A radar resource management and multi-task scheduling method and system based on projection geometry regulation

Abstract

The application discloses a radar resource management and multi-task scheduling method and system based on projection geometry regulation. The method maps multi-dimensional resources (time, frequency, energy, beam, calculation) of a radar system into a resource vector R_res in a projection angle space, represents each task demand as a projection demand vector D_k, calculates a task urgency projection index U_k (determined by a projection angle between D_k and R_res), solves a projection angle optimization problem minΣλ_k U_k to realize multi-task scheduling decision, and has a scheduling complexity O(D·K). The system comprises a resource projection space construction module, a task demand modeling module, an urgency calculation module, a projection angle optimization scheduler, a closed-loop feedback monitoring module, a mode switching controller and a scheduling instruction output interface. The application can realize scheduling decision in ≤100μs and mode switching in ≤1μs.

Description

Technical Field

[0001] This invention relates to the field of radar resource management and multi-task scheduling technology, specifically to a radar resource management and multi-task scheduling method and system based on projection geometry control. This invention can be applied to multi-functional phased array radars, airborne radars, shipborne radars, air defense early warning radars, multi-functional integrated radio frequency systems, and is particularly suitable for radar system scenarios involving multiple concurrent tasks, intense resource competition, and the need for real-time optimized scheduling. Background Technology

[0002] Modern phased array radars routinely face the challenge of concurrent multi-task execution: multiple tasks such as search, tracking, imaging, anti-jamming, communication, and electronic warfare support simultaneously compete for time, frequency, energy, beam, and computing resources. How to achieve efficient scheduling of multiple tasks under limited resources is the core challenge of radar resource management. Existing radar resource management and multi-task scheduling technologies mainly include the following solutions: (a) Priority-based static scheduling A fixed priority is preset for different tasks, and resources are allocated according to the priority order. This method has the following problems: (1) It cannot adapt to the dynamically changing battlefield environment. When a high-priority task suddenly occurs, the low-priority task is postponed indefinitely; (2) The resource utilization rate is low. Fixed allocation results in idle resources not being fully utilized; (3) It lacks global optimization and resource competition between tasks cannot be coordinated. (ii) Time-based polling scheduling The time is divided into fixed time slots, and each task polls and occupies the time slots. This method has the following problems: (1) low time utilization, and a large number of idle time slots are caused when the task load is uneven; (2) it cannot support the immediate response to sudden tasks; (3) the time granularity of different tasks is large, and it is difficult to adapt the fixed time slots. (III) Dynamic scheduling based on heuristic algorithms Heuristic methods such as genetic algorithms and particle swarm optimization are used for task scheduling. This method has the following problems: (1) large computational load, slow convergence, and difficulty in real-time response; (2) high algorithm complexity, making it difficult to implement on embedded platforms; (3) lack of ability to quickly track changes in resource status. (iv) Adaptive scheduling based on cognitive radar Using a cognitive radar architecture, task parameters are adaptively adjusted based on environmental perception results. This method has the following problems: (1) cognitive processing itself consumes a lot of computing resources; (2) there is a lack of a unified resource description framework, and each resource dimension is managed independently; (3) the resource coupling relationship between multiple tasks is difficult to model accurately. The common limitation of existing technologies is that: There is a lack of a unified resource description framework, and resource dimensions such as time, frequency, energy, and beam are managed separately; Scheduling decisions involve a large amount of computation and have poor real-time performance, making it difficult to complete rescheduling in milliseconds. It cannot support real-time dynamic changes in task priority and resource requirements; There is a lack of a closed-loop feedback optimization mechanism for resource utilization. In multi-task concurrent scenarios, resource contention leads to a severe decline in overall performance. To address the aforementioned problems, this invention provides a radar resource management and multi-task scheduling method and system based on projection geometry control. Summary of the Invention

