CN117171877A - Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game - Google Patents
Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game Download PDFInfo
- Publication number
- CN117171877A CN117171877A CN202311102259.2A CN202311102259A CN117171877A CN 117171877 A CN117171877 A CN 117171877A CN 202311102259 A CN202311102259 A CN 202311102259A CN 117171877 A CN117171877 A CN 117171877A
- Authority
- CN
- China
- Prior art keywords
- hypersonic aircraft
- interceptor
- penetration
- initial
- overload
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
技术领域Technical field
本发明属于飞行器技术领域,具体涉及一种高超声速飞行器机动突防策略设计方法。The invention belongs to the technical field of aircraft, and specifically relates to a method for designing a maneuvering penetration strategy for hypersonic aircraft.
背景技术Background technique
在高超声速飞行器的近距突防过程中,受自身机动能力约束,以及拦截弹机动能力和制导律的不确定性,对近距短时内的有效突防造成了很大的困难,所以如何利用攻方的机动主动性以及高速特性,并结合智能学习方法,提出高超声速飞行器智能机动突防策略,从而实现高超声速飞行器在上述约束和不确定场景下的成功突防,是拟解决的关键科学问题。During the short-range penetration process of hypersonic aircraft, the constraints of its own maneuverability and the uncertainty of the interceptor missile's maneuverability and guidance law have caused great difficulties to effective penetration in a short period of time. So how to Utilizing the maneuvering initiative and high-speed characteristics of the attacker, combined with intelligent learning methods, to propose an intelligent maneuvering penetration strategy for hypersonic aircraft, so as to achieve successful penetration of hypersonic aircraft under the above constraints and uncertain scenarios, is the key to be solved Scientific question.
综合国内外相关研究现状,针对高超声速飞行器的突防问题,更多集中在程序机动突防策略,而对于博弈机动突防方面,主要研究了基于单边最优理论的突防制导律以及基于微分对策的突防制导律。上述两种方法推导的高超声速飞行器机动突防策略将数学泛函的单/双边极值问题引进到高超飞行器与反导拦截器的攻防对抗模型中,将突防飞行器与拦截弹设为博弈双方,加入再入点位置和速度等终端约束、机动过载和控制变量约束,以终端脱靶量等为性能指标,结合突防飞行器和拦截弹的运动模型构建Hamilton函数,由极值的必要条件求解最优突防和拦截策略。基于单边最优理论的突防制导律需要事先假设敌方拦截弹的制导律,这在真实攻防对抗场景中敌方的制导信息是很难获取的,从而给工程实现造成了一定的困难;基于微分对策的突防制导律假设攻防双方同时采用最优机动策略,所设计的突防制导律较为保守,无法充分发挥高超声速飞行器的突防能力。而关于智能方法与突防策略的结合,在无人机、巡航导弹等对抗场景上取得了一定进展,但考虑到高超声速飞行器的攻防对抗场景具有高动态快时变等特点,现有针对无人机、巡航导弹等对抗场景的智能机动突防策略并不适用。Based on the current status of relevant research at home and abroad, for the penetration problem of hypersonic aircraft, more focus is on procedural maneuvering penetration strategies. For game maneuvering penetration, the penetration guidance law based on unilateral optimal theory and the penetration guidance law based on game maneuvering penetration are mainly studied. Penetration guidance law of differential countermeasures. The maneuvering penetration strategy of hypersonic aircraft derived by the above two methods introduces the unilateral/bilateral extreme value problem of mathematical functionals into the offensive and defensive confrontation model between hypersonic aircraft and anti-missile interceptors, and sets the penetration aircraft and interceptor missiles as two parties in the game. , adding terminal constraints such as re-entry point position and speed, maneuvering overload and control variable constraints, taking the terminal miss amount as performance indicators, and constructing a Hamilton function based on the motion model of the penetration aircraft and interceptor missile, and solving the optimal condition based on the necessary conditions for extreme values. Excellent penetration and interception strategies. The penetration guidance law based on the unilateral optimal theory requires the assumption of the guidance law of the enemy's interceptor missile in advance. This makes it difficult to obtain the enemy's guidance information in a real offensive and defensive confrontation scenario, thus causing certain difficulties in engineering implementation; The penetration guidance law based on differential game assumes that both the attacker and defender adopt optimal maneuver strategies at the same time. The designed penetration guidance law is relatively conservative and cannot fully utilize the penetration capability of the hypersonic aircraft. As for the combination of intelligent methods and penetration strategies, certain progress has been made in confrontation scenarios such as drones and cruise missiles. However, considering that the offensive and defensive confrontation scenarios of hypersonic aircraft have the characteristics of high dynamics and fast time-varying, the existing countermeasures against unmanned aerial vehicles have Intelligent mobile penetration strategies for confrontation scenarios such as man-machine and cruise missiles are not applicable.
考虑到突防策略的工程实现,真实攻防对抗场景中,往往存在获取信息延迟以及信息准确性差等问题。Considering the engineering implementation of penetration strategies, in real offensive and defensive confrontation scenarios, there are often problems such as delays in obtaining information and poor information accuracy.
