Vehicle stabilization system based on active airflow compensation and control method

Abstract

The invention relates to the technical field of vehicle active control, in particular to a vehicle stabilizing system based on active airflow compensation and a control method, and the system comprises a distributed airflow generation array unit, a multi-source sensing system unit, an intelligent control unit and an energy management unit. Directional compensation airflow is generated through the distributed airflow generation array unit, the multi-source sensing system unit monitors a flow field and motion parameters in real time, the intelligent control unit fuses a hybrid control strategy and reinforcement learning to achieve dynamic compensation and algorithm iteration, and the energy management unit recycles braking energy and optimizes distribution. The stability and response speed of the vehicle in a complex environment are remarkably improved, meanwhile, the airflow utilization rate is increased through the gradient porous material, energy consumption is reduced through reverse airflow power generation, and efficient and energy-saving active stable control is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of active vehicle control, and in particular to a vehicle stabilization system and control method based on active airflow compensation. Background Art

[0002] Active airflow compensation refers to a technical approach that achieves dynamic balance by actively generating controllable airflow to offset the effects of external environmental disturbances (such as turbulence, wind and waves) on vehicle stability in real time. In fields such as aviation and navigation, vehicle stability control is a core technology for ensuring safety and energy efficiency. Existing technologies, both passive stabilization schemes and traditional active control techniques have significant limitations: aircraft turbulence mitigation relies primarily on aerodynamic shape optimization and the passive response of flight control systems, resulting in a response delay of 300-500ms; ship roll stabilization devices often use gyro stabilizers or ballast water systems, consuming up to 15%-20% of the ship's power; and the active jet flow control system disclosed in CN1128747A suffers from low airflow utilization (less than 40%).

[0003] Based on this, the present invention provides a vehicle stabilization system and control method based on active airflow compensation to solve the above-mentioned technical problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a vehicle stabilization system and control method based on active airflow compensation. The distributed airflow generation array unit of the present invention generates directional compensation airflow, the multi-source perception system unit monitors the flow field and motion parameters in real time, the intelligent control unit integrates hybrid control strategy and reinforcement learning to achieve dynamic compensation and algorithm iteration, and the energy management unit recovers braking energy and optimizes its distribution, thereby significantly improving the stability and response speed of the vehicle in complex environments. At the same time, the airflow utilization rate is improved by gradient porous materials, and reverse airflow power generation is used to reduce energy consumption, thereby achieving efficient and energy-saving active stability control.

[0005] To achieve the above object, the present invention provides the following technical solutions: The present invention provides a vehicle stabilization system based on active airflow compensation, comprising a distributed airflow generation array unit, a multi-source sensing system unit, an intelligent control unit, and an energy management unit, wherein: The distributed airflow generating array unit is configured to include a rotatable vector nozzle and a three-stage centrifugal air compressor to generate a directional compensating airflow; The multi-source sensing system unit is used to integrate MEMS pressure sensors, laser velocimeters and inertial navigation systems to monitor flow field status and carrier motion parameters in real time; The intelligent control unit is used to adopt an LQR+H∞ hybrid control strategy to establish a predictive controller including a simplified model of the Bernoulli equation and the NS equation; The energy management unit is used to recover energy through reverse airflow power generation technology under braking conditions; The output of the multi-source sensing system unit is connected to the input of the intelligent control unit, and the output of the intelligent control unit drives the distributed airflow generating array unit to form a closed-loop control.

[0006] The distributed airflow generation array unit includes a vector nozzle module, an airflow generation module, and an array layout module, wherein: The vector nozzle module is used to perform efficient directional airflow injection and rapid direction adjustment through the gradient multi-hole vector nozzle and the universal joint structure; The airflow generation module is used to generate and accurately control high-pressure jet airflow by using a three-stage centrifugal air compressor in conjunction with a high-frequency solenoid valve; The array layout module is used to optimize the nozzle distribution according to the characteristics of the vehicle and dynamically adjust the working mode to improve the compensation efficiency and energy efficiency.