[0003] (a) Purpose of the invention This invention aims to address the technical problems in existing radar resource management and multi-task scheduling technologies, such as the lack of a unified resource description framework, large computational load and poor real-time performance of scheduling decisions, inability to support dynamic changes, and lack of closed-loop feedback optimization mechanisms. It provides a radar resource management and multi-task scheduling method and system based on projection geometry control. (II) Technical Principles This invention is based on the principles of projection geometry. In this invention, radar resource management and multi-task scheduling are achieved through the following physical chain: Unified mapping of resource projection space: The multidimensional heterogeneous resources (time, frequency, energy, beam, computation) of the radar system are uniformly mapped to resource vectors R_res = (θ_T, θ_F, θ_E, θ_B, θ_C) in the projection angle space. Each resource dimension is represented by the corresponding projection angle, satisfying the projection constraint β_r² + β_s² = 1. This mapping allows heterogeneous resources to be compared and optimized within the same geometric framework. Projective representation of task requirement vector: The requirement of each radar task is represented as a projected requirement vector D_k = (θ_Tk, θ_Fk, θ_Ek, θ_Bk, θ_Ck), where each component represents the degree of requirement of the task for various types of resources. The projective representation of task requirements allows the resource competition relationship between tasks to be directly quantified in the projection angle space. Geometric Measure of Task Urgency: Task urgency is determined by the angle between the projections of the task requirement vector and the current resource state vector. The smaller the angle, the higher the match between the task requirement and the current resource state, and the more urgent the task. This geometric measure transforms the resource matching problem into a vector angle calculation problem. Projection Angle Optimization Scheduling: This method transforms the multi-task scheduling problem into an optimization problem in the projection angle space. By solving the constraint optimization problem, the optimal resource allocation scheme is obtained with a complexity of only O(D·K), which is far superior to traditional heuristic algorithms. Closed-loop feedback and dynamic rescheduling: By monitoring the system resource utilization rate in real time, dynamic rescheduling is triggered when the resource status changes exceed the preset threshold or a new task arrives, so as to realize the real-time optimization of resource allocation. The fundamental difference from existing technologies: This invention is not a "fixed priority scheduling" or "heuristic search" solution, but rather achieves unified description and efficient optimized scheduling of heterogeneous resources through projective geometry. (III) Technical Solution A radar resource management and multi-task scheduling method based on projection geometry control, characterized by the following steps: (1) Construct a radar resource projection space and map the multidimensional resources of the radar system into a resource vector R_res = (θ_T, θ_F, θ_E, θ_B, θ_C, …) in the projection angle space. Each resource dimension θ_i = arctan(used resources / available resources) satisfies the projection constraint β_r² + β_s² = 1. (2) The requirements of each radar task k are represented as a projected requirement vector D_k = (θ_Tk, θ_Fk, θ_Ek, θ_Bk, θ_Ck, …), where each component represents the degree of requirement of the task for various types of resources; (3) Calculate the task urgency projection index U_k, which is determined by the projection angle between the projection demand vector D_k and the current resource state vector R_res, and multiplied by the task priority weight w_k: U_k = arccos(<D_k, R_res> / (||D_k|| · ||R_res||)) · w_k (4) The multi-task scheduling problem is transformed into an optimization problem in the projection angle space. Under the condition of satisfying the system constraints, the weighted sum of the urgency of all tasks is minimized by adjusting Δθ_res by the projection angle: min Σ λ_k · U_k(Δθ_res) st Resource availability constraints, total resource constraints, and task performance constraints; (5) By monitoring the system resource utilization and task completion quality in real time, update the resource state vector R_res. When the resource state changes exceed the preset threshold or a new task arrives, trigger dynamic rescheduling. (6) The scheduling strategy preference is automatically switched according to the system working mode, and the weight of the objective function of the projection angle optimization problem is automatically adjusted in each mode. A radar resource management and multi-task scheduling system based on projection geometry control, characterized in that it includes: The resource projection space construction module is used to map the multidimensional resources of the radar system into a resource vector R_res in the projection angle space; The task requirement modeling module is used to represent the requirements of each radar task as a projected requirement vector D_k; The urgency calculation module is connected to the resource projection space construction module and the task requirement modeling module, and is used to calculate the task urgency projection index U_k. The projection angle optimization scheduler, connected to the urgency calculation module, is used to solve the multi-task scheduling optimization problem in the projection angle space and output the scheduling scheme Δθ_res; The closed-loop feedback monitoring module is used to monitor the system resource utilization and task completion quality in real time. When the resource status changes exceed the preset threshold or a new task arrives, dynamic rescheduling is triggered. The mode switching controller is used to automatically switch scheduling strategy preferences according to the system's operating mode; The scheduling instruction output interface is used to output the scheduling scheme to each execution unit. (iv) Beneficial effects First, a unified resource description framework: heterogeneous resources such as time, frequency, energy, beam, and computing are uniformly mapped into a projection angle space, enabling unified management of multi-dimensional resources. Second, low-complexity scheduling decision: the solution complexity of the projection angle optimization problem is O(D·K), which is far better than O(K²) or higher of traditional heuristic algorithms. Third, real-time dynamic scheduling: projection angle modulation response ≤1μs, supporting real-time rescheduling when new tasks arrive. Fourth, closed-loop feedback optimization: dynamically adjust scheduling strategies based on resource utilization and task quality feedback. Fifth, multi-mode adaptive: supports automatic switching of scheduling preferences for different working modes such as search / tracking / imaging / electronic warfare. Detailed Implementation