发明内容Contents of the invention
为了克服现有技术的不足,本发明提供了一种基于时机博弈的高超声速飞行器机动突防策略设计方法,首先基于高超声速飞行器在近距对抗中的优劣势分析,提出基于突防窗口的博弈突防策略;然后在横侧向平面内建立近似逆轨拦截态势下的攻防对抗数学模型;接下来基于突防窗口的博弈突防策略,对工程应用中不同的作战态势、敌我机动策略和机动能力等进行蒙特卡洛仿真,生成博弈对抗数据库;最后通过神经网络对数据库进行离线学习,在线计算生成突防窗口指导高超声速飞行器完成对不同场景的突防。在线计算过程中,针对不确定的敌方信息和态势信息,可结合数据库将该输入项取极值或多个取值,最终在所生成的多个窗口中寻找公共区间,即为公共突防窗口。In order to overcome the shortcomings of the existing technology, the present invention provides a method for designing a maneuvering penetration strategy for hypersonic aircraft based on timing games. First, based on the analysis of the advantages and disadvantages of hypersonic aircraft in close confrontation, a game based on penetration windows is proposed. Penetration strategy; then establish a mathematical model of offensive and defensive confrontation in an approximate reverse-track interception situation in the lateral plane; and then develop a game penetration strategy based on the penetration window to analyze different combat situations, friendly and enemy maneuver strategies and maneuvers in engineering applications Monte Carlo simulation is performed on the capabilities and other capabilities to generate a game confrontation database; finally, the database is learned offline through neural network, and online calculation is performed to generate a penetration window to guide the hypersonic aircraft to complete penetration in different scenarios. During the online calculation process, for uncertain enemy information and situational information, the input item can be combined with the database to take the extreme value or multiple values, and finally find the common interval in the multiple generated windows, which is the public penetration window.
本发明解决其技术问题所采用的技术方案包括如下步骤:The technical solution adopted by the present invention to solve the technical problems includes the following steps:
步骤1:基于高超声速飞行器在近距对抗中的优劣势分析,提出突防窗口的作战概念;Step 1: Based on the analysis of the advantages and disadvantages of hypersonic aircraft in close confrontation, propose an operational concept of penetration window;
所述突防窗口是指高超声速飞行器通过满过载机动实现对拦截导弹的突防的相对距离区间;所述近距指在高超声速飞行器的探测范围内;The penetration window refers to the relative distance interval within which the hypersonic aircraft can penetrate the interceptor missile through full overload maneuvers; the short range refers to the detection range of the hypersonic aircraft;
步骤2:在横侧向平面内建立近似逆轨拦截态势下的攻防对抗数学模型;Step 2: Establish a mathematical model of offensive and defensive confrontation in an approximate reverse orbit interception situation in the lateral plane;
步骤3:利用蒙特卡洛仿真遍历工程应用中不同的作战态势、敌我机动策略、机动能力,生成博弈对抗数据库;Step 3: Use Monte Carlo simulation to traverse different combat situations, enemy and friendly maneuver strategies, and maneuver capabilities in engineering applications, and generate a game confrontation database;
步骤4:基于BP神经网络对步骤3中生成的博弈对抗数据库进行离线学习。Step 4: Perform offline learning on the game confrontation database generated in step 3 based on BP neural network.
进一步的,所述步骤2具体为:Further, the step 2 is specifically:
所述近似逆轨拦截态势是指特定的角度范围,当高超声速飞行器和拦截弹双方初始速度指向偏差大于等于该角度范围时,高超声速飞行器无需进行机动突防,利用自身速度即能完成突防;当双方初始速度指向偏差小于该角度范围时,则需要进行机动突防;The approximate reverse-orbit interception situation refers to a specific angular range. When the initial speed pointing deviation of both the hypersonic aircraft and the interceptor missile is greater than or equal to this angular range, the hypersonic aircraft does not need to maneuver for penetration and can complete the penetration using its own speed. ; When the initial speed pointing deviation of both parties is less than this angle range, a maneuverable penetration is required;
拦截弹在末制导段的数学描述如下:The mathematical description of the interceptor missile in the terminal guidance section is as follows:
式中,Vm(t)表示拦截弹运动速度矢量;γm表示拦截弹的弹道偏角;nzm表示拦截弹实际的飞行过载;g表示重力加速度;表示拦截弹在x轴上的速度分量;/>表示拦截弹在z轴上的速度分量;/>为工程实际中的拦截弹制导律,包括比例导引律PN、修正比例导引律APN和自适应滑模导引律ASMG;在横侧向平面内这三种拦截弹导引律表示如下:In the formula, V m (t) represents the velocity vector of the interceptor; γ m represents the ballistic deflection angle of the interceptor; n zm represents the actual flight overload of the interceptor; g represents the acceleration of gravity; Represents the velocity component of the interceptor on the x-axis;/> Represents the velocity component of the interceptor on the z-axis;/> It is the guidance law of interceptor missiles in actual engineering, including proportional guidance law PN, modified proportional guidance law APN and adaptive sliding mode guidance law ASMG; these three types of interceptor missile guidance laws in the lateral plane are expressed as follows:
比例导引律:Proportional guiding law:
修正比例导引律:Modified proportional guidance law:
自适应滑模导引律:Adaptive sliding mode guidance law:
