar networking detection. Background Art

[0002] In the field of multi-radar mission planning for moving platforms, when the detection area is large, it is necessary to use multiple radars on the moving platform for coordinated detection. When multiple radars are coordinated, task allocation is involved. Task allocation for moving platforms differs from task allocation for stationary radar networks. The former is more difficult and requires more real-time performance. Because the tracking quality of the same target on a moving platform varies at every moment, task allocation requires in-depth research. Summary of the Invention

[0003] Traditional intelligent optimization algorithms for multi-target allocation fail to consider the quality and effectiveness of each radar in assigning tracking tasks, resulting in a single task scheduling model. This paper proposes a multi-radar, multi-task adaptive scheduling method based on tracking effectiveness. First, the task tracking effectiveness is constructed. This objective function is then combined with factors such as efficiency, handover times, and resource load to create an objective function. This objective function is then solved using the Deep Q Network (DQN) algorithm, effectively reducing the complexity of multi-radar task allocation and improving its effectiveness.

[0004] To solve the above technical problems, the present invention proposes a multi-radar multi-task adaptive scheduling method based on tracking favorableness, which includes the following steps:

[0005] S1: Construct a mathematical model for task allocation; the radar configuration in the system is {R1, R2,…, R n}, n is the total number of radars, and the maximum tracking capacity of each radar is D i , In a certain time period, there are m targets to be tracked {M1, M2, ..., M m}, decompose each goal and transform the task into t subtask sequences {T 0, T 1, ...T t,}, after completing the task decomposition, the scheduling result is:

[0006] A=[A1,A2,...A t ]

[0007] in:

[0008] A t =[a ij ] n′m

[0009] In the above formula:

[0010]

[0011] S2: Construct the favorableness function of the tracking task;

[0012] The benefit function of a single task:

[0013]

[0014] Among them, h V It is related to the target's direction of motion relative to the radar and is calculated as follows:

[0015] h V =1-V / 90

[0016] In the above formula, V is the angle between the target's moving direction and the line connecting the radar and the target. That is, when the target moves radially, it is 1, which is the easiest to track. When it moves tangentially, it is 0. array It is related to the position of the target relative to the normal direction of the radar array, and its calculation formula is as follows:

[0017] h V =1-t azi / 60

[0018] In the above formula, t azi It is the angle between the target azimuth and the normal direction of the radar array. When the target is in the normal direction of the array, it is 1. At this time, the radar beam is the narrowest. When the target is at the side of the antenna array, it is 0 at 60 degrees. At this time, the tracking beam width is the largest and the angle measurement accuracy is low. snr It is related to the signal-to-noise ratio of the radar detecting the target, and its calculation formula is a typical tracking radar equation; the tracking advantage of each target at time t is S t ={S1, S2, ..., S M}, then the sum of the benefits of all tasks at all times is:

[0019]

[0020] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0021] S3: Construct the profit function of task allocation; the priority of each goal at time t is P t ={P1, P2...P m}, then the sum of the benefits of all tasks at all times is:

[0022]

[0023] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0024] S4: Construct task execution continuity function:

[0025]

[0026] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0027] S5 builds a task execution constraint function: Task allocation is subject to the following three constraints:

[0028] (1) The maximum number of targets that can be tracked by the radar is limited, namely:

[0029] n is the resource number

[0030] (2) Targets can only be assigned to the set of radars in which they can be detected, that is:

[0031] [a′ ij ] t =[a ij ] t ∧[V ij ] t

[0032] [V ij ] t Indicates whether the target numbered j at time t can be executed by the radar numbered i;

[0033] (3) In a single subtask sequence, only one radar is assigned to detect a single target:

[0034] , m is the target number

[0035] S6: Construct the task allocation objective function; the objective function of the task allocation problem scheduling model is:

[0036]

[0037] Among them, c1, c2, c3 are weight factors, indicating the emphasis of the detection resource allocation system on different influencing factors, and c1+c2+c3=1;

[0038] S7: Use the Deep Q Network (DQN) algorithm to solve the objective function.

[0039] This paper proposes a multi-radar multi-task adaptive scheduling method based on tracking advantage. This method establishes a resource scheduling function model based on tracking advantage, distributed detection benefit, and tracking. It then uses DQN to solve the task problem. This method can solve the multi-radar multi-task scheduling problem, effectively improving resource utilization, reducing resource consumption, and maximizing task execution success rate. This method can be widely applied to multi-sensor task allocation and has certain engineering reference value. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a processing flow chart of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] The preferred embodiment of the present invention comprises the following steps:

[0043] Assume that N radar sensors are maneuvering around a fixed area. Each radar can simultaneously track M targets. At a certain moment, several targets are detected. The settings for some target priorities and other related parameters are shown in Table 1, where each target priority is represented by a random number between (0, 1).