[0004] Example 1 (Multi-task scheduling system based on resource projection space) Step 1: Construct the resource projection space. Radar system configuration: Time resource T_total = 100% (normalized), Frequency resource F_total = 500MHz, Energy resource E_total = 1kW, Beam resource B_total = 16 (beams), Computational resource C_total = 1TFLOPS. The initial state of the five-dimensional resource projection space is R_res = (0.2, 0.3, 0.1, 0.4, 0.2) rad, corresponding to a resource utilization rate η_res = 0.24. Step 2: Define the task projection requirement vector. The current system needs to execute K=4 tasks: (1) Spatial search task (covering 90° of the airspace), D1 = (0.6, 0.3, 0.4, 0.8, 0.2); (2) High-priority tracking task (tracking target T1), D2 = (0.4, 0.2, 0.5, 0.6, 0.3); (3) Medium-priority tracking task (tracking target T2), D3 = (0.3, 0.2, 0.3, 0.5, 0.2); (4) Imaging task (synthetic aperture imaging), D4 = (0.7, 0.8, 0.6, 0.3, 0.6). The priority weight of each task is w = [0.7, 0.9, 0.6, 0.5]. Step 3: Calculate the urgency of the task. The urgency of each task is calculated using the formula: U1 = 0.42, U2 = 0.28, U3 = 0.38, U4 = 0.55. The overall urgency is U_total = Σλ_k U_k = 0.35. Step 4: Perform projection angle optimization scheduling. The optimizer solves for min Σλ_k U_k(Δθ_res) to obtain the optimal scheduling scheme: allocate 30% of time, 20% of frequency, 15% of energy, and 6 beams to the search task; allocate 40% of time, 25% of frequency, 30% of energy, and 4 beams to the tracking task; and allocate 30% of time, 55% of frequency, 55% of energy, and 2 beams to the imaging task. After optimization, the overall urgency is reduced to 0.22, an improvement of 37%. Step 5: Output scheduling instructions. The scheduling plan is output to each execution unit. The search task continues to execute, the tracking task updates every 50ms, and the imaging task completes data acquisition and imaging processing within 1 second. Example 2 (Dynamic Rescheduling - Sudden Target Tracking) Step 1: Initial scheduling run. The system is in search + tracking mode, and the scheduling scheme is running stably. Step 2: Sudden target detection. The search task detected two new targets, requiring immediate tracking. The system allocated resources from the idle resource pool: 0.1 units of time, 0.05 units of frequency, 0.1 units of energy, and 1 beam. However, due to the current high resource utilization rate, the overall urgency U_total increased from 0.22 to 0.31 after the initial resource allocation. Step 3: Automatically trigger rescheduling. If resource changes exceed the preset threshold δ_th = 0.05, the system automatically triggers rescheduling. The rescheduling decision is completed within <50μs. Step 4: Generate rescheduling scheme. The time allocation for the search task was reduced from 30% to 25%, the frequency from 20% to 15%, and the energy allocation from 15% to 10%, freeing up resources for two new target tracking tasks. After rescheduling, the overall urgency U_total of the four tasks decreased to 0.26, satisfying the system constraints. The total rescheduling time was <50μs (including computation <10μs + switching <40μs). The tracking task was established immediately, and target tracking was successful. Example 3 (Multi-mode adaptive scheduling - search → track → electronic warfare mode switching) Step 1: Search dominant mode. During peacetime, the system operates in search-dominated mode, with the search task weight λ_search = 1.0 in the scheduling objective function, while other tasks have lower weights. The system aims to maximize the coverage of the search spatial domain. Step 2: Track the switching of the dominant mode. When a high-value target (such as an enemy aircraft) is detected, the system automatically switches to a tracking-dominated mode. The search task weight λ_search decreases to 0.3, while the tracking task weight λ_track increases to 1.0. The projection angle optimization objective function is adjusted accordingly: the priority of the search task is reduced, the priority of the tracking task is increased, and beam resources are tilted towards the tracking task (two beams are allocated from the search beam and converted to the tracking beam). Step 3: Switching to electronic warfare-dominated mode. When enemy radar illumination is detected, the system switches to electronic warfare mode, with an electronic warfare task weight λ_EW = 1.0. Some radar resources are converted to electronic countermeasures (jamming suppression, chaff deployment control, etc.). The electronic warfare task requirement vector is introduced into the projection angle optimization objective function, and the scheduling algorithm automatically balances resource allocation between radar detection and electronic warfare. The beneficial effect of this embodiment is that by optimizing the weight reconfiguration of the objective function through projection angle, a smooth switching between different working modes can be achieved with a switching time ≤1μs. Example 4 (Closed-Loop Verification of Integrated Dispatch System) Step 1: System integration. The resource management and scheduling system of this invention is integrated into the radar platform and works in collaboration with each execution unit. Step 2: Coordinating scheduling and execution. The scheduling instructions output by the resource management and scheduling system are distributed to each execution unit: waveform switching response time <100ns; anti-interference mode switching response time <1μs; communication / radar mode switching response time <1μs; preprocessing parameter configuration and detection and tracking parameter configuration are completed synchronously. Step 3: Closed-loop performance verification. Verify the consistency between the scheduling scheme and the actual execution effect: system resource utilization η_res ≥ 85%, task completion rate ≥ 95%, and average scheduling response time < 100μs. Attached Figure Description

[0005] Figure 1 This is a schematic diagram of the overall structure of the radar resource management and multi-task scheduling system based on projection geometry control according to the present invention. Figure 2 This is a schematic diagram of the resource projection space and task projection requirement vector of the present invention. Figure 3 This is a flowchart of the projection angle optimization scheduling decision-making process of the present invention. Figure 4 This is a flowchart of the closed-loop resource feedback and dynamic rescheduling process of the present invention. Terminology Explanation

[0006] Resource projection space: The projection angle space that maps the multi-dimensional resources of the radar system in a unified manner. Resource vector R_res: A multidimensional projection angle vector describing the current resource state. Projection demand vector D_k: Describes the projection angle vector of task k for various resource requirements. Task urgency U_k: An indicator of the urgency of task k, determined by the angle between the projections of the demand vector and the resource vector. Projection Angle Optimization Scheduler: A hardware unit for solving multi-task scheduling optimization problems in projection angle space. Scheduling decision time: The total delay from the input of task requirements to the output of the scheduling plan. Mode switching controller: A hardware unit that controls the switching between different operating modes of the system.

Claims

1. Claim 1: A radar resource management and multi-task scheduling method based on projection geometry control, characterized in that, Includes the following steps: (1) Construct a radar resource projection space and map the multidimensional resources of the radar system into a resource vector R_res = (θ_T, θ_F, θ_E, θ_B, θ_C, …) in the projection angle space. Each resource dimension θ_i = arctan(used resources / available resources) satisfies the projection constraint β_r² + β_s² = 1. (2) The requirements of each radar task k are represented as a projected requirement vector D_k = (θ_Tk, θ_Fk, θ_Ek, θ_Bk, θ_Ck, …), where each component represents the degree of requirement of the task for various types of resources; (3) Calculate the task urgency projection index U_k, which is determined by the projection angle between the projection demand vector D_k and the current resource state vector R_res, and multiplied by the task priority weight w_k: U_k = arccos(<D_k, R_res> / (||D_k|| · ||R_res||)) · w_k; (4) The multi-task scheduling problem is transformed into an optimization problem in the projection angle space. Under the condition of satisfying the system constraints, the weighted sum of the urgency of all tasks is minimized by adjusting Δθ_res by the projection angle: min Σ λ_k · U_k(Δθ_res) st Resource availability constraints, total resource constraints, and task performance constraints; (5) By monitoring the system resource utilization and task completion quality in real time, update the resource state vector R_res. When the resource state changes exceed the preset threshold or a new task arrives, trigger dynamic rescheduling. (6) The scheduling strategy preference is automatically switched according to the system working mode, and the weight of the objective function of the projection angle optimization problem is automatically adjusted in each mode.

2. Claim 2: The method according to claim 1, characterized in that, The dimension D of the resource projection space mentioned in step (1) is greater than or equal to 5, including five dimensions: time, frequency, energy, beam and computation. It can be extended to D≤10 according to the system configuration.