式中:为从拦截弹视角出发的视线角速度;Vc为攻防双方接近速度;N,ε和δ均为制导律参数;ah表示高超声速飞行器飞行的实际加速度,该修正项保证了当目标做恒值机动时,拦截弹采取过载补偿措施;um表示使用不同导引律拦截弹的指令加速度;/>表示从拦截弹视角出发的方位角速度;In the formula: is the line-of-sight angular velocity from the perspective of the interceptor; V c is the approach speed of both attackers and defenders; N, ε and δ are all guidance law parameters; a h represents the actual acceleration of the hypersonic aircraft flight. This correction term ensures that when the target is a constant When maneuvering, the interceptor missile adopts overload compensation measures; u m represents the command acceleration of the interceptor missile using different guidance laws;/> Represents the azimuth angular velocity from the interceptor's perspective;
同时,拦截弹在整个拦截过程采用STT控制方式,其所受约束仅为最大可用过载约束nzmmax,表示为:At the same time, the interceptor adopts the STT control method during the entire interception process, and its constraints are only the maximum available overload constraint n zmmax , expressed as:
|nzm|≤nzmmax |n zm |≤n zmmax
对于高超声速飞行器,其具体描述如下:For hypersonic aircraft, its specific description is as follows:
其中,Vh(t)表示高超声速飞行器运动速度矢量,γh表示高超声速飞行器的弹道偏角,τh表示高超声速飞行器的一阶环节时间常数,nzhc表示高超声速飞行器的过载指令,nzh表示高超声速飞行器的实际飞行过载,表示高超声速飞行器在x轴上的速度分量,/>表示高超声速飞行器在z轴上的速度分量;Among them, V h (t) represents the velocity vector of the hypersonic aircraft, γ h represents the ballistic deflection angle of the hypersonic aircraft, τ h represents the first-order link time constant of the hypersonic aircraft, n zhc represents the overload command of the hypersonic aircraft, n zh represents the actual flight overload of the hypersonic vehicle, Represents the velocity component of the hypersonic aircraft on the x-axis,/> Represents the velocity component of the hypersonic aircraft on the z-axis;
高超声速飞行器所受控制约束为:The control constraints of a hypersonic aircraft are:
nzh≤nzhmax n zh ≤ n zhmax
根据已有仿真结果,近似逆轨拦截态势的角度范围设定为±3.2°,即:According to the existing simulation results, the angle range of the approximate reverse orbit interception situation is set to ±3.2°, that is:
|γh0-γm0+π|<3.2°|γ h0 -γ m0 +π|<3.2°
式中,nzhmax表示高超声速飞行器的最大可用过载;γh0,γm0分别为初始时刻高超声速飞行器和拦截弹的弹道偏角。In the formula, n zhmax represents the maximum available overload of the hypersonic aircraft; γ h0 and γ m0 are the ballistic deflection angles of the hypersonic aircraft and the interceptor at the initial moment, respectively.
优选地,所述步骤3具体如下:Preferably, the step 3 is as follows:
步骤3-1:以基准参数为初始条件、不进行参数拉偏情况下仿真:起始突防时刻取固定值,获取飞行器博弈过程中的状态参数变化,以获取后续双方初始距离、拦截弹过载、双方发射角度作为初始拉偏量参考量;Step 3-1: Simulate with the baseline parameters as the initial conditions and without parameter deviation: the initial penetration moment is set to a fixed value, and the state parameter changes during the game of the aircraft are obtained to obtain the initial distance between the two parties and the overload of the interceptor missile. , the launch angles of both sides are used as the reference for the initial deviation amount;
步骤3-2:以步骤3-1中的结论以及角度限制为初始条件,进行多个参数随机拉偏情况下仿真:通过对不同初始条件下多次机动突防仿真,构建攻防对抗数据库,遍历初始态势信息以及拦截方可能采取的机动策略,为后续通过神经网络对数据库进行离线学习提供训练、验证数据。Step 3-2: Taking the conclusion and angle limit in step 3-1 as the initial conditions, conduct a simulation under the condition of random deviation of multiple parameters: build an offensive and defensive confrontation database by simulating multiple maneuver penetrations under different initial conditions, and traverse The initial situation information and possible maneuvering strategies of the interceptor provide training and verification data for subsequent offline learning of the database through neural networks.
优选地,所述步骤4所述的BP神经网络,设有两层隐藏层并选择tansig激活函数,神经网络的具体参数设置如下表所示:Preferably, the BP neural network described in step 4 is provided with two hidden layers and a tansig activation function is selected. The specific parameter settings of the neural network are as shown in the following table:
表1神经网络参数设置Table 1 Neural network parameter settings
本发明的有益效果如下:The beneficial effects of the present invention are as follows:
本发明提出了一种基于主动博弈机动在真实攻防对抗场景中息获取延迟及不准确情况下高超声速飞行器突防策略的设计方法。具有两类优势。一是相较于最优控制和微分对策,己方机动策略主要依赖自身近距预警能力及机动能力,不依赖于对拦截弹制导律的假设,不确定性小,易于工程实现。二是设计的突防策略自身机动指令形式简单,无需复杂计算,没有实时性等障碍,易于工程实现。The present invention proposes a design method for a hypersonic aircraft penetration strategy based on active game maneuvers in real attack and defense confrontation scenarios where information acquisition is delayed and inaccurate. There are two types of advantages. First, compared with optimal control and differential countermeasures, one's own maneuver strategy mainly relies on its own short-range early warning capability and maneuverability, and does not rely on assumptions about the guidance law of interceptor missiles. It has low uncertainty and is easy to implement in engineering. Second, the designed penetration strategy has a simple form of maneuver instructions, does not require complex calculations, has no obstacles such as real-time performance, and is easy to implement in engineering.