[0007] The airflow generation module is used to generate and precisely control high-pressure jet airflow using a three-stage centrifugal air compressor in conjunction with a high-frequency solenoid valve. The specific operation is as follows: A1: The three-stage centrifugal air compressor compresses the air in three stages through the rotation of the impeller, raising the air pressure to 0.8MPa; A2: The high-frequency solenoid valve receives the pulse signal from the intelligent control unit to The high-speed response controls the on / off and flow rate of compressed air. The response time of the high-frequency solenoid valve makes the compensation delay of the system under turbulent conditions less than 80ms. A3: The regulated high-pressure airflow is delivered to the rotatable gradient multi-hole vector nozzle to form a directional compensation airflow.

[0008] The multi-source perception system unit includes a data acquisition module, a data processing module, and an extended perception module, wherein: The data acquisition module is used to obtain flow field pressure, speed and vehicle motion state data in real time using a micro MEMS pressure sensor array and a laser Doppler velocimeter; The data processing module is used to output accurate and reliable perception information through signal preprocessing, deep learning fusion, and verification and correction; The extended perception module is used to monitor the surface temperature distribution of the vehicle with the help of infrared thermal imaging sensors and acoustic sensors, and promptly detect local overheating and noise characteristics caused by abnormal airflow to assist system decision-making.

[0009] The data processing module outputs accurate and reliable perception information through signal preprocessing, deep learning fusion, and verification and correction. The specific operations are as follows: B1: Signal preprocessing uses a sliding median filter with a window width of 50ms, and the output signal-to-noise ratio

[0010] , median filter formula: ,in is the original signal, is the filtered signal, n is the filter window radius; B2: Deep learning fusion achieves multi-source data fusion through a deep neural network with three fully connected layers, with a fusion error of less than 0.1m / s. Deep learning fusion uses the fully connected layer weight update formula: ,in, is the weight of the first layer, is the learning rate, J is the loss function; B3: Data verification: When the Mahalanobis distance between MEMS and laser data is greater than 3σ, correction is triggered. The error judgment formula is used for data verification and correction: ,when The correction is triggered when is the homologous data from different sensors, and m is the number of data samples.

[0011] The micro-MEMS pressure sensor array has a spatial resolution of 5 cm², covering key areas on the surface of vehicles such as wings and hulls, and monitors the air / water pressure distribution in real time with an accuracy of ±0.1 kPa, which is used to identify local flow field disturbances and abnormal pressure fluctuations.

[0012] The intelligent control unit includes a control algorithm module, a strategy generation module, and a learning optimization module, wherein: The control algorithm module is used to use the LQR + H∞ hybrid control strategy and prediction model to deal with interference and respond quickly under complex working conditions.

[0013] The strategy generation module is used to generate a collaborative compensation control strategy by solving the flow field influence matrix and dynamically optimizing the nozzle parameters; The learning optimization module performs autonomous optimization and experience iteration of the control algorithm based on reinforcement learning and interaction with cloud data.

[0014] The learning optimization module performs autonomous optimization and experience iteration of the control algorithm based on reinforcement learning and interaction with cloud data. The specific operations are as follows: C1: State construction: Normalize the multi-source perception system data using the formula Generate standard state vector ; C2: Reinforcement Learning Decision Making: Computing Action-Value Functions via Deep Q-Networks ,in accordance with The greedy strategy selects actions where , t is the number of training cycles; C3: Reward calculation: According to the formula Calculate the single control reward, where is the change in the ship's roll angle, Add value to energy consumption; C4: Model update: using mean square error loss function ,in , update the network parameters through the Adam optimizer; C5: Cloud interaction: Every T = 60 minutes, compare the local model parameters with the cloud server. If the difference , then download the cloud model for fusion update.

