Target dynamic positioning three-dimensional guidance method based on visual identification
By combining visual recognition and sliding mode control, a multi-constraint guidance law is realized without the need for remaining flight time estimation, which solves the problem of guidance accuracy being affected by uncertain factors and improves the aircraft's hit accuracy and system stability in uncertain environments.
Patent Information
- Application Number
- CN202410351695.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-26
- Publication Date
- 2025-09-26
AI Technical Summary
The existing guidance law based on the optimal control principle is affected by multiple uncertain factors in actual flight, resulting in reduced guidance accuracy. It is also sensitive to the estimation of the remaining flight time. When the error is large, it will lead to a decrease in the missile's hit accuracy.
A three-dimensional guidance method for target dynamic positioning based on visual recognition is adopted. By setting a multi-constraint guidance law that does not rely on the remaining flight time estimation, combined with image recognition and sliding mode control, it is transformed into a travel distance control and target view tracking problem. Flow-Guided Feature Aggregation is used to obtain target motion information, and the sliding mode surface is used to control the aircraft acceleration command to achieve target hit.
The control system design is simplified, the system convenience and robustness are improved, the dependence on the accuracy of the remaining flight time is reduced, and the aircraft is ensured to be stably controlled and accurately hit the target in uncertain environments.
Smart Images
Figure CN120704346A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a target dynamic positioning three-dimensional guidance method based on visual recognition, and belongs to the technical field of aircraft control. Background Art
[0002] The most widely used guidance law at present is the optimal guidance law based on the optimal control principle. Although under ideal circumstances, the optimal guidance law can obtain the best trajectory, in actual flight, due to the influence of various uncertain factors, it will have a great impact on the flight performance of the missile.
[0003] Furthermore, the optimal guidance law is highly sensitive to estimates of target acceleration and flight time, placing high demands on the observation components. Large errors in the remaining flight time estimate can significantly reduce guidance accuracy.
[0004] Therefore, it is necessary to further study the existing guidance law to solve the above problems. Summary of the Invention
[0005] In order to overcome the above problems, the inventors have conducted in-depth research and proposed a three-dimensional guidance method for dynamic positioning of a target based on visual recognition, which includes the following steps:
[0006] S1. Set up a multi-constraint guidance law with range control and no need for remaining flight time estimation;
[0007] S2. Obtain target motion information based on the video captured by the aircraft using image recognition methods;
[0008] S3. Based on the obtained target motion information and the guidance law, an acceleration command is obtained by using a sliding mode control method;
[0009] S4. Use acceleration instructions to control the aircraft to achieve the aircraft's attack on the target.
[0010] In a preferred embodiment, in S1, the guidance law is set to:
[0011]
[0012] Among them, e r represents the remaining distance error, σ c is the desired angle command, V M Indicates the speed of the aircraft.
[0013] In a preferred embodiment, the desired angle command σ c Set to:
[0014]
[0015] Among them, k r is a coefficient greater than 0.
[0016] In a preferred embodiment, the desired angle command σ c Constraints are made to obtain the angular velocity command of the viewing angle Expressed as:
[0017]
[0018] Among them, σ0 is the non-zero initial viewing angle, σ max The maximum viewing angle of the aircraft.
[0019] In a preferred embodiment, in S2, target motion information is obtained based on a Flow-Guided Feature Aggregation method.
[0020] In a preferred embodiment, in S3, the sliding surface is set to:
[0021] s=σ-σ c
[0022]
[0023] Among them, s is the sliding surface, a M is the acceleration of the aircraft, and σ is the longitudinal viewing angle.
[0024] In a preferred embodiment, the obtained aircraft lateral acceleration instruction a mz for:
[0025]
[0026]
[0027]
[0028] in, is an equivalent controller, is a discontinuous controller, and M is the control gain.
[0029] In a preferred embodiment, the aircraft longitudinal acceleration instruction a is obtained. my for:
[0030]
[0031] Where ξ is the lateral viewing angle, ξ0 is the non-zero initial lateral viewing angle
[0032] The beneficial effects of the present invention include:
[0033] (1) Traditional time control problems usually require accurate estimation of the remaining flight time. The present invention adopts a multi-constraint guidance law that does not require remaining flight time estimation, which simplifies the design and implementation of the control system and reduces the dependence on the accuracy of the remaining flight time, thereby improving the convenience and practicality of the system.
[0034] (2) The use of a multi-constraint guidance law that does not require remaining flight time estimation can reduce the dependence on the accuracy of the remaining flight time and improve the robustness of the system. Even in the face of uncertain flight environments or changes in target motion, the guidance system can still stably control the aircraft and achieve target hits. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 A schematic flow chart of a target dynamic positioning and three-dimensional guidance method based on visual recognition according to a preferred embodiment of the present invention is shown;
[0036] Figure 2 A simulation diagram of the trajectory inclination angle of the aircraft in Example 1 is shown;
[0037] Figure 3 A simulation diagram of the trajectory inclination angle of the aircraft in Example 1 is shown;
[0038] Figure 4 The longitudinal overload simulation curve of the aircraft in Example 1 is shown;
[0039] Figure 5 The aircraft yaw overload simulation curve in Example 1 is shown;
[0040] Figure 6 The motion trajectory diagram of the aircraft and the target in Example 1 is shown. DETAILED DESCRIPTION
[0041] The present invention will be described in further detail below with reference to the accompanying drawings and examples, through which the features and advantages of the present invention will become more clearly understood.