附图说明Description of drawings
图1为本发明高超声速飞行器突防窗口计算流程图。Figure 1 is a flow chart for calculating the penetration window of a hypersonic aircraft according to the present invention.
图2为本发明高超声速飞行器突防窗口概念示意图,其中(a)高超声速飞行器在“突防窗口”内机动,(b)高超声速飞行器在“突防窗口”外过早机动,(c)高超声速飞行器在“突防窗口”外过晚机动。Figure 2 is a conceptual diagram of the penetration window of the hypersonic aircraft of the present invention, in which (a) the hypersonic aircraft maneuvers within the "penetration window", (b) the hypersonic aircraft maneuvers prematurely outside the "penetration window", (c) The hypersonic vehicle maneuvered too late outside the "penetration window".
图3本发明近似逆轨拦截态势示意图。Figure 3 is a schematic diagram of the approximate reverse orbit interception situation of the present invention.
图4本发明横侧向平面内近似逆轨拦截态势下的攻防对抗示意图。Figure 4 is a schematic diagram of the offensive and defensive confrontation in the approximate reverse orbit interception situation in the lateral plane of the present invention.
图5为本发明实施例不拉偏情况下“突防窗口”蒙特卡洛仿真示意图,(a)拦截弹制导律为比例导引律,(b)拦截弹制导律为修正比例导引律,(c)拦截弹制导律为自适应滑模导引律。Figure 5 is a schematic diagram of the Monte Carlo simulation of the "penetration window" without deflection according to the embodiment of the present invention. (a) The guidance law of the interceptor missile is a proportional guidance law, (b) the guidance law of the interceptor missile is a modified proportional guidance law. (c) The interceptor missile guidance law is an adaptive sliding mode guidance law.
图6为本发明实施例随机拉偏情况下“突防窗口”蒙特卡洛仿真示意图,(a)拦截弹制导律为比例导引律,(b)拦截弹制导律为修正比例导引律,(c)拦截弹制导律为自适应滑模导引律。Figure 6 is a schematic diagram of the Monte Carlo simulation of the "penetration window" under the condition of random deflection according to the embodiment of the present invention. (a) The guidance law of the interceptor missile is a proportional guidance law, (b) the guidance law of the interceptor missile is a modified proportional guidance law. (c) The interceptor missile guidance law is an adaptive sliding mode guidance law.
图7为本发明实施例当训练目标最小误差为0.001时“突防窗口”神经网络拟合效果,(a)拦截弹制导律为比例导引律,(b)拦截弹制导律为修正比例导引律,(c)拦截弹制导律为自适应滑模导引律。Figure 7 shows the fitting effect of the "penetration window" neural network when the minimum error of the training target is 0.001 according to the embodiment of the present invention. (a) The guidance law of the interceptor missile is a proportional guidance law, (b) the guidance law of the interceptor missile is a modified proportional guidance law. The guidance law, (c) the interceptor missile guidance law is an adaptive sliding mode guidance law.
具体实施方式Detailed ways
下面结合附图和实施例对本发明进一步说明。The present invention will be further described below in conjunction with the accompanying drawings and examples.
为了解决现有基于经典方法和智能方法设计的突防策略在针对真实攻防对抗场景中鲁棒性差的问题,结合神经网络所具有的不依赖显式数学模型、可以拟合复杂非线性映射关系的优势,通过神经网络拟合的高超声速飞行器满过载机动区间,不仅可以完成对敌方拦截弹的成功突防,同时可以加强突防策略对信息延迟及准确性差的鲁棒性。In order to solve the problem of poor robustness of existing penetration strategies designed based on classical methods and intelligent methods in real attack and defense confrontation scenarios, combined with the neural network, which does not rely on explicit mathematical models and can fit complex nonlinear mapping relationships, Advantages: The full overload maneuvering range of the hypersonic aircraft fitted through the neural network can not only complete the successful penetration of enemy interceptors, but also enhance the robustness of the penetration strategy to information delays and poor accuracy.
本发明的技术方案为:The technical solution of the present invention is:
步骤1,基于高超声速飞行器在近距对抗中的优劣势分析,提出“突防窗口”的概念。Step 1. Based on the analysis of the advantages and disadvantages of hypersonic aircraft in close confrontation, the concept of "penetration window" is proposed.