[0044] Table 1 Configuration table of some task parameters

[0045]

[0046] After the detection task is decomposed, the multi-objective allocation problem is transformed into a problem of gradually allocating subtasks according to a time slice sequence. The multi-objective allocation problem is to solve the optimal solution in all feasible solution spaces that meet the constraints of the detection system, so that the detection system can achieve the best working efficiency. The specific solution is as follows:

[0047] S1 builds a mathematical model for task allocation; the radar configuration in the system is {R1, R2,…, R n}, n is the total number of radars, and the maximum tracking capacity of each radar is D i , In a certain time period, there are m targets to be tracked {M1, M2, ..., M m}, decompose each goal and transform the task into t subtask sequences {T 0, T 1, ...T t,}, after completing the task decomposition, the scheduling result is:

[0048] A=[A1,A2,...A t ]

[0049] in:

[0050] A t =[a ij ] n′m

[0051] In the above formula:

[0052]

[0053] S2 constructs the favorableness function of the tracking task: The favorableness function of a single task:

[0054]

[0055] Among them, h V It is related to the target's direction of motion relative to the radar and is calculated as follows:

[0056] h V =1-V / 90

[0057] In the above formula, V is the angle between the target's moving direction and the line connecting the radar and the target. That is, when the target moves radially, it is 1, which is the easiest to track. When it moves tangentially, it is 0. array It is related to the position of the target relative to the normal direction of the radar array, and its calculation formula is as follows:

[0058] h V =1-t azi / 60

[0059] In the above formula, t azi It is the angle between the target azimuth and the normal direction of the radar array. When the target is in the normal direction of the array, it is 1. At this time, the radar beam is the narrowest. When the target is at the side of the antenna array, it is 0 at 60 degrees. At this time, the tracking beam width is the largest and the angle measurement accuracy is low. snr It is related to the signal-to-noise ratio of the radar detecting the target, and its calculation formula is a typical tracking radar equation; the tracking advantage of each target at time t is S t ={S1, S2, ..., S M}, then the sum of the benefits of all tasks at all times is:

[0060]

[0061] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0062] S3. Construct the benefit function of task allocation: the priority of each goal at time t is P t ={P1, P2...P m}, then the sum of the benefits of all tasks at all times is:

[0063]

[0064] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0065] S4. Construct task execution continuity function

[0066]

[0067] The meanings of the symbols and letters in the above formula are the same as those in S1;

[0068] S5. Constructing a task execution constraint function: Task allocation is subject to the following three constraints:

[0069] (1) The maximum number of targets that can be tracked by the radar is limited, namely:

[0070] n is the resource number

[0071] (2) Targets can only be assigned to the set of radars in which they can be detected, that is:

[0072] [a′ ij ] t =[a ij ] t ∧[V ij ] t

[0073] [V ij ] t Indicates whether the target numbered j at time t can be executed by the radar numbered i;

[0074] (3) In a single subtask sequence, only one radar is assigned to detect a single target:

[0075] , m is the target number

[0076] S6. Constructing the objective function of task allocation: The objective function of the task allocation problem scheduling model is:

[0077]

[0078] Among them, c1, c2, c3 are weight factors, indicating the emphasis of the detection resource allocation system on different influencing factors, and c1+c2+c3=1;

[0079] S7. Use the Deep Q Network (DQN) algorithm to solve the objective function.

Claims

1. A multi-radar multi-task adaptive scheduling method based on tracking advantage, characterized by: S1. Construct a mathematical model for task allocation. The radar configuration in the system is {R1, R2, …, R n }, n is the total number of radars, and the maximum tracking capacity of each radar is D i , In a certain period of time, there are m targets to be tracked {M1,M2,…,M m }, decompose each goal and transform the task into t subtask sequences {T 0, T 1, ...T t, }, after completing the task decomposition, the scheduling result is: <h2 style=";text-align:left;direction:ltr">A=[A1,A2,...A<h2 style=";text-align:left;direction:ltr"> t <h2 style=";text-align:left;direction:ltr"> ] in: A t =[a ij ] n′m In the above formula: S2. Construct the favorableness function of the tracking task: The favorableness function of a single task: Among them, h V It is related to the target's direction of motion relative to the radar and is calculated as follows: h V =1-V / 90 In the above formula, V is the angle between the target's moving direction and the line connecting the radar and the target, that is, when the target moves radially, it is 1, which is the easiest to track, and when it moves tangentially, it is 0; Among them, h array It is related to the position of the target relative to the normal direction of the radar array, and its calculation formula is as follows: h V =1-t azi / 60 In the above formula, t azi It is the angle between the target azimuth and the normal direction of the radar array. When the target is in the normal direction of the array, it is 1. At this time, the radar beam is the narrowest. When the target is at the side of the antenna array, it is 0 at 60 degrees. At this time, the tracking beam width is the largest and the angle measurement accuracy is low. snr It is related to the signal-to-noise ratio of the radar detecting the target, and its calculation formula is the typical tracking radar equation; The tracking advantage of each target at time t is S t ={S1, S2, ..., S M }, then the sum of the benefits of all tasks at all times is: The meanings of the symbols and letters in the above formula are the same as those in S1; S3. Construct the benefit function of task allocation: the priority of each goal at time t is P t = {P1, P2...P m }, then the sum of the benefits of all tasks at all times is: The meanings of the symbols and letters in the above formula are the same as those in S1; S4. Construct task execution continuity function: The meanings of the symbols and letters in the above formula are the same as those in S1; S5. Construct task execution constraint function: Task allocation is subject to the following three constraints: (1) The maximum number of targets that can be tracked by the radar is limited, that is, n is the resource number (2) Targets can only be assigned to the set of radars in which they can be detected, that is: [a′ ij ] t =[a ij ] t ∧[V ij ] t [V ij ] t Indicates whether the target numbered j at time t can be executed by the radar numbered i; (3) In a single subtask sequence, only one radar is assigned to detect a single target: m is the target number S6. Constructing the objective function of task allocation: The objective function of the task allocation problem scheduling model is: Among them, c1, c2, c3 are weight factors, indicating the emphasis of the detection resource allocation system on different influencing factors, and c1+c2+c3=1; S7. Use the Deep Q Network algorithm to solve the objective function.

Images