Unmanned aerial vehicle multi-polarized antenna array based on liquid metal dynamic reconstruction and switching control method

Abstract

The invention provides an unmanned aerial vehicle multi-polarized antenna array based on liquid metal dynamic reconstruction and a switching control method, and relates to the field of unmanned aerial vehicle communication systems, and the unmanned aerial vehicle multi-polarized antenna array comprises a liquid metal antenna unit array, a microfluidic driving system, an electromagnetic environment sensing module, an intelligent configuration decision unit and a control execution module. A three-layer structure design is adopted, specific arrangement and connection modes of a top layer micro-channel network, a middle layer electrode array and a bottom layer control circuit are included, dynamic reconstruction of the antenna form, the polarization mode and the working frequency is achieved through the liquidity and conductivity of liquid metal, and the communication adaptability and the anti-interference capacity of an unmanned aerial vehicle in a complex electromagnetic environment are improved. Liquid metal flow is accurately controlled in an electrostatic field and pressure field cooperative driving mode, an optimal antenna configuration decision is achieved based on a deep reinforcement learning algorithm, and the method has the advantages of being high in response speed, high in reconstruction precision, low in energy consumption and the like.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle (UAV) communication systems, and in particular to a multi-polarized UAV antenna array and a switching control method based on liquid metal dynamic reconstruction. Background Art

[0002] Drones are increasingly used in military reconnaissance, emergency communications, border patrols, and other fields. The reliability of their communication systems is crucial to mission execution. Antennas, as key components of drone communication systems, directly determine communication quality. Multi-polarization antennas can effectively reduce polarization mismatch losses and improve communication quality by adjusting polarization modes (vertical, horizontal, circular, etc.). Existing drone antenna systems have the following shortcomings:

[0003] 1. Lack of flexibility: Traditional fixed-structure antennas are difficult to adapt to complex and changing electromagnetic environments, especially in areas with dense low-altitude interference, where communication quality can be significantly affected. 2. Slow response speed: Although mechanically reconfigurable antennas can change the antenna shape, their response speed is slow (usually >100ms), making it difficult to respond to rapidly changing electromagnetic environments in real time, especially when the drone is flying at high speed. 3. Low integration: Existing reconfigurable antenna systems are usually large in size and complex in structure, making it difficult to achieve high integration on drone platforms with limited space. 4. High energy consumption: Technologies such as phased arrays require a large number of RF front-end circuits and signal processing units, resulting in high energy consumption (usually >10W), which is not suitable for battery-powered small drone systems. Reconfigurable antenna technology dynamically changes the physical form or electrical characteristics of the antenna to adapt it to different communication needs and electromagnetic environments, making it an effective way to solve the above problems.

[0004] 1. Mechanically reconfigurable antenna: The physical structure of the antenna is changed through physical mechanisms such as motors and actuators to achieve changes in antenna characteristics. For example, the reconfigurable antenna based on microelectromechanical systems (MEMS) developed by the University of Arizona in the United States changes the antenna shape through a tiny mechanical structure. 2. RF switch reconfigurable antenna: PIN diodes, MEMS switches, etc. are used to control the on and off of the antenna circuit to change the antenna working state. For example, the multi-band reconfigurable antenna based on PIN diodes developed by KAIST in South Korea. 3. Material property reconfigurable antenna: Special materials such as ferrites and liquid crystals are used to change the material properties through external conditions (electric fields, magnetic fields), thereby changing the antenna performance. For example, the reconfigurable antenna based on ferrite materials developed by the University of Electronic Science and Technology of China.

[0005] However, existing liquid metal antennas primarily focus on the reconfigurability of single antenna elements, with limited research on antenna arrays. In particular, research on multi-polarized antenna arrays for UAV platforms is lacking. Existing control methods often rely on a single drive mechanism, resulting in limited control accuracy and response speed. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a multi-polarization antenna array and switching control method for unmanned aerial vehicles based on liquid metal dynamic reconstruction to solve the problems raised in the above background technology. The present invention has small storage space requirements, high stability and long service life.

[0007] In order to achieve the above-mentioned objectives, the present invention is implemented through the following technical solutions: a multi-polarized antenna array for unmanned aerial vehicles based on dynamic reconstruction of liquid metal, comprising a liquid metal antenna unit array, a microfluidic drive system, an electromagnetic environment perception module, an intelligent configuration decision unit, a control execution module and a power supply and heat dissipation system, wherein the liquid metal antenna unit array is composed of multiple liquid metal antenna units in a specific arrangement, and each unit comprises a microchannel network, an electrode array and a liquid metal storage area; the microfluidic drive system comprises a micro pump, an electrode control circuit, an electrostatic drive unit and a pressure sensor; the electromagnetic environment perception module comprises a spectrum analyzer, a signal strength detector and an interference identification unit; the intelligent configuration decision unit is based on a deep reinforcement learning algorithm, and automatically generates the optimal antenna configuration scheme according to the environmental perception results and communication requirements; the control execution module receives the configuration decision instruction, converts it into a specific microfluidic drive command, controls the liquid metal flow to realize antenna reconstruction; the power supply and heat dissipation system provides a stable power supply for the entire system.