如图2所示,“突防窗口”是指高超声速飞行器可通过满过载机动实现对拦截导弹的突防的相对距离区间。当拦截弹已对可机动变轨飞行器形成近似逆轨拦截态势时,突防过程中可机动变轨飞行器的速度优势无法充分发挥,能利用主要优势是作为攻方的机动主动性和成功突防所需的脱靶量较小,而劣势是过载能力远远弱于拦截弹。因此,当机动过早时,会因过载劣势的暴露时间更长导致被拦截;反之机动过晚时,无法在侧向达成所需的脱靶量。当且仅当在“突防窗口”内进行突防机动时,可充分利用攻方机动主动性,结合双方飞快的接近速度,在短时间内创造出突防所需的1m以上脱靶量,同时又尽可能地回避了己方的过载劣势,在小范围内完成对拦截弹的有效突防。As shown in Figure 2, the "penetration window" refers to the relative distance range in which a hypersonic aircraft can achieve penetration of an interceptor missile through full overload maneuvers. When the interceptor missile has formed a similar reverse-orbit interception situation against the maneuverable orbit-changing aircraft, the speed advantage of the maneuverable orbit-changing aircraft cannot be fully utilized during the penetration process. The main advantage that can be utilized is the maneuvering initiative of the attacker and the successful penetration. The required miss amount is smaller, and the disadvantage is that the overload capability is much weaker than that of interceptors. Therefore, when the maneuver is too early, it will be intercepted due to the longer exposure time of the overload disadvantage; conversely, when the maneuver is too late, the required miss amount cannot be achieved laterally. When and only when the penetration maneuver is performed within the "penetration window", the attacker's maneuvering initiative can be fully utilized, combined with the rapid approach speed of both parties, to create a miss distance of more than 1m required for penetration in a short period of time. At the same time, It also avoided its own overload disadvantage as much as possible and completed effective penetration of interceptor missiles in a small area.
步骤2,在横侧向平面内建立近似逆轨拦截态势下的攻防对抗数学模型。Step 2: Establish a mathematical model of offensive and defensive confrontation in an approximate reverse orbit interception situation in the lateral plane.
高超声速飞行器的机动飞行按机动进行的方向可以分为纵向和横侧向机动,两种机动可以单独存在或同时存在。本发明优先选择通过横侧向的机动完成突防,原因在于:横侧向机动突防可在飞行器定高定速的情况下进行,避免了突防过程中由于速度和高度的变化,对于控制器的影响。The maneuvering flight of a hypersonic aircraft can be divided into longitudinal and lateral maneuvers according to the direction of maneuvering. The two maneuvers can exist alone or at the same time. The present invention preferentially chooses to complete the penetration through lateral maneuvers. The reason is that the lateral maneuvers can be carried out when the aircraft is at a fixed altitude and speed, which avoids the need for control due to changes in speed and altitude during the penetration process. influence of the device.
如图3所示,近似逆轨拦截态势是指特定的角度范围,当高超声速飞行器和拦截弹双方初始速度指向偏差大于该角度范围时,高超声速飞行器无需进行机动突防,利用自身速度即可完成突防;当双方初始速度指向偏差小于该角度范围时,则需要进行机动突防。As shown in Figure 3, the approximate reverse-orbit interception situation refers to a specific angular range. When the initial speed pointing deviation of both the hypersonic aircraft and the interceptor missile is greater than this angular range, the hypersonic aircraft does not need to maneuver for penetration and can use its own speed. Complete the penetration; when the initial speed direction deviation of both parties is less than this angle range, a maneuverable penetration is required.
步骤3,基于“突防窗口”的概念,利用蒙特卡洛仿真尽可能遍历工程应用中不同的作战态势、敌我机动策略、机动能力等,生成博弈对抗数据库。Step 3: Based on the concept of "penetration window", Monte Carlo simulation is used to traverse different combat situations, enemy and friendly maneuver strategies, maneuver capabilities, etc. in engineering applications as much as possible to generate a game confrontation database.
步骤4,基于BP神经网络对数据库进行离线学习。Step 4: Perform offline learning on the database based on BP neural network.
在新的攻防对抗场景中,将态势参数,飞行器能力参数,敌我机动策略等作为输入,其中未知参数可在边界范围内选取多组数值,生成多个窗口,离线训练会得到可靠的结果,在线输出突防机动窗口(多个窗口的公共区间),可机动变轨飞行器可凭借在该窗口内的自定义机动策略完成对拦截弹的小范围高效突防。“突防窗口”计算流程如附图1所示。In the new offensive and defensive confrontation scenario, situational parameters, aircraft capability parameters, enemy and friendly maneuver strategies, etc. are used as inputs. The unknown parameters can select multiple sets of values within the boundary range to generate multiple windows. Offline training will obtain reliable results. Online Output the penetration maneuver window (the common area of multiple windows), and the maneuverable orbit-changing aircraft can complete small-scale and efficient penetration of interceptors by relying on the customized maneuver strategy within this window. The “penetration window” calculation process is shown in Figure 1.
实施例:Example:
步骤1,在横侧向平面内建立近似逆轨拦截态势下的攻防对抗数学模型。Step 1: Establish a mathematical model of offensive and defensive confrontation in an approximate reverse orbit interception situation in the lateral plane.