[0015] The energy management unit includes a reverse airflow power generation device module, a supercapacitor energy storage group module, and a pulse width modulation circuit module, wherein: The reverse airflow power generation device module is used to utilize the reverse airflow to drive the worm gear to generate electricity and recover energy during braking; The supercapacitor energy storage module is used to store transient energy to meet peak power requirements; The pulse width modulation circuit module is used to adjust the timing of energy recovery and distribution to match the 10ms control cycle.

[0016] The present invention also proposes a vehicle stability control method based on active airflow compensation, comprising the following steps: S1: Using a micro-MEMS pressure sensor array, laser Doppler velocimeter, and inertial navigation system, the flow field pressure distribution, velocity vector, and vehicle attitude angle and acceleration motion parameters are acquired in real time. S2: Outputs accurate data that meets control requirements through filtering, deep learning fusion, and verification and correction processing; S3: Based on the hybrid control strategy and prediction model, the control parameters are calculated and optimized in combination with real-time data to generate a compensation strategy; S4: Control the airflow generating device to generate high-pressure directional airflow to compensate for the vehicle's posture; S5: Recover energy under braking conditions, store it and distribute it reasonably.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention generates directional compensation airflow through distributed airflow generation array units, multi-source perception system units monitor flow fields and motion parameters in real time, intelligent control units integrate hybrid control strategies and reinforcement learning to achieve dynamic compensation and algorithm iteration, and energy management units recover braking energy and optimize its distribution, thereby significantly improving the stability and response speed of vehicles in complex environments. At the same time, gradient porous materials are used to improve airflow utilization, reverse airflow power generation is used to reduce energy consumption, and efficient and energy-saving active stability control is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a system diagram of a vehicle stabilization system and control method based on active airflow compensation according to the present invention.

[0019] Figure 2 Flowchart of the vehicle stabilization system and control method based on active airflow compensation according to the present invention.

[0020] Description of Figure Numbers: 100. Distributed airflow generation array unit; 101. Vector nozzle module; 102. Airflow generation module; 103. Array layout module; 200. Multi-source perception system unit; 201. Data acquisition module; 202. Data processing module; 203. Extended perception module; 300. Intelligent control unit; 301. Control algorithm module; 302. Strategy generation module; 303. Learning optimization module; 400. Energy management unit; 401. Reverse airflow power generation device module; 402. Supercapacitor energy storage group module; 403. Pulse width modulation circuit module. DETAILED DESCRIPTION

[0021] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention. Example 1

[0022] like Figure 1 As shown, this embodiment provides a vehicle stabilization system based on active airflow compensation, including a distributed airflow generation array unit 100, a multi-source perception system unit 200, an intelligent control unit 300, and an energy management unit 400, wherein: the distributed airflow generation array unit 100: is used to include a rotatable vector nozzle and a three-stage centrifugal air compressor to generate a directional compensation airflow; the multi-source perception system unit 200: is used to integrate a MEMS pressure sensor, a laser velocimeter and an inertial navigation system to monitor the flow field state and the carrier motion parameters in real time; the intelligent control unit 300: is used to adopt an LQR+H∞ hybrid control strategy to establish a predictive controller including a simplified model of the Bernoulli equation and the NS equation; the energy management unit 400: is used to recover energy through the reverse airflow power generation technology of the braking condition; the output of the multi-source perception system unit 200 is connected to the input of the intelligent control unit 300, and the output of the intelligent control unit 300 drives the distributed airflow generation array unit 100 to form a closed-loop control.

[0023] In this embodiment, it should also be noted that the distributed airflow generation array unit 100 includes a vector nozzle module 101, an airflow generation module 102, and an array layout module 103. The vector nozzle module 101 is used to achieve efficient directional airflow injection and rapid directional adjustment through a gradient multi-hole vector nozzle and a universal joint structure. The airflow generation module 102 is used to generate and precisely control the high-pressure injection airflow using a three-stage centrifugal air compressor in conjunction with a high-frequency solenoid valve. The specific operations are as follows: A1: The three-stage centrifugal air compressor performs three-stage compression of air through the rotation of the impeller, raising the air pressure to 0.8 MPa. A2: The high-frequency solenoid valve receives pulse signals from the intelligent control unit 300 and controls the on / off and flow rate of the compressed air with a response speed of less than 10 ms. The response time of the high-frequency solenoid valve ensures that the system's compensation delay under turbulent conditions is less than 80 ms. A3: The regulated high-pressure airflow is delivered to the rotatable gradient multi-hole vector nozzle to form a directional compensation airflow. The array layout module 103 is used to optimize the nozzle distribution based on the characteristics of the vehicle and dynamically adjust the operating mode to improve compensation efficiency and energy efficiency.