[0008] Furthermore, it also includes a liquid metal antenna unit structure: each liquid metal antenna unit adopts a three-layer structure design:

[0009] The top layer is made of flexible polydimethylsiloxane (PDMS) and contains a network of microchannels for the flow of liquid metal. The microchannels are designed in a specific pattern to form the antenna shape required for different polarization modes.

[0010] Middle layer: electrode array layer, including electrostatic driving electrodes and conductivity detection electrodes arranged around the microchannel, used to control and monitor the flow of liquid metal;

[0011] Bottom layer: control circuit layer, which includes micro pumps, pressure control valves and drive circuits to achieve precise control of the flow of liquid metal.

[0012] Furthermore, it also includes the antenna array arrangement: the line units are arranged in a 4×4 matrix to form a planar array with an overall size of 50mm×50mm×5mm, and the entire array has a modular structure.

[0013] Furthermore, the microfluidic driving system adopts a coordinated driving mode of electrostatic field and pressure field, including:

[0014] Electrostatic drive unit: An electrode pair is set around the microchannel, and an electrostatic field is generated by applying voltage, thereby controlling the flow direction and speed of the liquid metal using the continuous electrowetting effect;

[0015] Pressure drive unit: This includes a micropump and pressure control valve, providing the base pressure required for liquid metal flow. It works in conjunction with the electrostatic drive. This collaborative drive method is based on a dynamic surface tension gradient model that takes into account the effects of electric, pressure, and temperature fields, and is used to accurately calculate the total force required to drive the liquid metal.

[0016] Pressure sensor: monitors the pressure distribution in the microchannel in real time, forms a closed-loop control, and improves control accuracy.

[0017] Furthermore, it also includes the control system architecture, which adopts a hierarchical design:

[0018] Low-level control: performs specific liquid metal flow control, including basic operations such as electrode voltage adjustment and pump pressure control;

[0019] Middle-level control: converts antenna configuration instructions into specific control sequences and coordinates the work of each drive unit;

[0020] Upper-layer control: Provides intelligent decision-making and selects antenna configuration based on environmental perception results.

[0021] A switching control method based on the above-mentioned antenna array includes state space reconstruction and action space optimization. The state space reconstruction includes: obtaining spectrum occupancy information of the current electromagnetic environment, polarization matching information between transmitting and receiving antennas, and current radiation direction information of the antenna array to form an environmental perception vector.

[0022] Furthermore, the state space reconstruction defines the environment perception vector S t for:

[0023]

[0024] Among them, S t is the state vector at time t; the first term is the spectrum sensing component, δ(·) is the Dirac function, f is the center frequency of the UAV communication channel, and f i is the center frequency of the i-th in-band or adjacent interference signal detected, N is the total number of interference signals detected; the second term is the polarization mismatch, where E rx and E tx The electric field strength vectors of the receiving antenna and the transmitted signal are respectively. The calculation result of this item is a scalar (range [0,1]) that characterizes the degree of polarization alignment, with a value of 0 indicating perfect alignment and a value of 1 indicating orthogonal mismatch. The third item is the main lobe direction, and AF(θ) is the normalized array factor of the antenna array in the direction θ. This item gives the direction of maximum gain at the moment.

[0025] The action space a t include:

[0026] a t =[{b1,b2,…,b m},{V1,V2,…,V k}]

[0027] Among them, b i ∈{0,1} is the binary on / off control signal for the i-th microchannel, with a total of m controllable channels; V j ∈[0,50V] is the voltage value applied to the jth driving electrode, with a total of k driving electrodes.

[0028] Furthermore, it also includes the spatiotemporal attention deep Q network (STA-DQN), the network structure is:

[0029] Input layer: St∈R128;

[0030] Spatiotemporal encoding layer: TransformerEncoder (4-head);

[0031] Strategy layer: dual-stream fully connected network;

[0032] Output layer: Q(s,a)eR2"x10;

[0033] The loss function is:

[0034]

[0035] Among them, r is the immediate reward of environmental feedback, s and s' are the current and next states respectively, a and a' are the current and next actions respectively, γ∈[0,1] is the reward discount factor, Q(s,a) is the action value output by the current Q network, Q target (s',a') is the action value output by the target Q network, λ is the regularization coefficient, D KL (π current ||π prior ) is the current policy π current The strategy π of the previous iteration prior The Kullback-Leibler divergence between is used to stabilize the training process.

[0036] Furthermore, it also includes polarization-spatial joint optimization: the polarization state is adjusted through a dynamic feedback mechanism to optimize the polarization matching degree. The calculation formula is:

[0037]

[0038] Where S = (S0, S1, S2, S3) is the Stokes vector of the transmitted signal, and S' = (S'0, S'1, S'2, S'3) is the Stokes vector of the signal effectively received by the receiving antenna. S0 and S'0 are the total intensity, S1, S2 and S'1, S'2 are the linearly polarized components, and S3 and S'3 are the circularly polarized components. The result of this formula is the dimensionless polarization matching efficiency.