3. Claim 3: The method according to claim 1, characterized in that, The solution complexity of the projection angle optimization problem in step (4) is O(D·K), where D is the number of resource dimensions and K is the number of tasks.

4. Claim 4: The method according to claim 1, characterized in that, The preset threshold in step (5) is δ_th = 0.05 (resource utilization change), and the rescheduling decision time is ≤100μs.

5. Claim 5: The method according to claim 1, characterized in that, The system operating modes described in step (6) include at least one of the following: search-dominated mode, tracking-dominated mode, imaging-dominated mode, and electronic warfare-dominated mode. The switching of each mode is achieved by adjusting the weights of the projection angle optimization objective function, and the switching time is ≤1μs.

6. Claim 6: The method according to claim 1, characterized in that, The resource availability constraint in step (4) is θ_res +Δθ_res ∈ [θ_min, θ_max], the total resource constraint is ΣD_k ≤ R_total, and the task performance constraint is U_k(Δθ_res) ≤ U_k_max.

7. Claim 7: A radar resource management and multi-task scheduling system based on projection geometry control, characterized in that, include: The resource projection space construction module is used to map the multidimensional resources of the radar system into a resource vector R_res in the projection angle space; The task requirement modeling module is used to represent the requirements of each radar task as a projected requirement vector D_k; The urgency calculation module is connected to the resource projection space construction module and the task requirement modeling module, and is used to calculate the task urgency projection index U_k. The projection angle optimization scheduler, connected to the urgency calculation module, is used to solve the multi-task scheduling optimization problem in the projection angle space and output the scheduling scheme Δθ_res; The closed-loop feedback monitoring module is used to monitor the system resource utilization and task completion quality in real time. When the resource status changes exceed the preset threshold or a new task arrives, dynamic rescheduling is triggered. The mode switching controller is used to automatically switch scheduling strategy preferences according to the system's operating mode; The scheduling instruction output interface is used to output the scheduling scheme to each execution unit.

8. Claim 8: The system according to claim 7 is characterized in that, The solution complexity of the projection angle optimization scheduler is O(D·K), and the scheduling decision time is ≤100μs, where D≤10 is the resource dimension and K≤32 is the number of concurrent tasks.

9. Claim 9: The method according to claim 1, characterized in that, The scheduling instruction output interface synchronously sends the scheduling scheme to the waveform generation unit, the anti-interference control unit, and the communication mode switching unit.

10. Claim 10: The system according to claim 7 is characterized in that, The switching modes supported by the mode switching controller include: Search-driven mode: Search task weight λ_search = 1.0; Tracking-dominated mode: Tracking task weight λ_track = 1.0; Imaging-dominated mode: Imaging task weight λ_imaging = 1.0; Electronic warfare-dominated mode: Electronic warfare mission weight λ_EW = 1.0; The switching time between modes is ≤1μs.

Images