[0024] Among them, it should be noted that the airflow generating module 102 generates and controls the high-pressure airflow through a three-stage centrifugal air compressor and a high-frequency solenoid valve, and realizes efficient directional injection and direction adjustment through the gradient porous nozzle and universal joint structure of the vector nozzle module 101. At the same time, the array layout module 103 optimizes the nozzle distribution and working mode according to the characteristics of the vehicle.

[0025] Furthermore, it should be noted that the gradient porous vector nozzle is manufactured from a porous material (such as carbon nanotube-reinforced composite) with a porosity that gradually changes from 15% to 35% along the jet direction. This utilizes the Coanda effect to increase airflow utilization to 68%. Furthermore, it possesses high strength and high-temperature resistance, capable of withstanding exhaust pressures of 0.8 MPa and jet velocities of 280 m / s, achieving directional airflow. The universal joint drive mechanism supports a 120° deflection range for the vector nozzle. Its compact and flexible joint design enables rapid response (response time <50 ms), precisely adjusting the airflow direction to compensate for changes in vehicle posture. The three-stage centrifugal air compressor (35 kW per unit, total mass <120 kg, aluminum alloy / titanium alloy components) uses CFD-optimized flow paths to deliver 0.8 MPa of high-pressure airflow, controlled by a high-frequency solenoid valve (response speed <10 ms). The array layout module 103 distributes four groups of vector nozzle arrays on the leading edge, trailing edge, flaps of the wing or the bow, stern, and both sides of the hull according to the type of vehicle, such as an airplane or a ship. The layout position is optimized through wind tunnel tests and numerical simulations to ensure that compensation force can be effectively generated under different flight / navigation conditions.

[0026] In this embodiment, it should also be noted that the multi-source perception system unit 200 includes a data acquisition module 201, a data processing module 202, and an extended perception module 203, wherein: the data acquisition module 201: is used to obtain flow field pressure, velocity and vehicle motion state data in real time using a micro-MEMS pressure sensor array and a laser Doppler velocimeter; the spatial resolution of the micro-MEMS pressure sensor array is 5cm², covering key areas of the vehicle surface such as wings and hulls, and monitoring the airflow / water flow pressure distribution in real time with an accuracy of ±0.1kPa, which is used to identify local flow field disturbances and abnormal pressure fluctuations. The data processing module 202: is used to output accurate and reliable perception information through signal preprocessing, deep learning fusion and verification correction; the specific operations are as follows: B1: Signal preprocessing uses a sliding median filter with a window width of 50ms to output the signal-to-noise ratio , median filter formula: ,in ; n is the filter window radius; B2: Deep learning fusion achieves multi-source data fusion through a deep neural network with 3 fully connected layers, with a fusion error of <0.1m / s. Deep learning fusion uses the fully connected layer weight update formula: ,in, is the weight of the first layer, is the learning rate, J is the loss function; B3: Data verification When the Mahalanobis distance between MEMS and laser data is greater than 3σ, correction is triggered. The data verification correction adopts the error judgment formula: ,when The correction is triggered when is the same source data from different sensors, and m is the number of data samples. Extended Perception Module 203: Used to monitor the surface temperature distribution of the vehicle using infrared thermal imaging sensors and acoustic sensors, and promptly detect local overheating and noise characteristics caused by abnormal airflow to assist system decision-making.

[0027] It should be noted that the data acquisition module 201 obtains flow field and motion data through a micro MEMS pressure sensor array and a laser Doppler velocimeter, and outputs accurate data after signal preprocessing, deep learning fusion, and verification and correction by the data processing module 202. At the same time, the extended perception module 203 uses infrared thermal imaging and acoustic sensors to supplement the monitoring of temperature and noise characteristics.

[0028] Furthermore, it should be noted that the laser Doppler velocimeter has a measurement range of 0-500m / s and a measurement accuracy of ±1%, and can obtain the three-dimensional velocity vector information of the external flow field in real time, accurately judging the air / water flow speed, direction and turbulence intensity. The inertial navigation system integrates MEMS inertial sensors (accelerometers, gyroscopes) and outputs high-precision attitude angles (pitch, roll, yaw), acceleration and position information through advanced filtering algorithms (such as Kalman filtering), and drift errors. , used to monitor the vehicle's own motion status.

[0029] In this embodiment, it should also be noted that the intelligent control unit 300 includes a control algorithm module 301, a strategy generation module 302, and a learning optimization module 303, wherein: the control algorithm module 301: is used to apply the LQR + H∞ hybrid control strategy and prediction model to deal with interference and respond quickly under complex working conditions. The strategy generation module 302: is used to dynamically optimize the nozzle parameters by solving the flow field influence matrix to generate a collaborative compensation control strategy; the learning optimization module 303: is based on reinforcement learning and cloud data interaction to perform autonomous optimization and experience iteration of the control algorithm. The specific operations are as follows: C1: State construction: normalize the multi-source perception system data, and use the formula

[0030] Generate standard state vector ; C2: Reinforcement Learning Decision: Calculating Action-Value Functions via Deep Q-Network ,in accordance with The greedy strategy selects actions where , t is the number of training cycles; C3: reward calculation: according to the formula Calculate the single control reward, where is the change in the ship's roll angle, Add value to energy consumption; C4: Model update: using mean square error loss function ,in , update the network parameters through the Adam optimizer ; C5: Cloud interaction: Every T = 60 minutes, compare the local model parameters with the cloud server. If the difference , then download the cloud model for fusion update.

[0031] Among them, it should be noted that the control algorithm module 301 uses a hybrid control strategy and a prediction model to achieve rapid response, the strategy generation module 302 dynamically optimizes the nozzle compensation strategy based on the flow field matrix, and the learning optimization module 303 autonomously iterates and optimizes the control algorithm through reinforcement learning and cloud interaction.

[0032] Furthermore, it should be noted that the LQR + H∞ hybrid control strategy is based on the linear quadratic regulator and the H∞ robust control algorithm. It optimizes the control input under the system linearization model, minimizes the state error and energy consumption, and copes with model uncertainty and strong external disturbances to ensure system stability.

[0033] The prediction model is combined with a large eddy simulation turbulence model or a three-dimensional hydrodynamic model to establish a high-precision prediction controller, which can predict the disturbance trend within the next 50ms and dynamically adjust the control strategy through a rolling optimization algorithm to achieve rapid response to complex working conditions.

[0034] In this embodiment, it should also be noted that the energy management unit 400 includes a reverse airflow power generation device module 401, a supercapacitor energy storage group module 402, and a pulse width modulation circuit module 403, wherein: the reverse airflow power generation device module 401: is used to use the reverse airflow to drive the worm gear to generate electricity and recover energy during braking; the supercapacitor energy storage group module 402: is used to store transient energy to meet peak power requirements; the pulse width modulation circuit module 403: is used to adjust the timing of energy recovery and distribution to match the 10ms control cycle.

[0035] Among them, it should be noted that the reverse airflow power generation device module 401 converts the reverse airflow kinetic energy into electrical energy during braking, the supercapacitor energy storage group module 402 stores the transient energy to meet the system peak power demand, and the pulse width modulation circuit module 403 accurately adjusts the timing of energy recovery and distribution to match it with the 10ms control cycle.