[0042] The word "exemplary" is used exclusively herein to mean "serving as an example, example, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.
[0043] According to the present invention, a three-dimensional guidance method for dynamic positioning of a target based on visual recognition is provided. Figure 1 As shown, the following steps are included:
[0044] S1. Set up a multi-constraint guidance law with range control and no need for remaining flight time estimation;
[0045] S2. Obtain target motion information based on the video captured by the aircraft using image recognition methods;
[0046] S3. Based on the obtained target motion information and the guidance law, an acceleration command is obtained by using a sliding mode control method;
[0047] S4. Use acceleration instructions to control the aircraft to achieve the aircraft's attack on the target.
[0048] Traditional guidance laws mostly use the remaining flight time estimation as the control basis, which requires a high degree of accuracy in the estimation of the remaining flight time. However, due to the influence of various uncertain factors, the estimation of the remaining flight time in some scenarios will have large errors, resulting in a significant reduction in guidance accuracy.
[0049] In S1, the guidance law no longer uses the traditional remaining flight time control, but instead uses the aircraft travel distance control.
[0050] Specifically, the remaining flight time control is converted into travel distance control, which is expressed as:
[0051] R c =V M (t d -t)
[0052] Set the remaining distance error:
[0053] e r =RR c
[0054] By differentiating the remaining range error, the guidance law is obtained:
[0055]
[0056] Among them, e r Indicates the remaining distance error, R indicates the remaining flight distance, R c represents the expected remaining flight distance, σ c is the desired angle command, V M Indicates the speed of the aircraft.
[0057] That is, the guidance law is set as:
[0058]
[0059] Among them, e r represents the remaining distance error, σ c is the desired angle command, V M Indicates the speed of the aircraft.
[0060] Preferably, the desired angle command σ cSet to:
[0061]
[0062] Among them, k r is a coefficient greater than 0.
[0063] According to the guidance law and the desired angle command setting, we can get:
[0064] e r (t) = e r0 exp(-k r t)
[0065] Among them, e r0 is the initial residual distance error. Obviously, e r0 is always less than zero, because the nominal forward distance should be greater than or equal to the missile's initial distance, indicating that the error system is asymptotically stable. If the true viewpoint angle can follow the target viewpoint angle command, the missile-target distance will converge to the nominal forward distance, and the flight time will equal the target impact time. Based on this, it can be proved that the above guidance law and the desired angle command setting can transform the remaining flight time control problem into a travel distance control problem, and then into a target viewpoint tracking problem.
[0066] Preferably, the desired angle command σ c Constraints are made to obtain the angular velocity command of the viewing angle Realize the target viewing angle tracking. The viewing angle angular velocity instruction Set to:
[0067]
[0068] Among them, σ0 is the non-zero initial viewing angle, σ max The maximum viewing angle of the aircraft.
[0069] In S2, the target motion information is obtained based on the Flow-Guided Feature Aggregation method.
[0070] The Flow-Guided Feature Aggregation method can be found in the literature Zhu X, Wang Y, Dai J, et al. Flow-Guided Feature Aggregation for Video Object Detection[J]. IEEE, 2017. DOI: 10.1109 / ICCV.2017.52, which is not described in detail in this invention.
[0071] Furthermore, in S2, the target motion information obtained includes: the target's speed, the target's trajectory inclination, and the target's trajectory deviation.
[0072] By using the Flow-Guided Feature Aggregation method based on image recognition, the target's motion information can be accurately acquired and used in the guidance system. This efficient method can improve the guidance accuracy and effectiveness, enabling the system to track and hit moving targets more quickly.
[0073] In S3, the sliding surface is set to:
[0074] s=σ-σ c
[0075]
[0076] Among them, s is the sliding surface, a M is the acceleration of the aircraft, and σ is the longitudinal viewing angle.
[0077] Preferably, the aircraft lateral acceleration command a is obtained mz for:
[0078]
[0079]
[0080]
[0081] in, is an equivalent controller, is a discontinuous controller, M is the control gain, and by increasing the control modulus M, the time interval to reach the sliding surface is reduced.
[0082] The aircraft lateral acceleration provided by the present invention can force sliding on the sliding surface. Once the sliding surface is reached, the system will remain on the sliding surface due to the equivalent control terms.
[0083] According to the sliding surface provided by the present invention, considering the Lyapunov function V L Taking the derivative, we can get:
[0084]
[0085] It can be seen that for any M≥0, V L ≤0, ensuring that the system is stable on the sliding surface.