[0039] Furthermore, it also includes a pattern synthesis algorithm: by introducing a multi-constraint convex optimization model, the array pattern is dynamically optimized, and the calculation formula is:

[0040]

[0041] stw H a(θ d ) = 1 (main lobe constraint)

[0042] w H a(θ j )=0(zero constraint)

[0043] |w i |≤1(amplitude constraint)

[0044] The frequency tuning equation is:

[0045]

[0046] Among them, L eff is the equivalent electrical length of the antenna after reconstruction, L0 is the initial equivalent electrical length, α=0.15 is the experimentally calibrated liquid metal deformation coefficient, ΔV is the change in driving voltage, and V0 is the initial or reference driving voltage.

[0047] Beneficial effects of the present invention:

[0048] 1. The antenna in this invention can be reconfigured in real time based on the electromagnetic environment. Its polarization, operating frequency, and directional pattern can be dynamically adjusted to adapt to complex electromagnetic environments. Experiments have shown that in strong interference environments, communication link reliability is improved by over 75%. The average response time of liquid metal flow control is 8.5ms, far superior to traditional mechanically reconfigurable antennas (typically over 100ms), enabling real-time response to rapidly changing electromagnetic environments.

[0049] 2. The UAV multi-polarization antenna array system based on liquid metal dynamic reconstruction has an overall size of only 50mm×50mm×5mm and weighs less than 50g, making it suitable for various small UAV platforms; the system's operating power consumption is less than 1.5W, which is more than 90% energy-saving than traditional phased array technology (usually more than 10W), significantly extending the UAV's flight time.

[0050] 3. The present invention supports dynamic switching of multiple polarization modes, including vertical polarization, horizontal polarization, and circular polarization, reducing polarization mismatch losses by 85%. The synergistic driving mechanism of the electrostatic field and the pressure field enables liquid metal position control with an accuracy of 50μm, far superior to the traditional single driving method.

[0051] 4. The deep reinforcement learning algorithm of the present invention enables the system to autonomously learn the optimal configuration strategy, improving the communication quality in complex environments by more than 60%. In addition, the liquid metal can be recycled, the system structure is simple, and the maintenance cost is reduced by more than 50% compared with traditional reconfigurable antennas. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a diagram of the overall system architecture of the present invention;

[0053] Figure 2 Schematic diagram of the three-layer structure of the liquid metal antenna unit;

[0054] Figure 3 Reconstructing the schematic for the liquid metal antenna;

[0055] Figure 4 This is a schematic diagram of the microfluidic driving principle of the present invention;

[0056] Figure 5 Dynamically reconfigure the flow chart for the antenna array;

[0057] Figure 6 Flowchart of decision-making algorithm for intelligent configuration;

[0058] Figure 7 This is a schematic diagram of multi-polarization mode switching;

[0059] Figure 8 Graphs showing experimental results of testing the signal-to-noise ratio (SINR) and bit error rate (BER) under vertical polarization, horizontal polarization, and circular polarization interference, respectively, in an embodiment of the present invention;

[0060] Figure 9 This is a comparison diagram of antenna reconstruction response speed in an embodiment of the present invention;

[0061] Figure 10 This is a comparison chart of energy consumption of different antenna systems in the embodiments of the present invention, where the integration difficulty is defined as three levels: Level 1 (can be directly mounted on the drone shell without modifying the main structure), Level 2 (requires minor structural modifications), and Level 3 (requires deeply customized integration). DETAILED DESCRIPTION

[0062] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0063] See also Figures 1 to 10, the present invention provides the following technical solutions:

[0064] Example 1

[0065] The multi-polarization antenna array for UAVs based on liquid metal dynamic reconstruction mainly consists of the following parts:

[0066] 1.1 Liquid Metal Antenna Unit Array: This array consists of multiple liquid metal antenna units arranged in a specific pattern. Each unit contains a microchannel network, an electrode array, and a liquid metal storage area. The liquid metal is made of gallium-indium alloy (GaIn), which has excellent conductivity and fluidity.

[0067] 1.2 Microfluidic drive system: includes a micro pump, electrode control circuit, electrostatic drive unit and pressure sensor, which is used to precisely control the flow and distribution of liquid metal in the microchannel.

[0068] 1.3 Electromagnetic environment perception module: includes spectrum analyzer, signal strength detector and interference identification unit, which is used to monitor the surrounding electromagnetic environment in real time and provide a basis for antenna configuration decision.

[0069] 1.4 Intelligent Configuration Decision Unit: Based on a deep reinforcement learning algorithm, it automatically generates the optimal antenna configuration plan according to environmental perception results and communication requirements.