[0036] Furthermore, it should be noted that under braking conditions, the reverse airflow generator utilizes the reverse airflow generated by the vehicle's deceleration to drive a turbine generator, converting kinetic energy into electrical energy and achieving energy recovery. The system's overall energy efficiency ratio reaches 1:2.3, reducing overall energy consumption by 18%-25%. The energy storage and distribution unit stores the recovered electrical energy in an energy storage device (such as a lithium battery pack) and distributes it appropriately based on system requirements (such as air compressor operation and sensor power supply) to optimize energy utilization efficiency. Example 2

[0037] like Figure 2 In this embodiment, the vehicle stability control method based on active airflow compensation specifically includes the following steps: S1: Real-time collection of multi-source data ① Collection objects: Flow field parameters: A micro-MEMS pressure sensor array (spatial resolution 5 cm², accuracy ±0.1 kPa) monitors the airflow / water flow pressure distribution on the wing / hull surface in real time to identify local disturbances (such as sudden pressure changes caused by turbulence). External flow field velocity: Use a laser Doppler velocimeter (measurement range 0-500m / s, accuracy ±1%) to obtain a three-dimensional velocity vector to determine the direction of air / water flow and turbulence intensity; Carrier motion state: The inertial navigation system integrates the accelerometer and gyroscope data to output pitch angle, roll angle, yaw angle and acceleration (drift error ); ②Data transmission: The acquisition frequency is synchronized with the control cycle (10ms), and the original data is transmitted to the intelligent control unit 300 via a high-speed bus; S2: Multi-source data fusion and verification Signal preprocessing: Use a sliding median filter with a window width of 50ms (formula: ,in is the original signal, is the filtered signal, n is the filter window radius) to remove noise and improve the signal-to-noise ratio to more than 20dB; Deep learning fusion: A deep neural network with three fully connected layers is used to fuse multi-source data. The network input is normalized feature vectors such as pressure, velocity, and posture, and the output is the fused flow field-motion state features (fusion error < 0.1m / s). Data verification and correction: Calculate the Mahalanobis distance between MEMS pressure data and laser velocity data (Formula: ),when Trigger correction when triggering the weighted average correction algorithm to ensure data reliability; S3: Intelligent control strategy generation and optimization ① Hybrid control strategy calculation: The control algorithm module 301 solves the optimal control input of the linearized system model based on the LQR+H∞ hybrid strategy. At the same time, it uses H∞ robust control to suppress external disturbances (such as gusts of wind and waves) to ensure that the system responds to disturbances within 50ms. ② Flow field prediction and parameter optimization: The strategy generation module 302 combines the Bernoulli equation and the simplified NS equation model to predict the flow field changes in the next 50ms and dynamically optimize the vector nozzle parameters by solving the flow field Jacobian matrix:

[0038] Injection pressure: 0.5-0.8MPa (dynamically adjusted according to disturbance intensity); ③ Reinforcement learning autonomous optimization: The learning optimization module 303 implements control strategy iteration through the deep Q network: State construction: Normalize the fused feature vector (Formula: ), as DQN input; Reward function:

[0039] ( is the change in the ship's roll angle, Add value to energy consumption); Cloud collaboration: Synchronize model parameters with the cloud server every 60 minutes, improving adaptability to multiple scenarios through federated learning; S4: Directional airflow compensation execution High-pressure airflow generation: A three-stage centrifugal air compressor (single unit power 35kW, total mass <120kg) compresses the air to 0.8MPa, and a high-frequency solenoid valve (response speed <10ms) adjusts the airflow and flow rate according to control instructions (accuracy ±5%). Dynamic adjustment of vector nozzles: Gradient porous nozzles (porosity 15%-35% gradient) use the Coanda effect to increase airflow utilization to 68%. The universal joint drive mechanism (response time <50ms) achieves 120° deflection of the nozzle to form a directional compensation airflow (injection speed ≤280m / s). The overall system compensates for delays. ; S5: Braking Energy Recovery and Management Energy recovery: Under braking conditions, the reverse airflow drives the turbine generator to generate electricity, and a single braking can recover 5-10kJ of energy (system energy efficiency ratio 1:2.3); Energy storage and distribution: The supercapacitor energy storage group stores transient electrical energy, and the pulse width modulation circuit module 403 adjusts energy distribution with a 10ms cycle, giving priority to supplying peak power devices such as high-frequency solenoid valves and sensors. The remaining energy is stored in the lithium battery pack, and the system's overall energy saving rate is 18%-25%.

[0040] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0041] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A vehicle stabilization system based on active airflow compensation, characterized in that: It comprises a distributed airflow generation array unit (100), a multi-source sensing system unit (200), an intelligent control unit (300), and an energy management unit (400), wherein: The distributed airflow generating array unit (100) is configured to include a rotatable vector nozzle and a three-stage centrifugal air compressor to generate a directional compensating airflow; The multi-source sensing system unit (200) is used to integrate a MEMS pressure sensor, a laser velocimeter, and an inertial navigation system to monitor the flow field state and carrier motion parameters in real time; The intelligent control unit (300) is used to adopt an LQR+H∞ hybrid control strategy to establish a predictive controller including a simplified model of the Bernoulli equation and the NS equation; The energy management unit (400) is used to recover energy through reverse airflow power generation technology under braking conditions; The output of the multi-source sensing system unit (200) is connected to the input of the intelligent control unit (300), and the output of the intelligent control unit (300) drives the distributed airflow generation array unit (100) to form a closed-loop control.

2. The vehicle stabilization system based on active airflow compensation according to claim 1, characterized in that: The distributed airflow generation array unit (100) comprises a vector nozzle module (101), an airflow generation module (102), and an array layout module (103), wherein: The vector nozzle module (101) is used to perform efficient directional airflow injection and rapid direction adjustment through a gradient multi-hole vector nozzle and a universal joint structure; The airflow generating module (102) is used to generate and precisely control the high-pressure jet airflow by using a three-stage centrifugal air compressor in conjunction with a high-frequency solenoid valve; The array layout module (103) is used to optimize the nozzle distribution according to the characteristics of the vehicle, and dynamically adjust the working mode to improve the compensation efficiency and energy efficiency.

3. The vehicle stabilization system based on active airflow compensation according to claim 2, characterized in that: The airflow generation module (102) is used to generate and accurately control the high-pressure jet airflow by using a three-stage centrifugal air compressor in conjunction with a high-frequency solenoid valve. The specific operation is as follows: A1: The three-stage centrifugal air compressor compresses the air in three stages through the rotation of the impeller, raising the air pressure to 0.8MPa; A2: The high-frequency solenoid valve receives the pulse signal from the intelligent control unit (300) and controls the on / off and flow rate of the compressed air with a response speed of <10ms. The response time of the high-frequency solenoid valve makes the compensation delay of the system under turbulent conditions less than 80ms; A3: The regulated high-pressure airflow is delivered to the rotatable gradient multi-hole vector nozzle to form a directional compensation airflow.

4. The vehicle stabilization system based on active airflow compensation according to claim 1, characterized in that: The multi-source perception system unit (200) comprises a data acquisition module (201), a data processing module (202), and an extended perception module (203), wherein: The data acquisition module (201) is used to obtain flow field pressure, speed and vehicle motion state data in real time using a micro MEMS pressure sensor array and a laser Doppler velocimeter; The data processing module (202) is used to output accurate and reliable perception information through signal preprocessing, deep learning fusion and verification correction; The extended perception module (203) is used to monitor the surface temperature distribution of the vehicle with the help of infrared thermal imaging sensors and acoustic sensors, and promptly detect local overheating and noise characteristics caused by abnormal airflow to assist system decision-making.