当存在单枚拦截弹时,横侧向平面的攻防对抗示意图及相关角度定义附图4所示。其中H为高超声速飞行器,M为拦截弹。在图4中,不失一般性,以H-M的初始视线方向作为X轴,Z轴在定高平面内与X轴垂直。图中相关的符号定义如表1所示。When there is a single interceptor missile, the schematic diagram of the offensive and defensive confrontation in the lateral plane and the definition of relevant angles are shown in Figure 4. Among them, H is the hypersonic aircraft and M is the interceptor missile. In Figure 4, without loss of generality, the initial line of sight direction of H-M is taken as the X-axis, and the Z-axis is perpendicular to the X-axis in the fixed height plane. The relevant symbol definitions in the figure are shown in Table 1.
表1攻防双方飞行器参数名称对照Table 1 Comparison of parameter names of aircraft on both sides of attack and defense
表中:i=h,m。h代表高超声速飞行器,m代表拦截弹。视线方位角λhm定义为从拦截弹指向高超声速飞行器,图4中λhm<0。In the table: i=h,m. h represents the hypersonic vehicle and m represents the interceptor missile. The line of sight azimuth angle λ hm is defined as from the interceptor missile to the hypersonic aircraft. In Figure 4, λ hm <0.
结合附图4所示二维平面攻防对抗形式,由此可以构建出拦截弹在末制导段的数学描述如下:Combined with the two-dimensional planar offensive and defensive confrontation form shown in Figure 4, the mathematical description of the interceptor missile in the terminal guidance section can be constructed as follows:
式中,为工程实际中典型的拦截弹制导律,包括比例导引律(ProportionalNavigation,PN)、修正比例导引律(Augmented Proportional Navigation,APN)和自适应滑模导引律(Adaptive Sliding Mode Guidance,ASMG)等。在横侧向平面内这3种拦截弹导引律可表示如下:In the formula, It is a typical interceptor missile guidance law in engineering practice, including Proportional Navigation (PN), Augmented Proportional Navigation (APN) and Adaptive Sliding Mode Guidance (ASMG). wait. These three interceptor missile guidance laws in the lateral plane can be expressed as follows:
比例导引律:Proportional guiding law:
修正比例导引律:Modified proportional guidance law:
自适应滑模导引律:Adaptive sliding mode guidance law:
式中:为从拦截弹视角出发的视线角速度;Vc为攻防双方接近速度;N,ε和δ均为制导律参数。In the formula: is the line-of-sight angular velocity from the perspective of the interceptor; V c is the approaching speed of both attacker and defender; N, ε and δ are all guidance law parameters.
同时,拦截弹在整个拦截过程采用STT控制方式,其所受约束仅为最大可用At the same time, the interceptor uses the STT control method during the entire interception process, and its constraints are only the maximum available
过载约束,表示为:Overload constraints, expressed as:
|nzm|≤nzmmax |n zm |≤n zmmax
对于高超声速飞行器,其具体描述如下:For hypersonic aircraft, its specific description is as follows:
高超声速飞行器所受控制约束为:The control constraints of a hypersonic aircraft are:
nzh≤nzhmax n zh ≤ n zhmax
根据已有仿真结果,近似逆轨拦截态势的角度范围设定为±3.2°,即:According to the existing simulation results, the angle range of the approximate reverse orbit interception situation is set to ±3.2°, that is:
|γh0-γm0+π|<3.2°|γ h0 -γ m0 +π|<3.2°
式中,γh0,γm0分别为初始时刻,高超声速飞行器和拦截弹的弹道偏角。In the formula, γ h0 and γ m0 are the ballistic deflection angles of the hypersonic aircraft and the interceptor missile at the initial moment, respectively.
步骤2,基于“突防窗口”的概念,利用蒙特卡洛仿真尽可能遍历工程应用中不同的作战态势、敌我机动策略、机动能力等,生成博弈对抗数据库。其中包含以下子步骤。Step 2: Based on the concept of "penetration window", Monte Carlo simulation is used to traverse as many different combat situations, enemy and friendly maneuver strategies, maneuver capabilities, etc. as possible in engineering applications to generate a game confrontation database. It contains the following sub-steps.
步骤2.1,以基准参数为初始条件、不进行参数拉偏情况下仿真:起始突防时刻取固定值,获取飞行器博弈过程中的状态参数变化,以获取后续双方初始距离、拦截弹过载、双方发射角度等初始拉偏量参考量。Step 2.1, use the benchmark parameters as the initial conditions and simulate without parameter deviation: the initial penetration moment is set to a fixed value, and the state parameter changes during the aircraft game are obtained to obtain the subsequent initial distance between the two parties, interceptor overload, and both sides. Reference quantity of initial pull deviation such as launch angle.
表2基于“突防窗口”概念的机动突防方法仿真相关参数设置Table 2 Parameter settings related to simulation of mobile penetration method based on the concept of "penetration window"
当拦截弹分别采用比例导引律、修正比例导引律以及自适应滑模导引律时,高超声速飞行器与拦截弹的攻防对抗情况如附图5所示。When the interceptor missile adopts proportional guidance law, modified proportional guidance law and adaptive sliding mode guidance law respectively, the offensive and defensive confrontation between the hypersonic aircraft and the interceptor missile is shown in Figure 5.