[0086] According to the present invention, the aircraft lateral acceleration command a M Similarly, preferably, the aircraft longitudinal acceleration instruction a is obtainedmy for:
[0087]
[0088] Where ξ is the lateral viewing angle and ξ0 is the non-zero initial lateral viewing angle.
[0089] The combination of the aircraft lateral acceleration command and the aircraft longitudinal acceleration command is the aircraft acceleration command.
[0090] The aircraft is controlled through the aircraft acceleration command to achieve the aircraft's target viewing angle tracking, thereby enabling the aircraft to accurately reach the target position.
[0091] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not a limitation herein.
[0092] Example
[0093] Example 1
[0094] Conducting simulation experiments to guide dynamic targets includes the following steps:
[0095] S1. Set up a multi-constraint guidance law with range control and no need for remaining flight time estimation;
[0096] S2. Set the target motion information to: constant speed of 30 m / s, target trajectory inclination of 5°, and trajectory deviation of 45°;
[0097] S3. Based on the obtained target motion information and the guidance law, an acceleration command is obtained by using a sliding mode control method;
[0098] S4. Use acceleration instructions to control the aircraft to achieve the aircraft's attack on the target.
[0099] In S1, the guidance law is set as:
[0100]
[0101] Desired angle command σ c Set to:
[0102]
[0103] Viewing angle velocity command for:
[0104]
[0105] Among them, k r =3;
[0106] In S3, the sliding surface is set to:
[0107] s=σ-σ c
[0108]
[0109] The obtained aircraft lateral acceleration command a mz for:
[0110]
[0111]
[0112]
[0113] Get the aircraft longitudinal acceleration command a my for:
[0114]
[0115] During the simulation, 6 aircraft (missile 1 to missile 6) are set to attack the target at different positions. The simulation results are as follows: Figure 2-6 As shown, Figure 2 This is a simulation diagram of the aircraft's trajectory inclination angle. Figure 3 This is a simulation diagram of the aircraft's trajectory inclination angle. Figure 4 is the aircraft longitudinal overload simulation curve, Figure 5 is the aircraft yaw overload simulation curve, Figure 6 This is the trajectory diagram of the aircraft and the target.
[0116] from Figure 3-6 It can be seen that the angle changes of different aircraft during flight are small, the pitch overload converges smoothly to near 1 gravity acceleration, the yaw overload converges smoothly to 0, and the trajectory is smooth and flat.
[0117] from Figure 6 It can be seen that the aircraft launched from different positions can all reach the target position, ensuring the accuracy of the guidance method.
[0118] The present invention has been described above with reference to preferred embodiments, but these embodiments are merely exemplary and serve only as illustrations. On this basis, various replacements and improvements can be made to the present invention, all of which fall within the scope of protection of the present invention.
Claims
1. A three-dimensional guidance method for dynamic positioning of a target based on visual recognition, characterized in that: The following steps are involved: S1. Set up a multi-constraint guidance law with range control and no need for remaining flight time estimation; S2. Obtain target motion information based on the video captured by the aircraft using image recognition methods; S3. Based on the obtained target motion information and the guidance law, an acceleration command is obtained by using a sliding mode control method; S4. Use acceleration instructions to control the aircraft to achieve the aircraft's attack on the target.
2. The target dynamic positioning three-dimensional guidance method based on visual recognition according to claim 1 is characterized in that: In S1, the guidance law is set as: Among them, e r represents the remaining distance error, σ c is the desired angle command, V M Indicates the speed of the aircraft.
3. The target dynamic positioning three-dimensional guidance method based on visual recognition according to claim 2 is characterized in that: Desired angle command σ c Set to: Among them, k r is a coefficient greater than 0.
4. The target dynamic positioning three-dimensional guidance method based on visual recognition according to claim 3 is characterized in that: The desired angle command σ c Constraints are made to obtain the angular velocity command of the viewing angle Expressed as: Among them, σ0 is the non-zero initial viewing angle, σ max The maximum viewing angle of the aircraft.
5. The target dynamic positioning three-dimensional guidance method based on visual recognition according to claim 1 is characterized in that: In S2, the target motion information is obtained based on the Flow-Guided Feature Aggregation method.
6. The target dynamic positioning three-dimensional guidance method based on visual recognition according to claim 1 is characterized in that: In S3, the sliding surface is set to: s=σ-σ c Among them, s is the sliding surface, a M is the acceleration of the aircraft, and σ is the longitudinal viewing angle.
7. The method for dynamic positioning and three-dimensional guidance of a target based on visual recognition according to claim 6, characterized in that: The obtained aircraft lateral acceleration command a mz for: in, is an equivalent controller, is a discontinuous controller, and M is the control gain.
8. The method for dynamic positioning and three-dimensional guidance of a target based on visual recognition according to claim 6, characterized in that: Get the aircraft longitudinal acceleration command a my for: Where ξ is the lateral viewing angle and ξ0 is the non-zero initial lateral viewing angle.