5. The vehicle stabilization system based on active airflow compensation according to claim 4, characterized in that: The data processing module (202) outputs accurate and reliable perception information through signal preprocessing, deep learning fusion, and verification and correction. The specific operations are as follows: B1: Signal preprocessing uses a sliding median filter with a window width of 50ms, and the output signal-to-noise ratio , median filter formula: ,in is the original signal, is the filtered signal, n is the filter window radius; B2: Deep learning fusion achieves multi-source data fusion through a deep neural network with three fully connected layers, with a fusion error of less than 0.1m / s. Deep learning fusion uses the fully connected layer weight update formula: ,in, is the weight of the first layer, is the learning rate, J is the loss function; B3: Data verification: When the Mahalanobis distance between MEMS and laser data is greater than 3σ, correction is triggered. The error judgment formula is used for data verification and correction: ,when The correction is triggered when is the homologous data from different sensors, and m is the number of data samples.

6. The vehicle stabilization system based on active airflow compensation according to claim 4, characterized in that: The micro-MEMS pressure sensor array has a spatial resolution of 5 cm², covering key areas on the surface of vehicles such as wings and hulls, and monitors the air / water pressure distribution in real time with an accuracy of ±0.1 kPa, which is used to identify local flow field disturbances and abnormal pressure fluctuations.

7. The vehicle stabilization system based on active airflow compensation according to claim 1, characterized in that: The intelligent control unit (300) comprises a control algorithm module (301), a strategy generation module (302), and a learning optimization module (303), wherein: The control algorithm module (301) is used to use the LQR + H∞ hybrid control strategy and prediction model to deal with interference and perform rapid response under complex working conditions; The strategy generation module (302) is used to generate a collaborative compensation control strategy by dynamically optimizing nozzle parameters by solving the flow field influence matrix; The learning optimization module (303) performs autonomous optimization and experience iteration of the control algorithm based on reinforcement learning and interaction with cloud data.

8. The vehicle stabilization system based on active airflow compensation according to claim 7, characterized in that: The learning optimization module (303) performs autonomous optimization and experience iteration of the control algorithm based on reinforcement learning and interaction with cloud data. The specific operations are as follows: C1: State construction: Normalize the multi-source perception system data using the formula Generate standard state vector ; C2 : Reinforcement Learning Decision Making: Computing Action-Value Functions via Deep Q-Networks ,in accordance with The greedy strategy selects actions where , t is the number of training cycles; C3: Reward calculation: According to the formula Calculate the single control reward, where is the change in the ship's roll angle, Add value to energy consumption; C4: Model update: using mean square error loss function ,in , update the network parameters through the Adam optimizer; C5: Cloud interaction: Every T = 60 minutes, compare the local model parameters with the cloud server. If the difference , then download the cloud model for fusion update.

9. The vehicle stabilization system based on active airflow compensation according to claim 1, characterized in that: The energy management unit (400) comprises a reverse airflow power generation device module (401), a supercapacitor energy storage group module (402), and a pulse width modulation circuit module (403), wherein: The reverse airflow power generation device module (401) is used to utilize the reverse airflow to drive the worm gear to generate electricity and recover energy during braking; The supercapacitor energy storage module (402) is used to store transient energy to cope with peak power demand; The pulse width modulation circuit module (403) is used to adjust the timing of energy recovery and distribution, matching the 10ms control cycle.

10. A vehicle stability control method based on active airflow compensation, characterized in that: A vehicle stabilization system based on active airflow compensation according to any one of claims 1 to 9 is used, comprising the following steps: S1: Using a micro-MEMS pressure sensor array, laser Doppler velocimeter, and inertial navigation system, the flow field pressure distribution, velocity vector, and vehicle attitude angle and acceleration motion parameters are acquired in real time. S2: Outputs accurate data that meets control requirements through filtering, deep learning fusion, and verification and correction processing; S3: Based on the hybrid control strategy and prediction model, the control parameters are calculated and optimized in combination with real-time data to generate a compensation strategy; S4: Control the airflow generating device to generate high-pressure directional airflow to compensate for the vehicle's posture; S5: Recover energy under braking conditions, store it and distribute it reasonably.

Images