[0070] 1.5 Control execution module: receives configuration decision instructions, converts them into specific microfluidic drive commands, and controls the flow of liquid metal to achieve antenna reconstruction.

[0071] 1.6 Power supply and cooling system: Provides stable power supply for the entire system and ensures that the system operating temperature is within the appropriate range of liquid metal.

[0072] This embodiment also provides a structural description of the above system, which is as follows:

[0073] 2.1 Liquid Metal Antenna Unit Structure

[0074] Each liquid metal antenna unit adopts a three-layer structure design:

[0075] The top layer is made of flexible polydimethylsiloxane (PDMS) and contains a network of microchannels for the liquid metal to flow. The microchannels are designed in a specific pattern to form the antenna shape required for different polarization modes.

[0076] Middle layer: electrode array layer, including electrostatic driving electrodes and conductivity detection electrodes arranged around the microchannel, which are used to control and monitor the flow of liquid metal.

[0077] Bottom layer: control circuit layer, which includes micro pumps, pressure control valves and drive circuits to achieve precise control of the flow of liquid metal.

[0078] The unit size is 10mm×10mm×2mm, and the miniaturized design is suitable for integration into UAV platforms.

[0079] 2.2 Antenna Array Arrangement

[0080] The antenna elements are arranged in a 4×4 matrix, forming a planar array with an overall size of 50mm×50mm×5mm (including control circuitry). The array adopts a modular design, which can be expanded or adjusted according to the needs of different drone platforms.

[0081] The antenna unit of the present invention is manufactured using multi-layer soft lithography. First, a mold for the microchannel network is created on a silicon wafer by spin-coating SU-8 photoresist. PDMS prepolymer is then poured onto the mold and thermally cured to form the top microchannel layer. The middle electrode array is fabricated on another PDMS film using magnetron sputtering or screen printing. The bottom layer of control circuitry and micropump valves is implemented using flexible printed circuit (FPC) technology. Finally, oxygen plasma bonding is used to align and permanently bond the three layers, and a gallium-indium alloy is injected through a reserved injection port.

[0082] 2.3 Microfluidic drive system

[0083] The microfluidic drive system uses a synergistic drive method of electrostatic field and pressure field, including:

[0084] Electrostatic drive unit: Electrode pairs are placed around the microchannel. Voltage is applied to generate an electrostatic field, leveraging the continuous electrowetting (EWOD) effect to control the direction and speed of liquid metal flow. Pressure drive unit: Consists of a micropump and pressure control valve, providing the base pressure required for liquid metal flow and working in conjunction with the electrostatic drive. Pressure sensor: Real-time monitoring of pressure distribution within the microchannel creates a closed-loop control loop, improving control accuracy.

[0085] 2.4 Control System Architecture

[0086] The control system adopts a hierarchical design:

[0087] Bottom-level control: responsible for executing specific liquid metal flow control, including basic operations such as electrode voltage adjustment and pump pressure control.

[0088] Middle-level control: responsible for converting antenna configuration instructions into specific control sequences and coordinating the work of each drive unit.

[0089] Upper-layer control: Implements intelligent decision-making and selects the optimal antenna configuration based on environmental perception results.

[0090] This embodiment explains the working principle of the liquid metal antenna unit to facilitate understanding of the technical solution of the present invention:

[0091] The present invention proposes a dynamic electric-mechanical synergistic driving model, which breaks through the limitations of the traditional single driving mode. It combines the multi-field synergistic effects of electric field, pressure field and temperature field to achieve precise dynamic adjustment of the liquid metal antenna shape.

[0092] The liquid metal antenna unit uses gallium indium tin (GaInSn) alloy with a conductivity of 3.8×10 6 The liquid metal, with a surface tension of 580 mN / m, remains liquid at room temperature and has electrical conductivity close to that of copper. By reconfiguring the topology of the microchannel network, the flow of the liquid metal can be precisely controlled to form different antenna configurations, enabling dynamic polarization and directionality adjustments.

[0093] Microchannel Network Design

[0094] The present invention adopts a fractal topology (Hilbert curve) and realizes different antenna configuration switching through a multi-level microchannel network (width of 200μm). This design can realize the switching of multiple polarization modes and frequency bands.

[0095] Vertical polarization: activates the central straight channel;

[0096] Circular polarization: Activate the ring channel (the ring circumference is matched to λ / 4);

[0097] Multi-band: segmented filling of channels of different lengths (such as λ / 4, 3λ / 4);

[0098] The above-mentioned microchannel design can dynamically generate vertical, circularly polarized and multi-band antennas, greatly expanding the application range of liquid metal antennas.

[0099] This example also proposes a dynamic surface tension gradient model, modifies the traditional electrowetting (EWOD) formula, combines the synergistic effects of the electric field, pressure field, and temperature field, and proposes a new liquid metal flow driving force model:

[0100] F total =F EWOD +F Marangoni +F Pressure -F viscous

[0101]

[0102] To facilitate understanding by those skilled in the art, the above formula is a simplified model based on standard theory.

[0103] F total The combined force that drives the liquid metal to flow.

[0104] The first term is the electrowetting (EWOD) driving force, where ∈0 is the vacuum dielectric constant, ∈r is the relative dielectric constant of the medium, V is the applied voltage, W is the droplet width, and d is the thickness of the dielectric layer.

[0105] The second term is the thermocapillary driving force caused by the Marangoni effect, where C m is the coefficient related to geometry, is the rate of change of surface tension γ with temperature T, is the temperature gradient.

[0106] The third term is the pressure driving force, where ΔP is the pressure difference across the microchannel and A is the channel cross-sectional area.

[0107] The fourth term is the viscous resistance, where η is the dynamic viscosity of the liquid metal, L is the droplet length, W is the channel width, and h is the channel height. is the average flow velocity of the liquid metal.

[0108] The control equation proposed above greatly improves the accuracy and response speed of antenna shape adjustment by combining the electro-thermal coupling effect with liquid metal flow control.

[0109] Experimental verification:

[0110] Through experimental verification, the driving speed is increased to 5mm / s, which is four times higher than the traditional method of 1.2mm / s, and the average response time can reach 8.5ms.

[0111] This embodiment also provides a multi-polarization array switching control method:

[0112] This embodiment proposes a deep reinforcement learning framework for spatiotemporal joint coding (ST-DRL), which breaks the limitations of traditional fixed polarization modes and implements dynamic and intelligent configuration of antenna arrays. The state space reconstruction defines the environment perception vector St as:

[0113] Action space a t Defined as a tuple consisting of two parts: a t =(b,V)where:

[0114] b=(b1,b2,…,b m ) is an m-dimensional binary vector, where b i ∈{0,1} represents the on / off control of the i-th microchannel, which determines the flow path of the liquid metal and thus forms different antenna shapes.

[0115] V=(V1,V2,…,V k ) is a k-dimensional voltage vector, where V j ∈[0,50V] represents the specific voltage value applied to the jth driving electrode. This is used to precisely control the position and shape of the liquid metal.

[0116] Where δ(·) is the Dirac function and AF(θ) is the array factor, which represents the gain of the antenna array in different directions.

[0117] The action space consists of:

[0118]

[0119] Among them, b I is a binary signal (channel on / off status), Vi is the driving voltage, supports 2 m Antenna shape combination.

[0120] This embodiment also provides a spatiotemporal attention deep Q network (STA-DQN), which introduces spatiotemporal coding technology to enhance the performance of traditional deep Q learning (DQN) in dynamic environments. The network structure is:

[0121] Input layer: St∈R128;

[0122] Spatiotemporal encoding layer: TransformerEncoder (4-head);

[0123] Strategy layer: dual-stream fully connected network;

[0124] Output layer: Q(s,a)eR2"x10;

[0125] The loss function is:

[0126]

[0127] Among them, λ is the regularization term, KL is the Kullback-Leibler divergence, π current is the current policy, π prior For the target strategy.

[0128] The above solution optimizes the reinforcement learning strategy by introducing the spatiotemporal attention mechanism, enabling the system to quickly adapt to changes in a dynamic environment and improve performance.

[0129] Experimental data:

[0130] The spatiotemporal encoding layer is composed of four stacked Transformer encoder units, each containing a four-head multi-head self-attention module and a feedforward neural network. The strategy layer uses a two-stream Q network structure, one of which outputs the state value V(s) through a three-layer fully connected network (with 64, 32, and 1 neurons, respectively), and the other outputs the advantage function A(s, a) through a three-layer fully connected network (with 64, 32, and |A| neurons, respectively, where |A| is the action space dimension). Training uses the Adam optimizer with a learning rate of 0.001. In a dynamic interference environment, the communication interruption rate is reduced by 62% and spectrum utilization is improved by 45%.

[0131] This embodiment also provides an antenna performance optimization algorithm: a polarization-spatial domain joint optimization theory is proposed, which solves the problems of traditional single-dimensional optimization and improves the overall performance of the antenna.

[0132] Polarization matching optimization:

[0133] The polarization state is adjusted through a dynamic feedback mechanism to optimize the polarization matching degree. The calculation formula is:

[0134]

[0135] Where: S′i is the Stokes parameter at the transmitter, S ′ i ′ The Stokes parameters at the receiving end are adjusted dynamically by distributing the liquid metal to ensure that the matching degree ηp>0.9, thereby improving the polarization matching and optimizing signal reception.

[0136] This embodiment also provides a directional pattern synthesis algorithm

[0137] By introducing a multi-constraint convex optimization model, the array pattern is dynamically optimized, and the calculation formula is:

[0138]

[0139] stw H a(θ d ) = 1 (main lobe constraint)

[0140] w H a(θ j )=0(zero constraint)

[0141] |w i |≤1(amplitude constraint)

[0142] The frequency tuning equation is:

[0143]

[0144] Among them, Leff is the equivalent electrical length of the antenna after reconstruction, L0 is the initial equivalent electrical length, α = 0.15 is the experimentally calibrated liquid metal deformation coefficient, ΔV is the change in driving voltage, and V0 is the initial or reference driving voltage

[0145] Wherein, α=0.15 is the deformation coefficient of liquid metal, and ΔV is the driving voltage change.

[0146] Measured data:

[0147] At 1.2GHz, the frequency tuning accuracy reaches ±0.5%, demonstrating the high-precision frequency tuning capability of this technology.

[0148] This embodiment also provides an innovative system integration design: a three-dimensional heterogeneous integration architecture is adopted to achieve deep integration of the metasurface and the drone, thereby optimizing the performance and integration of the system.

[0149] Cooling solution:

[0150] A microchannel phase change heat dissipation system was designed to optimize the thermal management of the system. The thermal resistance model of the heat dissipation system is:

[0151]

[0152] The symbols are defined as follows:

[0153] R total : Total thermal resistance of the system (°C / W).

[0154] T j : Junction temperature (°C) in critical areas of the antenna system (such as the chip or high power density area).

[0155] T a : Ambient temperature (℃).

[0156] P diss : Total power dissipation of the system (W).

[0157] R cond : Thermal conduction resistance (℃ / W), which indicates the resistance encountered when heat is conducted through a solid material.

[0158] R conv : Convective thermal resistance (℃ / W), which represents the resistance encountered when heat is dissipated from a solid surface through a fluid (such as air) by convection.

[0159] R phase : Phase change thermal resistance (℃ / W), specifically refers to the thermal resistance equivalent to the heat absorbed by the material during the phase change process when using phase change materials for heat dissipation

[0160] This cooling solution uses phase change materials to improve the stability of the system at high power through efficient thermal management.

[0161] Test results:

[0162] At 5W system power consumption, the temperature rise of key areas on the antenna surface relative to the ambient temperature is controlled within 15°C, demonstrating the system's superior heat dissipation performance.

[0163] This technical solution combines innovative technologies such as liquid metal antennas, multi-polarization array switching control, space-time coded deep reinforcement learning algorithms, and polarization matching optimization to form a complete closed-loop technology loop, demonstrating both high innovation and breakthroughs. The proposed algorithmic formula, control method, and system architecture all demonstrate significant innovation and superior anti-interference performance in interference environments.

[0164] In order to verify the effectiveness of the present invention, the following comparative experiments were conducted:

[0165] Experiment 1: Comparison of anti-interference performance under different polarization modes

[0166] This experiment compares the performance of the multi-polarization antenna array of the present invention and the traditional fixed polarization antenna under different polarization interferences.

[0167] Experimental setup:

[0168] Test environment: semi-anechoic chamber

[0169] Communication frequency: 2.4GHz

[0170] Interference source: directional antenna, output power 30dBm

[0171] Test plan: Test the signal-to-noise ratio (SINR) and bit error rate (BER) under vertical polarization, horizontal polarization, and circular polarization interference respectively.

[0172] Experimental results show that the multi-polarized antenna array of the present invention maintains a high SINR under various polarization interference conditions, significantly exceeding that of traditional fixed-polarization antennas. In terms of bit error rate (BER), the present invention's solution also significantly outperforms traditional solutions. Specific comparative data is shown in Table 1 below.

[0173] Table 1: Comparison of anti-interference performance under different polarization interference

[0174]

[0175]

[0176] As can be seen from Table 1, the present invention can effectively combat interference from different polarization directions through dynamic polarization reconstruction, and the reliability of the communication link is greatly improved.

[0177] Experiment 2: Antenna Reconfiguration Response Speed ​​Comparison

[0178] This experiment compared the performance differences in response speed between the liquid metal dynamic reconfigurable antenna of the present invention and the traditional mechanical reconfigurable antenna. Specific data are shown in Table 2 below.

[0179] Table 2: Antenna reconfiguration response speed comparison

[0180] Refactoring Type Liquid metal antenna of the present invention Traditional mechanical reconfigurable antenna Speed ​​improvement rate Polarization switching 8.5ms (average) 135ms (average) ～93.7％ Frequency tuning 12.0ms (average) 150ms (average) ～92.0％

[0181] Experimental setup:

[0182] Test environment: laboratory environment

[0183] Reconstruction type: Polarization switching (vertical polarization → horizontal polarization → circular polarization)

[0184] Test indicator: time required to complete reconstruction

[0185] Experimental results show that the response time of the present invention reaches millisecond level, which is more than an order of magnitude faster than the traditional mechanical reconstruction method, and can meet the real-time reconstruction requirements in highly dynamic environments.

[0186] Experiment 3: Energy Consumption Comparison of Different Antenna Systems

[0187] This experiment compares the system power consumption (excluding RF transmission power) of the present invention, a traditional phased array antenna, and a mechanically reconfigurable antenna under typical operating conditions. Specific data is shown in Table 3 below.

[0188] Table 3: Comparison of antenna system energy consumption, weight and integration difficulty

[0189] Performance indicators Liquid metal antenna of the present invention Traditional phased array antenna Traditional mechanical reconfigurable antenna Power consumption (W) 1.2W 12.5W 3.8W Weight (g) 48g ～200g ～150g Integration difficulty Level 1 (low) Level 3 (High) Level 2 (Medium)

[0190] Experimental setup:

[0191] Test environment: laboratory environment

[0192] Test conditions: The antenna system operates in the 2.4GHz frequency band, and the transmit power is fixed at 20dBm

[0193] Test indicators: Antenna system power consumption (excluding transmitted RF power).

[0194] Experimental results show that the present invention has significant advantages in energy consumption and lightweight, saving about 90% of energy (compared to phased array), low integration difficulty, and is more suitable for UAV platforms with limited power consumption and space.

[0195] Example 2 Electromagnetically driven liquid metal antenna

[0196] Alternative solution 1 uses electromagnetic drive instead of the electrostatic and pressure field synergistic drive mechanism. Specifically, a solenoid coil is placed around the microchannel to generate a magnetic field, which uses the Lorentz force to drive the liquid metal flow. This solution has the advantage of high driving force, making it suitable for controlling larger liquid metals. However, it has the disadvantages of high energy consumption and increased system complexity.

[0197] In specific implementation, the driving force calculation formula is:

[0198] F=IL×B

[0199] Where I is the current in the liquid metal, L is the effective length, and B is the magnetic induction intensity. The control algorithm needs to be adjusted accordingly, but the basic framework of intelligent configuration decision-making can remain unchanged.

[0200] Example 3: Thermally Driven Liquid Metal Antenna

[0201] Alternative embodiment 2 uses a thermally driven mechanism to control the flow of liquid metal. Specifically, heating resistors are placed within the microchannel network. Localized heating changes the surface tension of the liquid metal, creating the Marangoni effect and driving the liquid metal flow. This approach has the advantages of a simple structure and easy control circuit implementation; however, it has the disadvantages of slow response, difficulty in precise control, and high energy consumption.

[0202] Thermal driving force calculation formula:

[0203]

[0204] Where: A is the effective surface area of ​​the liquid metal, is the rate of change of liquid metal surface tension γ with temperature T (Marangoni coefficient), To adjust the temperature gradient on the liquid metal surface, in this scheme, the microchannel network and antenna morphology design can remain unchanged, and only the driving mechanism and the corresponding control algorithm need to be replaced.

[0205] The basic principles, main features and advantages of the present invention are shown and described above. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention.

[0206] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A multi-polarization antenna array for drones based on liquid metal dynamic reconfiguration, characterized by: It includes a liquid metal antenna unit array, a microfluidic drive system, an electromagnetic environment perception module, an intelligent configuration decision unit, a control execution module and a power supply and heat dissipation system. The liquid metal antenna unit array is composed of multiple liquid metal antenna units in a specific arrangement, and each unit contains a microchannel network, an electrode array and a liquid metal storage area; the microfluidic drive system includes a micro pump, an electrode control circuit, an electrostatic drive unit and a pressure sensor; the electromagnetic environment perception module includes a spectrum analyzer, a signal strength detector and an interference identification unit; the intelligent configuration decision unit is based on a deep reinforcement learning algorithm and automatically generates the optimal antenna configuration scheme according to the environmental perception results and communication requirements; the control execution module receives the configuration decision instructions, converts them into specific microfluidic drive commands, controls the liquid metal flow to achieve antenna reconstruction; the power supply and heat dissipation system provides a stable power supply for the entire system.

2. The multi-polarization antenna array for unmanned aerial vehicles based on liquid metal dynamic reconstruction according to claim 1 is characterized in that: It also includes a liquid metal antenna unit structure: Each liquid metal antenna unit adopts a three-layer structure design: The top layer is made of flexible polydimethylsiloxane (PDMS) and contains a network of microchannels for the flow of liquid metal. The microchannels are designed in a specific pattern to form the antenna shape required for different polarization modes. Middle layer: electrode array layer, including electrostatic driving electrodes and conductivity detection electrodes arranged around the microchannel, used to control and monitor the flow of liquid metal; Bottom layer: control circuit layer, which includes micro pumps, pressure control valves and drive circuits to achieve precise control of the flow of liquid metal.

3. The multi-polarization antenna array for unmanned aerial vehicles based on liquid metal dynamic reconstruction according to claim 2 is characterized in that: It also includes the antenna array arrangement: the line units are arranged in a 4×4 matrix to form a planar array with an overall size of 50mm×50mm×5mm, and the entire array has a modular structure.

4. The multi-polarization antenna array for unmanned aerial vehicles based on liquid metal dynamic reconstruction according to claim 1 is characterized in that: The microfluidic driving system adopts a coordinated driving mode of electrostatic field and pressure field, including: Electrostatic drive unit: An electrode pair is set around the microchannel, and an electrostatic field is generated by applying voltage, thereby controlling the flow direction and speed of the liquid metal using the continuous electrowetting effect; Pressure drive unit: This includes a micropump and pressure control valve, providing the base pressure required for liquid metal flow. It works in conjunction with the electrostatic drive. This collaborative drive method is based on a dynamic surface tension gradient model that takes into account the effects of electric, pressure, and temperature fields, and is used to accurately calculate the total force required to drive the liquid metal. Pressure sensor: monitors the pressure distribution in the microchannel in real time, forms a closed-loop control, and improves control accuracy.

5. The multi-polarization antenna array for unmanned aerial vehicles based on liquid metal dynamic reconstruction according to claim 1 is characterized in that: It also includes the control system architecture, which adopts a hierarchical design: Low-level control: performs specific liquid metal flow control, including basic operations such as electrode voltage adjustment and pump pressure control; Middle-level control: converts antenna configuration instructions into specific control sequences and coordinates the work of each drive unit; Upper-layer control: Provides intelligent decision-making and selects antenna configuration based on environmental perception results.

6. A switching control method based on the antenna array according to claim 1, characterized in that: It includes state space reconstruction and action space optimization. The state space reconstruction includes: obtaining spectrum occupancy information of the current electromagnetic environment, polarization matching information between transmitting and receiving antennas, and current radiation direction information of the antenna array to form an environment perception vector.

7. The switching control method according to claim 6, characterized in that , the state space reconstruction defines the environment perception vector T as: Where T is the state vector at time t; the first term is the spectrum sensing component, δ(·) is the Dirac function, f is the center frequency of the UAV communication channel, and f i is the center frequency of the i-th in-band or adjacent interference signal detected, N is the total number of interference signals detected; the second term is the polarization mismatch, where E rx and E tx The electric field strength vectors of the receiving antenna and the transmitted signal are respectively. The calculation result of this item is a scalar (range [0,1]) that characterizes the degree of polarization alignment, with a value of 0 indicating perfect alignment and a value of 1 indicating orthogonal mismatch. The third item is the main lobe direction, and AF(θ) is the normalized array factor of the antenna array in the direction θ. This item gives the direction of maximum gain at the moment. The action space a t include: <h2 style=";text-align:left;direction:ltr">a<h2 style=";text-align:left;direction:ltr"> t <h2 style=";text-align:left;direction:ltr"> =[{b1,b2,…,b<h2 style=";text-align:left;direction:ltr"> m <h2 style=";text-align:left;direction:ltr">},{V1,V2,…,V<h2 style=";text-align:left;direction:ltr"> k <h2 style=";text-align:left;direction:ltr">}] Among them, b i ∈{0,1} is the binary on / off control signal for the i-th microchannel, with a total of m controllable channels; V j ∈[0,50V] is the voltage value applied to the jth driving electrode, with a total of k driving electrodes.

8. The switching control method according to claim 6, wherein: It also includes the spatiotemporal attention deep Q network (STA-DQN), the network structure is: Input layer: St∈R128; Spatiotemporal encoding layer: TransformerEncoder (4-head); Strategy layer: dual-stream fully connected network; Output layer: Q(s,a)eR2"x10; The loss function is: Among them, r is the immediate reward of environmental feedback, s and s' are the current and next states respectively, a and a' are the current and next actions respectively, γ∈[0,1] is the reward discount factor, Q(s,a) is the action value output by the current Q network, Q target (s',a') is the action value output by the target Q network, λ is the regularization coefficient, D KL (π current ||π prior ) is the current policy π current The strategy π of the previous iteration prior The Kullback-Leibler divergence between is used to stabilize the training process.

9. The switching control method according to claim 6, wherein: It also includes polarization-spatial joint optimization: the polarization state is adjusted through a dynamic feedback mechanism to optimize the polarization matching degree. The calculation formula is: Where S = (S0, S1, S2, S3) is the Stokes vector of the transmitted signal, and S' = (S'0, S'1, S'2, S'3) is the Stokes vector of the signal effectively received by the receiving antenna. S0 and S'0 are the total intensity, S1, S2 and S'1, S'2 are the linearly polarized components, and S3 and S'3 are the circularly polarized components. The result of this formula is the dimensionless polarization matching efficiency.

10. The switching control method according to claim 6, wherein: It also includes a pattern synthesis algorithm: by introducing a multi-constraint convex optimization model, the array pattern is dynamically optimized. The calculation formula is: stw H a(θ d ) = 1 (main lobe constraint) w H a(θ j )=0(zero constraint) |w i |≤1(amplitude constraint) The frequency tuning equation is: Among them, L eff is the equivalent electrical length of the antenna after reconstruction, L0 is the initial equivalent electrical length, α=0.15 is the experimentally calibrated liquid metal deformation coefficient, ΔV is the change in driving voltage, and V0 is the initial or reference driving voltage.

Images