步骤2.2,以步骤2.1中结论以及角度限制为初始条件、进行多个参数随机拉偏情况下仿真:通过对不同初始条件下多次机动突防仿真,构建攻防对抗数据库,遍历初始态势信息以及拦截方可能采取的机动策略,为后续通过神经网络对数据库进行离线学习提供训练、验证数据。Step 2.2, use the conclusion and angle limit in step 2.1 as the initial conditions, and conduct a simulation with multiple parameters randomly biased: by simulating multiple maneuver penetrations under different initial conditions, build an offensive and defensive confrontation database, traverse the initial situation information and interception The possible maneuvering strategies can be adopted to provide training and verification data for subsequent offline learning of the database through neural networks.
表3相关拉偏参数设置Table 3 Relevant bias parameter settings
蒙特卡洛仿真结果如附图6所示。The Monte Carlo simulation results are shown in Figure 6.
步骤3,利用步骤2.2中随机拉偏参数得到的500组“突防窗口”数据作为训练集,利用BP神经网络进行拟合,网络设两层隐藏层并选择tansig激活函数。多次训练调整参数后保留结果最好的网络用于后续测试。Step 3: Use the 500 sets of "penetration window" data obtained by randomly pulling the bias parameters in step 2.2 as the training set, and use the BP neural network for fitting. The network has two hidden layers and selects the tansig activation function. After multiple trainings and adjusting parameters, the network with the best results is retained for subsequent testing.
表4神经网络参数设置Table 4 Neural network parameter settings
“突防窗口”神经网络拟合结果如附图7所示。The fitting results of the "penetration window" neural network are shown in Figure 7.
最终,训练好的神经网络可用于预测给定初始条件时高超声速飞行器机动“突防窗口”情况。假设敌方拦截弹初始状态如下:初始相对距离R0=11km,弹道偏角γm0=181°,最大可用过载ummax=7g;Ultimately, the trained neural network can be used to predict the "penetration window" of hypersonic vehicle maneuvers given initial conditions. Assume that the initial state of the enemy interceptor is as follows: initial relative distance R0 = 11km, ballistic deflection angle γ m0 = 181°, maximum available overload u mmax = 7g;
1)当敌方拦截弹制导律采用比例导引律时,高超声速飞行器实际“突防窗口”上界为10km,下界为1.3km,采用最终训练好的BP神经网络进行窗口预测,预测“突防窗口”上界为9.99km,下界为1.33km,误差分别为0.01与0.03。1) When the enemy interceptor missile guidance law adopts the proportional guidance law, the upper bound of the actual "penetration window" of the hypersonic aircraft is 10km and the lower bound is 1.3km. The final trained BP neural network is used for window prediction to predict the "penetration window". The upper bound of "anti-window" is 9.99km, the lower bound is 1.33km, and the errors are 0.01 and 0.03 respectively.
2)当敌方拦截弹制导律采用修正比例导引律时,高超声速飞行器实际“突防窗口”上界为5.3km,下界为1..4km,采用最终训练好的BP神经网络进行窗口预测,预测“突防窗口”上界为5.215km,“突防窗口”下界为1.422km,误差分别为0.085与0.022。2) When the enemy interceptor missile guidance law adopts the modified proportional guidance law, the upper bound of the actual "penetration window" of the hypersonic aircraft is 5.3km and the lower bound is 1..4km. The final trained BP neural network is used for window prediction. , the upper bound of the "penetration window" is predicted to be 5.215km, and the lower bound of the "penetration window" is 1.422km, with errors of 0.085 and 0.022 respectively.
3)当敌方拦截弹制导律采用自适应滑模导引律时,高超声速飞行器实际“突防窗口”上界为8.1km,下界为1.4km,采用最终训练好的BP神经网络进行窗口预测,预测“突防窗口”上界为8.175km,“突防窗口”下界为1.475km,误差分别为0.075与0.075。3) When the enemy interceptor missile guidance law adopts the adaptive sliding mode guidance law, the upper bound of the actual "penetration window" of the hypersonic aircraft is 8.1km and the lower bound is 1.4km. The final trained BP neural network is used for window prediction. , the upper bound of the "penetration window" is predicted to be 8.175km, and the lower bound of the "penetration window" is 1.475km, with errors of 0.075 and 0.075 respectively.
结合上述针对敌方拦截弹采用不同制导律计算出的三个“突防窗口”,最终求得其公共“突防窗口”为△R=[1.475,5.215],高超声速飞行器在此区间内进行满过载机动,即可对敌方拦截弹成功突防。Combining the above three "penetration windows" calculated using different guidance laws for enemy interceptors, the final public "penetration window" is △R=[1.475,5.215], and the hypersonic aircraft operates within this interval With full overload maneuver, you can successfully penetrate enemy interceptors.
结论:本发明基于高超声速飞行器在近距对抗中的优劣势分析,提出“突防窗口”的作战概念;构建攻防对抗数据库,遍历初始态势信息以及拦截方可能采取的机动策略,通过BP神经网络对数据库进行离线学习,在线计算生成“突防窗口”指导高超声速飞行器完成对不同场景的突防。在线计算过程中,针对不确定的敌方信息和态势信息,可结合数据库将该输入项取极值或多个取值,最终在所生成的多个窗口中寻找公共区间,即为公共“突防窗口”。公共“突防窗口”的设计理念,避免了对不确定信息的主观估计,使所设计的突防策略具有普适性和鲁棒性。Conclusion: Based on the analysis of the advantages and disadvantages of hypersonic aircraft in close confrontation, this invention proposes the operational concept of "penetration window"; builds an offensive and defensive confrontation database, traverses the initial situation information and possible maneuvering strategies of the interceptor, and uses BP neural network to The database is studied offline, and the online calculation generates a "penetration window" to guide the hypersonic aircraft to complete penetration in different scenarios. During the online calculation process, for uncertain enemy information and situational information, the input item can be combined with the database to take the extreme value or multiple values, and finally find a common interval in the multiple generated windows, which is the common "sudden moment". Anti-window". The design concept of the public "penetration window" avoids subjective estimation of uncertain information, making the designed penetration strategy universal and robust.
Claims (4)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311102259.2A CN117171877A (en) | 2023-08-30 | 2023-08-30 | Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311102259.2A CN117171877A (en) | 2023-08-30 | 2023-08-30 | Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117171877A true CN117171877A (en) | 2023-12-05 |
Family
ID=88932909
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311102259.2A Pending CN117171877A (en) | 2023-08-30 | 2023-08-30 | Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117171877A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117852309A (en) * | 2024-03-06 | 2024-04-09 | 西北工业大学 | A penetration effectiveness evaluation method based on index hierarchy |
CN117852415A (en) * | 2024-03-07 | 2024-04-09 | 西北工业大学 | Method and device for solving maneuvering space of ultra-high-speed aircraft based on variable stepping withdrawal method |
CN118862315A (en) * | 2024-09-24 | 2024-10-29 | 西北工业大学 | A method of aircraft penetration based on deep reinforcement learning |
-
2023
- 2023-08-30 CN CN202311102259.2A patent/CN117171877A/en active Pending
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117852309A (en) * | 2024-03-06 | 2024-04-09 | 西北工业大学 | A penetration effectiveness evaluation method based on index hierarchy |
CN117852309B (en) * | 2024-03-06 | 2024-05-24 | 西北工业大学 | A penetration effectiveness evaluation method based on index hierarchy |
CN117852415A (en) * | 2024-03-07 | 2024-04-09 | 西北工业大学 | Method and device for solving maneuvering space of ultra-high-speed aircraft based on variable stepping withdrawal method |
CN117852415B (en) * | 2024-03-07 | 2024-05-24 | 西北工业大学 | Method and device for solving maneuvering space of ultra-high-speed aircraft based on variable stepping withdrawal method |
CN118862315A (en) * | 2024-09-24 | 2024-10-29 | 西北工业大学 | A method of aircraft penetration based on deep reinforcement learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117171877A (en) | Design method of maneuvering penetration strategy for hypersonic aircraft based on timing game | |
Garcia et al. | Cooperative strategies for optimal aircraft defense from an attacking missile | |
CN108362171B (en) | A kind of Guidance constrained with attack time and angle-of-attack | |
Prokopov et al. | Linear quadratic optimal cooperative strategies for active aircraft protection | |
CN112859921B (en) | Three-dimensional simultaneous attack robust cooperative guidance law design method | |
CN105716470B (en) | A kind of differential game is counter to intercept Maneuver Penetration/precision strike guidance method | |
Fonod et al. | Estimation enhancement by cooperatively imposing relative intercept angles | |
CN107818219A (en) | A multi-missile cooperative trajectory planning method for defense penetration | |
CN111142382B (en) | Maneuvering control method, device, equipment and storage medium of anti-interceptor missile | |
Weintraub et al. | Optimal guidance strategy for the defense of a non‐manoeuvrable target in 3‐dimensions | |
CN114722701A (en) | Method for obtaining war and chess deduction cooperation strategy based on deep reinforcement learning model | |
Xu et al. | Application of situation function in air combat differential games | |
CN106054612A (en) | BTT missile flight trajectory automatic control method | |
CN115145295A (en) | Online autonomous flight path optimization control method for unmanned aerial vehicle in dynamic environment | |
Li et al. | Three‐Dimensional Impact Time and Angle Control Guidance Based on MPSP | |
CN113139331A (en) | Air-to-air missile situation perception and decision method based on Bayesian network | |
CN118331058A (en) | Advanced reinforcement learning-based intelligent sudden-defense maneuver decision-making method for hypersonic vehicle | |
Xu et al. | Virtual motion camouflage based phantom track generation through cooperative electronic combat air vehicles | |
Shi et al. | Predictive guidance strategies for active aircraft defense | |
CN116774714B (en) | Multi-constraint collaborative guidance method based on event triggering mechanism | |
CN115357051B (en) | Deformation and maneuvering integrated avoidance and defense method | |
Shuyang et al. | Cooperative guidance law with maneuverability awareness: A decentralized solution | |
CN116360500A (en) | Missile burst prevention method capable of getting rid of controllable distance | |
Li et al. | Generation method of autonomous evasive maneuver strategy in air combat | |
Wang et al. | Research on naval air defense intelligent operations on deep reinforcement learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |