A low-delay millimeter wave transmission method for communication measurement and control of a drone swarm

Abstract

The application relates to the technical field of unmanned aerial vehicle (UAV) millimeter wave communication, and particularly discloses a low-delay millimeter wave transmission method for UAV group communication measurement and control, which comprises the following steps: constructing a parallel conventional service data transmission channel and a physical layer bypass transmission channel, wherein the conventional service data transmission channel comprises a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module. According to the application, a bypass transmission channel independent of conventional data is constructed at the physical layer, so that emergency measurement and control instructions directly bypass time-consuming channel coding and interleaving modules, the waiting time of data processing is greatly reduced at the bottom hardware level, the instructions are directly output to the flight control system of the UAV after being parsed, and the instantaneous dynamic response requirement of the UAV when facing a sudden obstacle or emergency tactical action is met.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) millimeter-wave communication technology, specifically a low-latency millimeter-wave transmission method for UAV swarm communication and control. Background Technology

[0002] UAV swarms widely employ millimeter-wave communication systems for data exchange when performing complex tasks. To combat the high path loss and susceptibility to physical blockage inherent in the high-frequency millimeter-wave band, existing communication systems incorporate standardized channel coding and interleaving modules in the baseband processing stage. All control commands and routine service data are treated equally, undergoing complex error correction coding calculations, interleaving and packetization, and centralized channel scheduling at the upper layer before being sent to the radio frequency interface for over-the-air transmission.

[0003] In highly dynamic operational scenarios requiring rapid evasion or quick formation changes, existing transmission architectures face a trade-off between command transmission reliability and extremely low latency:

[0004] To ensure the success rate of emergency telemetry and control commands in complex air interface environments, traditional methods must rely on deep channel coding and complex centralized air interface scheduling, which inevitably introduces a long signal processing delay. If the scheduling layer is directly abandoned for data concurrency in order to shorten the delay, multiple UAV nodes will cause serious air interface channel collisions when issuing emergency commands at the same time, resulting in the destruction or loss of critical telemetry and control commands in the conflict. Summary of the Invention

[0005] This invention aims to at least partially solve one of the technical problems in related technologies. Therefore, the objective of this invention is to propose a low-latency millimeter-wave transmission method for communication and control of unmanned aerial vehicle (UAV) swarms, thereby achieving a balance between low latency and high reliability in communication.

[0006] To achieve the above objectives, a first aspect of the present invention proposes a low-latency millimeter-wave transmission method for communication and control of unmanned aerial vehicle (UAV) swarms, comprising:

[0007] Construct a parallel regular service data transmission channel and a physical layer bypass transmission channel, wherein the regular service data transmission channel includes a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module;

[0008] A deterministic frame structure is configured for the physical layer bypass transmission channel, wherein the front of the deterministic frame structure includes an arbitration-specific preamble field carrying air interface arbitration information;

[0009] In listening mode, the transmitter decodes the arbitration preamble field of the received signal, performs distributed priority arbitration according to the preset arbitration rules, and after the arbitration is successful, it preempts the physical layer bypass transmission channel to send high-priority measurement and control commands.

[0010] The receiving end detects the arriving signal through a multi-channel parallel preamble detection circuit, latches only the highest priority high-priority telemetry and control command, and outputs it to the flight control system of the UAV node.

[0011] Extract the environmental and link status features of the UAV node, and perform cross-layer linkage based on the status features and the preset arbitration rules to dynamically adjust the arbitration parameters and transmission strategies in the arbitration rules.

[0012] To achieve the above objectives, a second aspect of the present invention provides a low-latency millimeter-wave transmission system for communication, measurement, and control of unmanned aerial vehicle (UAV) swarms, comprising:

[0013] The channel construction module is used to construct a parallel regular service data transmission channel and a physical layer bypass transmission channel, wherein the regular service data transmission channel includes a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module;

[0014] A frame structure configuration module is used to configure a deterministic frame structure for the physical layer bypass transmission channel. The front of the deterministic frame structure includes an arbitration-specific preamble field that carries air interface arbitration information.

[0015] The transmitting arbitration module is used to decode the arbitration-specific preamble field of the received signal in monitoring mode, perform distributed priority arbitration according to preset arbitration rules, and preempt the physical layer bypass transmission channel to send high-priority measurement and control commands after the arbitration is successful.

[0016] The receiver detection output module is used to detect the arriving signal through a multi-channel parallel preamble detection circuit, latch only the highest priority high-priority telemetry and control command, and output it to the flight control system of the UAV node.

[0017] The cross-layer linkage adjustment module is used to extract the environmental and link status characteristics of the UAV node, and perform cross-layer linkage based on the status characteristics and the preset arbitration rules to dynamically adjust the arbitration parameters and transmission strategies in the arbitration rules.

[0018] To achieve the above objectives, a third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the aforementioned low-latency millimeter-wave transmission method for UAV swarm communication and control.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] This invention constructs a bypass transmission channel independent of conventional data at the physical layer, enabling emergency telemetry and control commands to bypass time-consuming channel coding and interleaving modules. This significantly reduces data processing latency at the underlying hardware level, allowing commands to be directly output to the flight control system of the associated UAV after parsing. This meets the instantaneous maneuverability requirements of UAVs when facing sudden obstacles or emergency tactical maneuvers. Simultaneously, by utilizing a deterministic frame structure and distributed priority arbitration rules, the conflict between abandoning complex scheduling and multi-node air interface collisions is cleverly resolved. This allows UAV nodes to autonomously determine transmission rights simply by detecting multiple preambles at the hardware level, ensuring that the highest priority emergency telemetry and control commands can always preempt the channel without hindrance during dense concurrent communication in a swarm. Combined with a cross-layer linkage environmental perception mechanism and dynamic correction strategy, this invention achieves a reasonable balance between low latency and high reliability in harsh actual operating environments. Attached Figure Description

[0021] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals refer to the same parts. Wherein:

[0022] Figure 1 This is a flowchart illustrating the low-latency millimeter-wave transmission method for communication and control of unmanned aerial vehicle swarms provided by the present invention.

[0023] Figure 2 This is a comparison curve of the signal spectrum before and after power amplifier nonlinear distortion compensation in the low-latency millimeter-wave transmission method for UAV swarm communication and control provided by the present invention.

[0024] Figure 3 This is a graph of the peak output data of sliding cross-correlation of multiple preamble parallel detection in the low-latency millimeter-wave transmission method for UAV swarm communication and telemetry provided by the present invention.

[0025] Figure 4 This is a three-dimensional response surface plot of rotor pitch angle and angular velocity to link blocking time margin in the low-latency millimeter-wave transmission method for UAV swarm communication and control provided by the present invention.

[0026] Figure 5 This is a heat map of the baseband Doppler frequency shift limit distribution under different relative maneuvering speeds and spatial angles in the low-latency millimeter-wave transmission method for UAV swarm communication and telemetry provided by the present invention.

[0027] Figure 6 This invention provides a comparison of the received signal constellation diagrams before and after baseband Doppler frequency offset despin compensation in the low-latency millimeter-wave transmission method for UAV swarm communication and telemetry provided by this invention.

[0028] Figure 7 This is a peak value estimation diagram of the spatial angle of arrival spectrum based on the baseband phase difference of the phase array antenna in the low-latency millimeter-wave transmission method for UAV swarm communication and telemetry provided by the present invention.

[0029] Figure 8 This is a schematic diagram illustrating the implementation of the low-latency millimeter-wave transmission system for communication and control of unmanned aerial vehicle swarms provided by the present invention.

[0030] Figure 9 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0031] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0032] The following description, with reference to the accompanying drawings, illustrates a low-latency millimeter-wave transmission method, system, and electronic device for unmanned aerial vehicle (UAV) swarm communication, telemetry, and control according to embodiments of the present invention.

[0033] Example 1:

[0034] This embodiment provides a low-latency millimeter-wave transmission method for communication and control of unmanned aerial vehicle (UAV) swarms. It relies on a millimeter-wave software radio architecture mounted on the UAV node and covers the underlying configuration of the baseband digital signal processing core, radio frequency front-end, and field-programmable gate array (FPGA).

[0035] When drone swarms perform highly dynamic flight missions, such as dense formations, high-speed shuttles, or collaborative operations, the relative spatial positions and channel states between nodes change rapidly. The solution in this embodiment aims to solve the timeliness problem faced by high-priority telemetry and control commands when transmitting them over the air interface.

[0036] Specifically, such as Figure 1 As shown, the method in this embodiment includes the following:

[0037] This method first constructs a parallel regular service data transmission channel and a physical layer bypass transmission channel in the underlying logic of the transceiver. The regular service data transmission channel includes a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module.

[0038] In practical applications, the regular business data transmission channel is mainly responsible for processing broadband services such as high-definition video streams from UAVs and large-capacity point cloud mapping data. Since this type of data is sensitive to bit error rate but not to latency, this channel is configured with a complete forward error correction algorithm (such as low-density parity-check codes or polar codes) forming a channel coding module, as well as block interleaving or convolutional interleaving modules to break up burst errors. Data encoding and interleaving typically consume a large number of clock cycles, introducing processing delays on the order of milliseconds. The physical layer bypass transmission channel, on the other hand, is a hardware direct path specifically designed to carry telemetry and control commands such as emergency obstacle avoidance and attitude takeover. Bypassing means that in the FPGA's data flow graph, the data flow of this channel directly skips the memory arrays and computational logic resources of the encoders and interleavers, retaining only the most basic synchronization and framing modules, thereby compressing the baseband processing latency to the nanosecond level.

[0039] For example, to ensure the high efficiency of the bypass channel, the preemption of the physical layer bypass transmission channel to send a high-priority measurement and control command specifically includes: interrupting the current service data transmission in the regular service data transmission channel and saving the transmission breakpoint address of the current service data; inserting a breakpoint identifier into the current service data stream and sending the high-priority measurement and control command to the Serdes interface for transmission; and after the high-priority measurement and control command is transmitted, resuming the interrupted service data transmission according to the transmission breakpoint address.

[0040] At the hardware implementation level, this preemption mechanism is jointly implemented by a multiplexer and a finite state machine deployed in the data link control layer. When the master control system issues a high-priority measurement and control command, the finite state machine immediately triggers an interrupt signal, causing the data pointer in the regular business data channel to stop incrementing, and temporarily storing the physical memory address of the current start and end pointers (i.e., the transmission breakpoint address) in static random access memory.

[0041] Subsequently, the hardware logic automatically appends a specific bit sequence as a breakpoint marker to the end of the unsent regular data stream, indicating changes in subsequent data segments to the receiving end. Next, the multiplexer switches the data source, directly routing high-priority telemetry and control command packets to the transmitter's SerDes interface. The SerDes interface converts parallel baseband digital signals into a high-speed serial bit stream and feeds it to the digital-to-analog converter. After the high-priority command is fully processed by the SerDes interface, the finite state machine reactivates the regular channel, retrieves the previously saved transmission breakpoint address from memory, and continues to drive the serialized output of regular service data, thus achieving seamless insertion of telemetry and control commands without causing broadband data loss.

[0042] Optionally, to ensure that the data stream in the bypass channel is quickly identified by the receiving end, this scheme configures a deterministic frame structure for the physical layer bypass transmission channel. The front of the deterministic frame structure includes an arbitration-specific preamble field carrying air interface arbitration information. The total length of the deterministic frame structure is a first preset length, and the length of the arbitration-specific preamble field is a second preset length, wherein the second preset length is less than the first preset length.

[0043] It is necessary to clearly define a deterministic frame structure, which refers to a frame in which the number of bits occupied by each field and their order are fixed. This allows the receiving end to read key information in real time by slicing the data using a fixed offset, without waiting for the entire data frame to be received or parsing complex packet length identifiers. The deterministic frame structure includes, in sequence: a bit synchronization header and a frame synchronization header for signal synchronization; a priority identifier for defining instruction priority; a node unique identifier for globally uniquely identifying the sending node; a collision avoidance random number generated in real time by hardware; an instruction length identifier and instruction data carrying the actual measurement and control content; a hardware checksum for parallel verification; and an acknowledgment request identifier for indicating feedback from the receiving end.

[0044] In actual protocol mapping, the first preset length is limited to a compact total time domain span, for example, 95 bits, to control the channel occupancy time of a single transmission; the second preset length is limited to the prefix length covering the synchronization header, priority, node identifier, and random number, for example, 37 bits. The bit synchronization header used for signal synchronization employs a high-frequency alternating level sequence to help the receiver's phase-locked loop quickly acquire the clock frequency; the frame synchronization header uses Barker codes with steep autocorrelation characteristics to mark the starting boundary of valid data on the time axis. The priority identifier divides the UAV's commands into multiple digital levels, with smaller values ​​indicating a higher urgency of maneuver intervention. The unique node identifier is a statically assigned network interface card (NIC) physical address based on the network configuration before the fleet takes off. Collision avoidance random numbers are directly interfaced to a thermal noise-based true random number generator inside the FPGA, ensuring that the values ​​are different for each frame to cope with air interface collisions in extreme cases. The command length identifier defines the specific payload size of the following command data. The hardware checksum is generated using a cyclic redundancy check (CRC) algorithm for all the aforementioned fields, and is used to detect discrete flip errors caused by spatial links. The acknowledgment request flag is a single-bit switch; when it is set, the receiving end is required to return an acknowledgment packet after successful verification.

[0045] Specifically, to avoid severe channel congestion caused by multiple nodes transmitting simultaneously in a high-dynamic cluster, the transmitter decodes the arbitration-specific preamble field of the received signal in monitoring mode, performs distributed priority arbitration according to preset arbitration rules, and preempts the physical layer bypass transmission channel to send high-priority measurement and control commands after successful arbitration; the preset arbitration rules include a three-level arbitration mechanism executed sequentially:

[0046] Level 1 Priority Comparison: When a signal with a higher priority than the instruction to be sent by this node is detected in the air, this node performs a backoff operation;

[0047] Level 2 node identifier comparison: When a signal with the same priority as the instruction to be sent by this node is detected in the air, the unique node identifiers of the two are compared, and the node with the smaller value obtains the right to send;

[0048] Level 3 random number comparison: When the priority and the node's unique identifier are the same, the random number is compared to avoid conflict, and the node with the smaller value gets the right to send.

[0049] The practical significance of this distributed arbitration lies in the fact that drone swarms do not need to rely on a centralized base station or cluster head to allocate time slots. When all nodes have a transmission requirement, they first listen to the signal spectrum in the current space through their antenna arrays. Within a very short listening window, the node hardware extracts the arbitration-specific preamble that is floating in the air. When entering the first level of arbitration, the node sends the parsed air priority value and the priority value of the local command to be sent to the comparator unit. If the air priority is higher (smaller value), the node's transmission state machine enters sleep mode and performs a backoff operation, that is, it calculates a delay time based on the duration of the air signal and remains silent during this time period, yielding the channel.

[0050] If both have equal priorities, Level 2 arbitration is triggered. At this point, the nodes compare their physical layer identities, i.e., their unique identifiers. Since the unique identifiers of nodes across the entire network are non-overlapping, the node with the smaller value has higher physical execution rights. However, in complex electromagnetic reflection environments or under malicious replay attacks, a node might receive an identifier identical to its own. Therefore, a Level 3 arbitration mechanism is introduced. Because the local collision avoidance random number is generated instantly by a hardware true random number generator before each transmission, the probability of two nodes generating the same random number is extremely low. This is used as the final decision criterion, comparing the values ​​again. The node with the smaller value breaks its local silence and immediately enters the transmission state. All three levels of comparison are implemented in the gate-level combinational logic of the FPGA, typically taking less than 100 nanoseconds, ensuring low latency.

[0051] It should also be noted that in the millimeter-wave band, RF power amplifiers exhibit significant memory effects and nonlinear distortion during high-frequency switching. For example, the transmitting end performs RF direct-drive compensation while sending high-priority measurement and control commands, specifically including:

[0052] The drain current of the power amplifier in the transmitter is collected in real time as the sampling current value; according to the current link blocking state, the target mapping table is matched with several pre-stored current and bias voltage mapping tables; the sampling current value is input into the target mapping table to obtain the bias voltage correction amount, and the bias voltage correction amount is applied to the bias terminal of the RF driver amplifier in the transmitter through digital-to-analog conversion.

[0053] In practical radio frequency (RF) circuit applications, the high-speed movement of drones causes dynamic pulling of the antenna load impedance, which in turn causes fluctuations in the drain current of the power amplifier. Without intervention, these fluctuations can lead to spectral broadening of the transmitted signal and adjacent channel power leakage, reducing the demodulation success rate at the effective receiver. Therefore, this embodiment configures a high-speed current sampling module in the RF front-end to acquire the drain current in real time. Simultaneously, multiple sets of empirical lookup tables (i.e., mapping tables) for different external temperatures and link fading conditions (such as line-of-sight and non-line-of-sight) are pre-programmed into the baseband's read-only memory.

[0054] To quantify the compensation process, this solution implements a bias adjustment algorithm based on polynomial fitting in hardware, and its specific calculation formula is as follows:

[0055] ;

[0056] In the formula: This represents the calculated bias voltage correction to be applied to the RF driver amplifier, expressed in volts. This indicates the base reference bias voltage for the initial hardware configuration; This represents the i-th order voltage transformation weight coefficient corresponding to the currently matched target mapping table; This indicates the real-time feedback of the sampled current value collected by the sensor; This represents the reference quiescent current of the power amplifier under standard load conditions.

[0057] In practical applications, this formula works by calculating the deviation between the actual current and the ideal quiescent current, and then using first- to third-order polynomial weights for nonlinear expansion to dynamically determine the bias voltage correction. This correction is then converted to an analog level via a digital-to-analog converter, dynamically adjusting the operating bias point of the RF amplifier to maintain it within a near-linear operating range, thus avoiding clipping distortion of the millimeter-wave signal due to amplifier saturation.

[0058] like Figure 2 The graph shows a comparison of the signal spectrum before and after nonlinear distortion compensation for the power amplifier. The horizontal axis represents frequency shift in megahertz, and the vertical axis represents power spectral density in decibels and milliwatts per hertz.

[0059] The red pre-compensation spectrum curve in the figure shows a significant spectrum broadening phenomenon. Outside the main frequency band of ±100 MHz, its out-of-band transmit power spectral density remains at a high level of around -30 dBmW / Hz. This objectively indicates that when the high-speed movement of the UAV causes dynamic traction of the load impedance, the nonlinear distortion of the power amplifier produces serious adjacent channel power leakage.

[0060] After RF direct-drive compensation, the blue compensated spectrum curve shows a significant convergence effect. By real-time acquisition of the drain current and application of bias voltage correction, the amplifier was effectively tuned to a near-linear operating region. In this state, the main frequency band shape remained stable, while the out-of-band frequency domain shoulder energy was effectively suppressed, the adjacent channel leakage power spectral density rapidly decreased to below -50 dBmW / Hz, and the out-of-band attenuation was improved by approximately 20 dB.

[0061] This data and waveform transformation results confirm that the dynamic current mapping compensation scheme can effectively suppress clipping distortion of millimeter-wave signals caused by amplifier saturation, ensure the spectral purity of the transmitted signal, and provide a physical guarantee at the radio frequency layer for the reliable transmission of high-priority telemetry and control commands in complex air interface environments.

[0062] Subsequently, the receiving end detects the arriving signal through a multi-channel parallel preamble detection circuit, latches only the highest priority telemetry and control command, and outputs it to the flight control system of the corresponding UAV node; the receiving end performs multi-preamble parallel detection, specifically including:

[0063] The received signal is deserialized into a baseband bitstream, and the baseband bitstream is synchronously input into multiple independent preamble detection circuits, each of which corresponds to a priority level. The correlation value between the baseband bitstream and the preamble corresponding to each priority level is calculated independently. When the correlation value exceeds the first correlation threshold, the latch signal of the corresponding channel is triggered, and the priority levels of each channel are compared in real time. Only the latch state of the highest priority channel is maintained to extract the instruction, while all low priority channels are closed.

[0064] In the receiving baseband architecture, the high-frequency analog signal received by the antenna is down-converted and then converted from analog to digital, before being recovered into a parallel baseband bitstream via the SerDes interface. Given that signals from different nodes may be superimposed simultaneously in the air interface environment, this solution instantiates multiple sliding correlator hardware resources within the FPGA, equal to the number of priority levels. Each preamble detection circuit has a pre-set standard preamble template for the corresponding priority level.

[0065] For example, the hardware calculation process for the correlation value strictly follows the following discrete cross-correlation formula:

[0066] ;

[0067] In the formula: This represents the sliding correlation value calculated at the current sampling time; Indicates the total length of the sequence of sample points for the standard preamble; This indicates the current discrete-time index number of the receiver; This indicates the baseband bitstream after deserialization at the sliding window time. The specific complex sampling amplitude; This indicates that the preamble sequence of a specific priority level is in the index. The conjugate mapping value at that location.

[0068] In the actual demodulation process, the baseband data stream passes through a sliding window sequentially, and each shift involves parallel multiplication and addition operations according to the formula above. This is true if and only if the input baseband bit stream contains the corresponding priority preamble. The output amplitude will show a spike. When the detection circuit determines that the spike exceeds the preset hardware judgment benchmark (i.e., the first correlation threshold), a latch level is generated. If multiple correlators are triggered simultaneously, the hardware priority encoder intervenes, forcibly selecting the physical channel representing the highest priority, and cutting off the subsequent signal input buffer of other channels to ensure that the underlying processing resources are fully dedicated to serving the most urgent instructions.

[0069] like Figure 3 The graph shows the peak output data of sliding cross-correlation for parallel detection of multiple preambles. The horizontal axis represents the sampling sequence index (in units), and the vertical axis represents the cumulative cross-correlation amplitude (in quantization units).

[0070] In the diagram, the red curve represents the correlation value trend of the highest priority detection branch, the blue curve represents the correlation value trend of the medium priority detection branch, the green curve represents the correlation value trend of the lowest priority detection branch, and the black dashed line represents the hardware-preset first correlation threshold.

[0071] As can be seen from the waveform transformation, when the sampling sequence index reaches around 500, the red highest priority detection branch curve rises rapidly, producing a sharp correlation peak with an amplitude of about 40. This peak significantly breaks through the black first correlation threshold dashed line set at 30.

[0072] Meanwhile, the blue and green non-match priority detection branch curves consistently maintained low-level random fluctuations around an amplitude of 10, failing to reach the threshold benchmark.

[0073] This data objectively reflects the actual physical effect of each circuit independently performing parallel multiplication and addition operations when the baseband data stream is synchronously input into multiple independent preamble detection circuits. Matched channels with high phase consistency can effectively accumulate signal energy and trigger the latch signal of the corresponding channel, while mismatched channels and environmental background noise cancel each other out due to phase orthogonality or random dispersion. This demonstrates that this multi-channel parallel detection mechanism can accurately capture and identify target priority commands in complex air interface environments, ensuring the sensitivity and anti-interference capability of the underlying signal identification in the telemetry and control link.

[0074] Optionally, after the receiving end latches the highest priority telemetry and control command, it further includes: calculating the checksum of the latched received frame in parallel by hardware logic and comparing it with the hardware checksum of the transmitting end; when the comparison matches, directly outputting the command data to the flight control system; when the comparison does not match, discarding the received frame; if the high-priority telemetry and control command contains a response request identifier and the comparison matches, automatically generating a hardware-level response frame and feeding it back to the transmitting end through the physical layer bypass transmission channel.

[0075] After completing the relevant capture, the hardware-level cyclic redundancy check (CRC) module performs parallel shift-and-XOR operations on the received bits. Once the comparison is successful, the hardware controller uses a direct memory access mechanism, such as the AXI-Lite bus, to write the instruction data into a specific register address in the flight control system. This design completely eliminates the complex processes of traditional architectures that rely on CPU interrupts, protocol stack unpacking, and operating system context switching, achieving direct data transmission from the antenna to the controller. If the verification matches and the preceding parsed response request is obtained, the underlying logic will autonomously construct a simple acknowledgment packet containing its own node identifier and feed it back along the original transmission link to complete the closed-loop acknowledgment.

[0076] Building upon the low-latency hardware channel, this embodiment further introduces a cross-layer linkage mechanism between the underlying environmental perception and communication layers.

[0077] For example, the environmental and link state characteristics of the UAV node are extracted, and cross-layer linkage is performed based on the state characteristics and the preset arbitration rules to dynamically adjust the arbitration parameters and transmission strategies in the arbitration rules. Specifically, the extraction of the UAV node's environmental and link state characteristics and the cross-layer linkage include linkage with the rotor blocking prediction module: based on the UAV node's rotor parameters, relative position, and attitude angle, the link blocking state within the next first preset time period is predicted; when it is predicted that the link will be blocked by the rotor, the priority of all pending commands of this node is increased by a preset level, and the air interface listening window length is shortened; the system switches to a current and bias voltage mapping table pre-stored in the transmitter and applicable to non-line-of-sight environments, and the transmitter's transmission power is increased.

[0078] Because millimeter waves have a strong line-of-sight propagation dependency, the carbon fiber rotor blades of the UAV, when rotating at high speed, will periodically cut, absorb, and scatter the millimeter wave rays, causing severe rotor blockage fading. To address this, this method establishes a kinematics-based blockage prediction and evaluation model, which performs forward-looking calculations by collecting data on the mechanical state of the UAV.

[0079] For example, the prediction model relies on the following geometric and kinematic relationship formulas:

[0080] ;

[0081] In the formula: This represents the relative time margin for predicting future link congestion, calculated by the cross-layer linkage module. This represents the initial spatial angle deviation between the current UAV node's main beam line-of-sight transmission path and the rotor blade that is about to enter that path; This indicates the current real-time physical angular velocity of the rotor as read by the onboard electronic speed controller, in radians per second. This indicates the real-time pitch angle of the drone as fed back by the gyroscope of the flight control system.

[0082] In actual flight operations, this formula serves a predictive function. When the calculated... When the system falls into the danger zone, specifically within the first preset time period, it anticipates that the communication link will face deep fading of hundreds of milliseconds. At this point, it is crucial to deliver critical data before it is blocked. Therefore, the hardware logic proactively modifies the frame header configuration of the pending command, increasing its priority attribute. Simultaneously, it increases the success rate of preempting the channel by shortening the listening backoff time and invokes the radio frequency's preset non-line-of-sight high-power mapping table for high-power transmission, utilizing multipath reflection to achieve penetration transmission.

[0083] like Figure 4 This diagram displays a three-dimensional response surface plot of rotor pitch angle and angular velocity to link blocking time margin. The horizontal axis represents the rotor's physical angular velocity in radians per second, the vertical axis represents the flight pitch angle in degrees, and the vertical axis represents the link blocking time margin in milliseconds. The three-dimensional surface in the diagram gradually changes from dark blue to dark red, with the color intensity visually reflecting the numerical distribution of the time margin under different parameter combinations.

[0084] Observing the spatial variation of the curved surface reveals that when the flight pitch angle approaches 45 degrees and the rotor physical angular velocity is at a low level of 100 radians per second, the link blocking time margin reaches a relatively high level of about 14 milliseconds, which is shown as a red high value area in the figure.

[0085] As the rotor's physical angular velocity gradually climbs to the high range of 500 radians per second, regardless of changes in the flight pitch angle, the three-dimensional surface rapidly sinks and converges to the dark blue area. At this point, the link blocking time margin drops rapidly to around 2 milliseconds. This data trend illustrates in this invention that when a UAV performs high-speed, high-maneuver flight maneuvers, the rotor cutting-off time window for the communication link is greatly compressed.

[0086] The response surface objectively confirms the rationality of cross-layer blocking prediction based on kinematic parameters. Based on this nonlinear attenuation law, the system can trigger the shortening of the air interface listening window and switch the non-line-of-sight mapping table in advance within a very short time before the actual physical blockage of the link occurs, thereby ensuring the reliability of high-priority telemetry and control commands in penetrating transmission when facing rotor depth blockage.

[0087] It is also important to note that in dense drone swarms, link degradation caused by electronic interference or component aging is common. Specifically, the extraction of environmental and link state characteristics of drone nodes and cross-layer linkage also includes linkage with the physical layer anomaly detection module:

[0088] The system monitors the link synchronization status, sudden changes in received signal strength, and command error rate of the SerDes interface in the receiving end in real time, generating corresponding levels of abnormal status. When the attenuation of the received signal strength reaches the first attenuation threshold and is determined to be a level 1 abnormality, the priority of the command to be sent is increased and the transmission power of the transmitting end is increased. When the command error rate reaches the first error threshold and is determined to be a level 2 abnormality, the backup bypass transmission channel is activated to execute the primary and backup dual-transmission mode. When the link synchronization status is detected to be disconnected and is determined to be a level 3 abnormality, the main channel transmission is stopped, the command to be sent is completely switched to the backup bypass transmission channel, and re-node pairing and beam allocation are triggered.

[0089] The aforementioned linkage strategy constructs a tiered fault degradation and self-healing system. A sudden drop in Received Signal Strength (RSSI) usually indicates an increase in flight distance or slight air interface obstacle interference, which is defined as a Level 1 anomaly. In this case, it is only necessary to adjust the RF power of the transmitter within a range of minimal resource consumption and slightly increase the priority in the arbitration mechanism.

[0090] When the baseband demodulation error increases dramatically due to multipath effect or hostile frequency sweeping interference, causing the instruction bit error rate caused by checksum mismatch to exceed the preset security threshold (the first bit error threshold), it is defined as a level 2 anomaly. The system no longer relies on a single frequency or a single spatial polarization direction, but directly wakes up the backup transceiver resources and implements a primary and backup dual-transmission mechanism for high-priority telemetry and control instructions to be sent simultaneously and redundantly through two physical links.

[0091] If the SerDes interface completely loses its bit synchronization signal due to severe clock drift or hard damage to RF components, triggering a Level 3 anomaly, the main channel resource is deemed invalid. The cross-layer linkage module directly blocks the injection of new data into the failed module, activates the entire hot-standby bypass unit, and commands the UAV antenna array to re-search and establish a communication topology with adjacent reliable nodes through underlying signaling interaction, thereby ensuring the transmission reliability of the telemetry and control link in harsh environments.

[0092] Optionally, to further enhance the communication potential in the spatial domain, the extraction of environmental and link state features of UAV nodes and cross-layer linkage also includes linkage with the node beam joint optimization module:

[0093] The default priority of each node is dynamically adjusted periodically based on the historical collision probability and the frequency of emergency command transmission. Antenna beams with corresponding beamwidths are assigned to commands of different priorities, with the highest priority command assigned the first beamwidth, the medium priority command assigned the second beamwidth, and the lowest priority command assigned the third beamwidth. The first beamwidth is smaller than the second beamwidth, and the second beamwidth is smaller than the third beamwidth. The switching of the antenna beamwidth is directly triggered and controlled by the physical layer hardware based on the priority level of the command.

[0094] In the actual configuration of millimeter-wave phased array antennas, beamwidths of different sizes possess drastically different physical properties. Smaller beamwidths, such as the first beamwidth, have extremely high equivalent isotropic radiated power and strong resistance to lateral space interference due to highly concentrated energy, but their alignment requirements are extremely stringent. This solution assigns this high-gain configuration only to the highest priority commands. For commands carrying secondary telemetry or status broadcasts, a wider beamwidth (the third beamwidth) is selected, sacrificing some transmission distance for a larger coverage area. This priority-based beamwidth fast lookup switching is directly driven by the underlying programmable gate array pins to the phase shifter network, eliminating the hysteresis of configuring RF registers in the upper-layer protocol stack.

[0095] Furthermore, the importance of each drone node within the group is not static. Changes in the roles of the lead drone, wingman, or node performing reconnaissance or strike missions require corresponding changes in their underlying communication privileges.

[0096] To enable nodes to adapt to dynamic networks, the system periodically executes a priority update algorithm based on historical evaluations, with adjustments made according to the following formula:

[0097] ;

[0098] In the formula: This represents the dynamically adjusted default priority scalar that is overwritten and updated in the basic register of this node after a round of periodic calculations. This indicates the initial priority value of the configuration issued by the preset planning table when the drone joins the network; This represents the penalty coefficient constant configured by the administrator for air interface collisions. This indicates the number of collisions recorded by this node within the historical statistical window during the past monitoring period, which involved air interface transmission but did not receive a verification response. This indicates the sliding time length for performing historical status statistics, i.e., the duration of the historical statistics window; An emergency weighting coefficient representing the current environmental urgency level of the assessment node; This indicates the frequency at which emergency intervention commands are sent by this node within the same monitoring period.

[0099] In practical operation, the logic of this formula is as follows: if a UAV frequently encounters air interface collisions recently, or issues emergency obstacle avoidance and attack commands at high frequency due to being on the front line of combat, it indicates that the node is in a severe or critical mission state. The last two factors of the formula will apply the dynamic deduction to its base priority, so that when it executes the above distributed priority arbitration again, it can obtain a smaller priority value benchmark, establish an absolute inherent advantage in the first-level comparison, and obtain more air interface preemption rights.

[0100] In summary, the method described in detail in this embodiment breaks away from the traditional approach in existing technologies that relies on the Media Access Control (MAC) layer for complex channel detection and centralized scheduling. The drawback of existing technologies is that when the application scenario shifts to a swarm of UAVs with extremely high node density and rapid relative movement, traditional time-division multiple access or carrier sense architectures can lead to uncontrollable queue waiting and significant signaling overhead.

[0101] This embodiment employs a physical layer dual-channel decoupling design, separating regular services from bypass direct access, thus avoiding the computational time consumption of error correction coding at the source. Combined with a bit-level simplified deterministic frame structure and a purely hardware-implemented distributed three-level arbitration circuit, nanosecond-level channel awareness and preemption are achieved.

[0102] Meanwhile, taking into full account the physical characteristics of the millimeter-wave band, the hardware-level nonlinear compensation and RF parameter configuration are deeply coupled and adjusted across layers with the aircraft's kinematic attitude and rotor blockage status. This overall technical solution ensures that, in the face of rapidly changing battlefield or industrial operating environments, the critical telemetry and control communications used by UAVs to maintain life and mission baseline are no longer constrained by broadband data queuing and centralized scheduling bottlenecks. It exhibits comprehensive technical advantages such as congestion resistance, blockage resistance, and extremely low transmission latency, possessing extremely high practical applicability and industrialization value.

[0103] Example 2:

[0104] When unmanned aerial vehicle (UAV) swarms perform tasks such as formation flying, high-speed evasion, or air combat, the high-speed relative motion between transmitting and receiving nodes induces a significant Doppler frequency shift in the millimeter-wave band. This frequency shift causes a rapid phase rotation in the constellation diagram of the baseband signal, preventing the preamble detection circuit at the receiver from obtaining sufficient integration energy, thus leading to the serious problem of missed detection of high-priority telemetry and control commands. To address this, this embodiment provides a Doppler frequency shift pre-compensation mechanism based on kinematic parameter cross-layer feedforward.

[0105] Specifically, the extraction of environmental and link state features of UAV nodes and the cross-layer linkage also includes linkage with the Doppler frequency offset pre-compensation module:

[0106] The 3D velocity vector and acceleration from the flight control system are extracted in real time as kinematic features. In the hardware architecture of an unmanned aerial vehicle (UAV), the flight control system is the core hub for sensing the physical motion state of the aircraft. This system is typically equipped with an inertial measurement unit (IMU) composed of a microelectromechanical system (MEMS) and a high-precision real-time dynamic differential GPS receiver. To obtain extremely low-latency physical attitude data, this solution does not poll data in the application layer software. Instead, at the underlying hardware driver level, the baseband processing unit reads the navigation output data fused by the extended Kalman filter algorithm from within the flight control system in real time via a dedicated internal bus.

[0107] The navigation output data mentioned above includes the UAV's independent three-axis velocity components and corresponding three-axis acceleration components in the northeast coordinate system. These components together constitute the kinematic characteristics representing the node's current maneuvering trend. Since this data is not encapsulated by the operating system's protocol stack, the hardware bus delay from the flight control system to the baseband Doppler frequency offset pre-compensation module is strictly controlled within the microsecond range, providing valuable time margin for subsequent baseband signal processing.

[0108] Optionally, after acquiring the aforementioned underlying data, the system calculates the relative velocity extreme range between the current node and the expected communication node in real time based on the kinematic characteristics and the position and velocity information of neighboring nodes, and maps the relative velocity extreme range to the baseband Doppler frequency shift boundary. During the collaborative networking process of UAV swarms, each node broadcasts its own kinematic characteristics in periodic heartbeat packets in the regular service data transmission channel. After decoding these heartbeat packets, the receiving node establishes a local air interface topology maintenance table in its local cache. When the system anticipates receiving a high-priority telemetry and control command from a specific node, the Doppler frequency offset pre-compensation module first extracts the velocity vector and acceleration vector of the expected communication node in the topology maintenance table, and combines it with the 3D velocity vector and acceleration extracted in real time by the current node to perform a three-dimensional spatial vector difference operation. Since the heartbeat packet broadcast has a fixed time interval, and the acquired neighboring node data has a slight time delay, the calculation module uses the acceleration vectors of both parties to perform a first-order linear extrapolation to calculate the maximum approach velocity and maximum departure velocity that may occur between the two parties within the current detection window. These two scalar values ​​together constitute the relative velocity extreme range.

[0109] Subsequently, the Doppler frequency shift pre-compensation module needs to transform the motion boundary of the physical space into frequency constraints in the radio frequency communication domain. The process of mapping the relative velocity extreme range to the baseband Doppler frequency shift boundary is achieved through the following underlying conversion formula:

[0110] ;

[0111] In the formula: This represents the absolute value of the maximum limit of the baseband Doppler frequency shift boundary after mapping calculation, in Hertz; This represents the velocity scalar with the largest absolute value within the range of extreme relative velocities calculated in the preceding steps, expressed in meters per second. It represents the equivalent propagation speed constant of electromagnetic waves in the air medium of the current working environment, and the unit is meters per second; This indicates the millimeter-wave radio frequency center carrier frequency configured in the current drone transceiver, measured in Hertz. This represents the relative azimuth angle between the receiving node and the intended sending node in three-dimensional space. This angle is calculated based on the three-dimensional spatial coordinates of both parties.

[0112] This formula allows the system to accurately assess the maximum distortion effect of the current mechanical motion state on the baseband signal frequency, thus providing a quantitative reference for subsequent baseband compensation.

[0113] like Figure 5A heatmap showing the baseband Doppler shift limit distribution under different relative maneuvering speeds and spatial angles is presented. The horizontal axis represents the extreme values ​​of relative maneuvering speeds in meters per second, and the vertical axis represents the spatial angle in degrees. The color scale on the side of the graph represents the specific value of the baseband Doppler shift limit in kilohertz.

[0114] As can be seen from the color of the thermal distribution and the trend of data change in the figure, when the physical angle between the two sides of the space is close to zero degrees and the extreme value of the relative maneuver speed reaches 80 meters per second, the thermal map shows a dark red area representing high frequency bias. At this time, the baseband Doppler frequency shift limit exceeds 16 kHz.

[0115] As the spatial physics angle increases or the relative maneuvering velocity extreme value decreases, the color of the heatmap gradually transitions to a deep blue, representing low-frequency bias. When the spatial physics angle approaches 90 degrees, regardless of how drastically the relative maneuvering velocity extreme value changes, the baseband Doppler frequency shift limit converges to the deep blue state of 0 kHz.

[0116] The color and data distribution characteristics of this two-dimensional plane intuitively reflect that when a drone swarm performs high-dynamic maneuvering operations such as high-speed oncoming flight, due to the extremely short wavelength of millimeter waves, huge carrier frequency drift that exceeds the phase tolerance threshold of conventional baseband hardware is easily generated between the transmitting and receiving nodes.

[0117] This heatmap objectively demonstrates the physical layer communication loss-of-lock risk boundary caused by Doppler frequency offset in high dynamic scenarios, thus proving the necessity and rationality of introducing a multi-branch parallel frequency offset search strategy based on kinematic feature feedforward in the underlying baseband processing of this system to estimate the frequency offset boundary in advance and perform multi-path parallel pre-compensation.

[0118] It is important to note that when the absolute value of the baseband Doppler frequency shift boundary exceeds the preset phase tolerance threshold, a multi-branch parallel frequency offset search strategy is triggered for the preamble detection circuit detecting the highest priority measurement and control command. In the working principle of the baseband correlator, the accumulation of correlation peaks depends on the consistency of the phase between the local preamble template and the received signal. When a Doppler frequency offset exists, the phase of the received signal rotates linearly with time. If, within the entire integration time window of the preamble, this accumulated phase deviation exceeds a specific critical point, the peak energy output by the correlator will decay sharply, preventing the hardware latch from being triggered. The phase tolerance threshold here refers to the maximum acceptable accumulated phase deflection tolerance preset by the system based on the preamble sequence length and the physical limits of the correlator's detection sensitivity hardware.

[0119] When the UAV is hovering or cruising at low speed, the calculated Doppler frequency shift boundary value is extremely small and does not reach the phase tolerance threshold. The system only maintains the operation of the single-channel preamble detection circuit as in Example 1 to reduce the dynamic power consumption within the field-programmable gate array. However, when the UAV enters a high-speed maneuvering countermeasure posture, once the baseband Doppler frequency shift boundary exceeds the phase tolerance threshold, the single-channel detection circuit will face a high risk of missed detections. At this time, the system hardware controller will directly pull up the enable signal and forcibly activate the multi-branch parallel frequency offset search strategy, which consumes a large amount of multiplier resources, for the highest priority telemetry and control command channel, sacrificing space resources for extremely high reception reliability under dynamic conditions.

[0120] For example, the multi-branch parallel frequency offset search strategy specifically involves: after deserializing the baseband signal at the receiving end, dynamically generating a set of discrete local compensation frequency values ​​based on the baseband Doppler frequency shift boundary; and performing hardware parallel multiplication and addition operations on the baseband data stream and the conjugate complex stream of the discrete local compensation frequency values ​​to generate multiple phase pre-compensated baseband branch streams.

[0121] In the specific execution of this strategy, the hardware logic uses the calculated baseband Doppler frequency shift boundary as the upper and lower limits, and divides multiple frequency grid nodes at equal intervals within the allowed frequency offset range according to the frequency step resolution determined by the main lobe width of the correlator. Each frequency grid node corresponds to a discrete local compensation frequency value.

[0122] Subsequently, multiple numerically controlled oscillator arrays within the system generate conjugate complex waveform data streams with opposite rotation directions in real time based on these locally compensated frequency values. The raw baseband data stream output from the Serdes interface is copied into multiple identical copies after entering the field-programmable gate array. Each data copy is input into an independent hardware complex multiplier, where it undergoes sample-by-sample multiplication and accumulation operations with the conjugate complex stream generated by the corresponding channel. This physical-level complex multiplication process essentially artificially performs frequency pre-deflection operations on the raw baseband data to varying degrees, causing the data stream, which originally carried unknown Doppler frequency offsets, to be discretized and despinned, thereby generating multi-channel phase-compensated baseband branch streams containing various pre-compensation possibilities.

[0123] Specifically, after generating multiple branches, the multiple phase-compensated baseband branch streams are synchronously input into the corresponding highest-priority preamble detection circuit for parallel correlation calculation. When the correlation value of any phase-compensated baseband branch stream exceeds the second correlation threshold, the target compensation frequency value that generated the correlation value is extracted, and the target compensation frequency value is used to perform baseband phase despinning on subsequent instruction data. To achieve microsecond-level ultra-fast detection, the system is configured with multiple parallel sliding correlator arrays. Each phase-compensated baseband branch stream is pushed into its dedicated correlator in real time and convolved with the local highest-priority preamble template.

[0124] The second correlation threshold here is a hardware decision reference voltage calibrated separately for the pre-compensation mechanism, used to filter false alarm levels caused by ambient background noise in multi-branch parallel computing. When an actual measurement and control command arrives, there will inevitably be a branch current whose compensation frequency is closest to the actual physical Doppler frequency offset, which can maximally offset the carrier rotation effect. This optimal branch current will generate the most concentrated correlation peak in the correlator, thus breaking through the second correlation threshold first. At this time, the hardware priority encoder immediately locks the winning channel and extracts the currently used discrete frequency from the corresponding numerically controlled oscillator configuration register, establishing it as the target compensation frequency value.

[0125] After obtaining the target compensation frequency value, the system will perform despinning compensation on the core data portion of the instruction. The specific despinning process depends on the following baseband compensation formula:

[0126] ;

[0127] In the formula: This represents the pure baseband symbol complex variable output to the subsequent analysis module after phase despinning compensation at discrete time points; This represents the original input baseband symbol complex variable with frequency offset that has not yet had its Doppler effect eliminated after deserialization at the receiver front end; This represents the determined target compensation frequency value extracted after triggering the second correlation threshold in the aforementioned parallel correlation calculation, in Hertz; This represents the duration of a single sampling clock cycle in a baseband analog-to-digital conversion system, in seconds. This indicates the discrete-time series index number of the payload portion of the currently being processed valid instruction data.

[0128] The practical application of this formula in the system is to directly construct the corresponding negative rotation complex factor using the precise frequency deviation value associated with the successfully captured preamble. This factor is then used by a hardware multiplier to eliminate residual frequency deviations in subsequent telemetry and control command payload data point by point. This operation ensures that core command data containing obstacle avoidance or control parameters can be accurately mapped to the decision region of the constellation diagram during demodulation, effectively preventing sudden errors in subsequent data.

[0129] like Figure 6 This diagram shows a comparison of the received signal constellation before and after baseband Doppler frequency offset despin compensation. The horizontal axis represents the in-phase components (in quantization units), and the vertical axis represents the quadrature components (in quantization units).

[0130] In the figure, the red dots represent the received signal before compensation, and the blue dots represent the received signal after compensation. Observing the red data points in the figure, it can be found that in an environment where the high-speed maneuvering of the UAV swarm causes a significant Doppler frequency shift, the original baseband signal after deserialization at the receiver cannot be stabilized in a fixed quadrant region due to the linear phase rotation accumulated over time. Instead, it is distributed in a circular pattern on the complex plane of the coordinate system. At this time, the receiver cannot make correct quadrant boundary determination of the signal, which easily leads to burst decoding errors.

[0131] After applying a multi-branch parallel frequency offset search strategy and obtaining the target compensation frequency value, the system performed a phase despinning compensation operation on the baseband data stream. At this time, the blue data points underwent a significant spatial position transformation. All the blue scattered points escaped the rotational divergence state and converged highly towards the four diagonals of the coordinate system, stably clustering around the four standard modulation decision centers near the positive and negative one quantization unit of the in-phase and quadrature components, respectively.

[0132] The significant change in the spatial distribution of this data point objectively confirms that the hardware multiplication and addition operation based on the negative rotation complex factor of the target compensation frequency value adopted by this system can effectively offset the millimeter wave carrier rotation effect, significantly reduce the residual frequency offset in the command data, and thus ensure the physical layer demodulation accuracy of high-priority telemetry and control commands, including obstacle avoidance parameters, under harsh conditions such as high-speed maneuvering and combat.

[0133] In terms of existing technology, traditional radio systems mostly employ closed-loop feedback tracking mechanisms such as Costas loops or phase-locked loops to cope with Doppler frequency shifts. These traditional closed-loop mechanisms typically require thousands of symbol cycles for state convergence when facing frequency steps. This millisecond-level loop convergence delay significantly compromises the timeliness of low-latency transmission, easily leading to packet loss in high-speed, oncoming flight scenarios.

[0134] The technical solution disclosed in this embodiment reduces the unknown frequency offset search space by using cross-layer feedforward of the kinematic characteristics of the flight control system; then, it adopts a pure hardware multi-branch parallel multiply-accumulate strategy and related threshold capture mechanism to transform the original serial closed-loop time delay into parallel logical space resource consumption.

[0135] The overall solution ensures that the reception latency of high-priority telemetry and control commands through the physical layer bypass transmission channel is strictly controlled within the microsecond range in harsh physical environments such as high-speed maneuvering and combat without adding any additional air interface detection overhead. It also significantly improves the anti-Doppler distortion capability of the receiving link under complex three-dimensional relative motion conditions, and has significant practical application value.

[0136] Example 3:

[0137] In actual UAV swarm network operations, due to the extremely high preemption priority of the physical layer bypass transmission channel provided by this invention, if a malicious node captured by the enemy, or a friendly node that malfunctions due to a deadlock in its flight control program, continuously broadcasts the highest-priority telemetry and control command preamble into space, normal nodes will remain in a backoff state during distributed listening arbitration. This phenomenon is known as priority deadlock or preamble flooding attack, which can paralyze the entire low-latency telemetry and control network.

[0138] To address this issue, this embodiment constructs a low-level hardware anomaly defense system independent of the media access control layer software.

[0139] Specifically, the extraction of environmental and link state features of UAV nodes and cross-layer linkage also includes linkage with the physical layer abnormal node suppression module, specifically including:

[0140] The receiving end continuously monitors the frequency of high-priority commands sent by each node's unique identifier and extracts the spatial angle of arrival of the arriving signal through an antenna array. In traditional network security architectures, the statistics of abnormal transmission frequencies are usually processed by the software firewall of the central processing unit after receiving complete data packets, which is lagging and inefficient when facing physical layer flooding attacks.

[0141] This solution instantiates hardware logic circuitry for transmitting frequency monitoring within the receiver baseband pipeline of a field-programmable gate array (FPGA). When the receiver's preamble detection circuit captures an arbitration-specific preamble containing a unique identifier for a legitimate node, the hardware statistics module immediately triggers the counter corresponding to that node identifier at the underlying level, without waiting for subsequent instruction data payload decoding to complete. To objectively and dynamically measure the transmitting frequency, this solution introduces a hybrid statistical algorithm of leaky bucket and token bucket, constructed entirely from digital logic, at the hardware level.

[0142] For example, the real-time state update of this hardware token bucket relies on the following digital logic recursive formula:

[0143] ;

[0144] In the formula: This represents the hardware count value of the currently available tokens corresponding to the unique identifier of a specific node during the k-th system clock sampling period; This represents the maximum capacity of the hardware token bucket for the highest priority burst of instructions allowed on this node. This represents the available token count value latched in the register during the previous clock cycle, i.e., the (k-1)th cycle. This parameter represents the constant recovery rate parameter by which the hardware linearly replenishes tokens into the bucket over time. This represents the physical time length of a single clock drive cycle in the baseband digital signal processing system. It should be noted that the system clock drive cycle and the front-end baseband sampling clock cycle belong to different digital clock domains, and their frequencies are independently determined by the runtime sequence of the FPGA logic gate circuits. It is a Boolean trigger variable. When a high-priority instruction is sent by the unique identifier of the specific node during the k-th period, the value of this variable is a preset consumption weight constant; otherwise, the value is zero.

[0145] In practical applications, this formula serves to maintain a dynamic credit limit for each legitimate node in the network. Under normal flight conditions, drones' emergency evasive maneuvers are limited by aerodynamic physics, making it impossible for them to perform massive operations that defy the laws of physics within a very short time. Therefore, the normal transmission frequency will be... Stable replenishment; however, in the event of malicious flooding, high-frequency... Consumption will quickly deplete tokens, making By resetting the frequency to zero, the source of the abnormal frequency signal can be accurately exposed in the time dimension.

[0146] Optionally, while performing statistical analysis in the time dimension, the receiver also needs to cross-verify the authenticity of the signal source from a spatial physical dimension. The extraction of the spatial angle of arrival of the arriving signal via the antenna array refers to calculating the wavefront phase difference of the incident electromagnetic wave using a millimeter-wave large-scale phased array antenna configured on the receiving node. When a signal claiming to be from a "friendly" source arrives at the receiving antenna array, due to the fixed geometric arrangement distance between the microstrip antenna elements in the antenna array, the arrival time of the same electromagnetic wavefront at different elements will produce a small physical time difference. This time difference manifests as a significant baseband phase difference in the millimeter-wave high-frequency band. The underlying digital beamforming network of the receiver can infer the true transmission location of the signal in three-dimensional physical space by extracting the phase deflection of adjacent RF channels.

[0147] For example, the process of calculating the spatial angle of arrival is based on the following spatial spectrum phase conversion formula:

[0148] ;

[0149] In the formula: It represents the true spatial angle of arrival of the target arrival signal after being calculated and output by the bottom phase array, that is, the spatial physical angle between the incident electromagnetic wave ray and the antenna array normal; This represents the millimeter-wave center operating wavelength parameter that carries the signal in the current space wireless channel; This represents the phase difference of the baseband complex signal between the receiving links of two adjacent antenna array elements, extracted in real time by the baseband phase detector. This represents the fixed distance between the physical geometric centers of two adjacent antenna elements in the phased array antenna at the receiving end.

[0150] Using this formula, the baseband processing unit can objectively determine the true physical spatial location of the current signal transmitter within tens of nanoseconds of receiving the preamble, employing a hardware-level arcsine lookup table without relying on any upper-layer positioning data exchange. When facing deceptive jamming by an enemy that forges friendly node identifiers, since the enemy jammer cannot physically occupy the exact location of the forged friendly UAV, the spatial angle measurement provided by this formula becomes a key physical feature for identifying the forged identity.

[0151] like Figure 7 The figure shows the peak value estimation of the spatial angle of arrival spectrum based on the baseband phase difference of the phased array antenna. The horizontal axis represents the spatial angle of arrival in degrees, and the vertical axis represents the normalized amplitude of the spatial spectrum in decibels.

[0152] The blue spatial spectrum amplitude curve in the figure represents the actual calculation result of the spatial incident signal orientation by the underlying digital beamforming network of the receiving node.

[0153] From the waveform transformation trend, it can be observed that the blue curve rises rapidly at the 30-degree position within the monitoring range of -90 degrees to +90 degrees on the horizontal axis, forming a significant energy peak with a normalized amplitude of 0 dB. In other spatial regions on both sides of this angle, the curve amplitude decreases sharply and remains in the noise floor range of about -30 dB.

[0154] The energy spike in this concentration objectively reflects that the system can accurately determine the true physical spatial orientation of the current electromagnetic wave ray as 30 degrees by extracting the phase difference of the baseband signal between the receiving links of adjacent antenna array elements.

[0155] Combined with the physical layer abnormal node suppression mechanism, if the expected angle of arrival corresponding to the unique identifier of a node resolved by the system in the topology maintenance table is -20 degrees, while the physical angle of arrival measured by the underlying hardware through the spatial spectrum curve is 30 degrees, the significant deviation of 50 degrees between the two obviously exceeds the preset angle tolerance range.

[0156] The attached data demonstrates that the system can directly identify spoofed signals with severely deviated positions from the physical space dimension by utilizing the spatial spectrum peak value of the output of the underlying phase array. This provides objective physical quantitative support for the subsequent generation of hardware blacklist masks and blocking of abnormal flooding nodes, enhancing the operational security of the low-latency telemetry and control network in complex interference environments.

[0157] It is also important to note that after acquiring the aforementioned spatiotemporal two-dimensional features, the abnormal node suppression module will enter the decision-making stage. Specifically, when the frequency of a node's unique identifier sending high-priority instructions exceeds the burst token threshold, or when the spatial angle of arrival deviates from the expected angle of arrival of the node's unique identifier in the preset topology maintenance table beyond the preset angle tolerance range, the source of the arriving signal will be determined as an abnormal flooding node. The burst token threshold here is the trigger boundary when the token bucket count value drops to the danger lower limit (e.g., zero) in the aforementioned formula, representing that the node exhibits abnormally high-frequency transmission behavior that exceeds physical norms in the time dimension. The preset topology maintenance table is a dynamic mapping table stored in the high-speed static random access memory at the receiving node's underlying layer.

[0158] Because drone swarms exchange position information during normal flight, this table records the unique three-dimensional coordinates of each legitimate node in the swarm in real time. Based on its own coordinates and the target coordinates recorded in the table, the receiving node continuously calculates the expected angle of arrival at which its signal should theoretically be transmitted. If the difference between the actual spatial angle of arrival measured by the aforementioned formula and the expected angle of arrival exceeds the preset angle tolerance range determined by system positioning errors and minor attitude jitter of the aircraft, it indicates that the signal source claiming to be a legitimate node has a significant deviation between its actual transmission position and its theoretical position, constituting a typical spatial spoofing signal. As long as the arriving signal triggers an out-of-bounds condition (i.e., a logical OR relationship) in either the time-frequency statistics or spatial angle verification dimension, the hardware decision-maker will immediately lock its source and classify it as an abnormal flooding node.

[0159] For example, after anomaly detection, a robust blocking mechanism must be established at the underlying level. The underlying hardware generates a hardware blacklist mask corresponding to the abnormal flooding nodes. This mask contains the unique identifier of the disabled nodes and their restricted spatial angle of arrival range. This is a purely digital logic-based security policy distribution mechanism. Due to the delay in upper-layer software intervention, the anomaly suppression module in this scheme directly writes this mask data to the configuration register bus of the multi-channel parallel preamble detection circuit. The hardware blacklist mask is a binary control string with a specific width. The high-order bits of this string record the unique identifier of the node determined to be abnormal, while the low-order bits define the restricted spatial angle of arrival range (i.e., the physical sector where the interference source is located) of the abnormal signal source. This masking mechanism does not involve complex routing table updates or media access control layer reconfiguration; instead, it hardwires the blacklist information to the physical layer's signal correlation matching front-end, giving it underlying physical-level filtering capabilities.

[0160] Specifically, the final blocking action of this scheme is executed as follows: the hardware blacklist mask is input to the multi-parallel preamble detection circuit; when the parsed features of a subsequent arriving signal match the hardware blacklist mask, the relevant calculated value of the subsequent arriving signal is forcibly cleared at the underlying logic, blocking it from entering the preset arbitration rule, so as to restore the preemption right of the normal UAV node to the physical layer bypass transmission channel. In the multi-parallel preamble detection circuit, the input baseband signal stream will continuously be convolved and integrated with the local preamble template in the sliding correlator to generate an energy correlation peak used to trigger subsequent reception actions.

[0161] In this embodiment, a mask comparator based on a hardware blacklist mask is added before the correlator. When a new air signal arrives, the mask comparator first quickly extracts the unique node identifier carried in its preamble and outputs the current real-time angle of arrival using the antenna array. If either of these two parsed features, or a combination of both, logically matches the hardware blacklist mask stored in the register, the comparator will immediately output a high-level interrupt suppression signal.

[0162] This interrupt suppression signal is directly connected to the reset pin of the end-of-line accumulator register of the hardware multiply-accumulate tree inside the sliding correlator. Therefore, by forcibly clearing the relevant calculated value of the subsequently arriving signal at the underlying logic level, it means that even if this abnormal signal has an extremely regular waveform and the highest priority level, the energy peak accumulated by integration in the hardware correlator will be forcibly pulled down to absolute zero by the physical circuit at the moment the decision is about to be output. Since the relevant calculated value is cleared, the signal can never trigger the relevant energy threshold set in the preamble detection circuit, and the system state machine will consider the spatial signal as merely invalid background white noise. This forced underlying signal shielding operation completely blocks malicious flooding signals, preventing them from entering the preset arbitration rules. In this way, the continuous backoff commands issued by the abnormal node become ineffective, the arbitration logic of the air interface channel is purified, and other normal UAV nodes are no longer suppressed by false high-priority signals when sending emergency commands, thus enabling them to successfully regain control of the physical layer bypass transmission channel even in extremely harsh electromagnetic countermeasures environments.

[0163] Currently, anti-interference and anti-flooding strategies for drone networks generally rely on upper-layer protocol stacks. Existing technologies typically establish a security gateway at the data link layer (MAC layer) or network layer. The receiving end must first perform complete baseband demodulation and channel decoding on the arriving radio frequency signal before passing it to the central processing unit (CPU). The CPU then runs security verification software to identify the legitimacy of the sending source, discarding the data packet if an anomaly is detected. This conventional method has a significant limitation: even if abnormal data packets are ultimately discarded by the CPU, the lengthy reception and decoding process objectively occupies valuable air interface channel transmission time slots and baseband hardware processing resources. When drone swarms with low latency requirements face high-frequency flooding attacks, this "receive first, discard later" software-layer filtering can cause high-priority instructions from normal nodes to time out due to channel congestion.

[0164] The technical solution disclosed in this embodiment sinks and solidifies complex network security verification into the front-end logic of the correlator of the physical baseband chip. It utilizes hardware token buckets and spatial spectrum estimation formulas to obtain spatiotemporal two-dimensional features, and uses mask comparison to force zeroing of correlation values ​​at the underlying level, thus blocking malicious signals at the hardware level during the physical layer correlation capture stage. The overall solution requires no software protocol stack intervention, endowing the low-latency millimeter-wave transmission architecture with strong anti-flooding capabilities and self-healing fault tolerance without increasing system response latency, significantly improving the survivability and mission reliability of large-scale UAV swarms in complex interference environments.

[0165] Example 4:

[0166] like Figure 8As shown, this embodiment provides a low-latency millimeter-wave transmission system for communication and control of UAV swarms. In the complex electromagnetic environment of UAV swarm collaborative operations, this system is typically integrated into each UAV node as an embedded payload terminal. The core processing architecture of this system is based on a combination of a field-programmable gate array (FPGA) and a high-performance millimeter-wave radio frequency front-end (RFFront-end). Through highly parallelized hardware design, it achieves low-latency transmission and highly reliable arbitration of control commands at the physical layer.

[0167] The system provided in this embodiment includes the following functional modules:

[0168] Specifically, the system includes a channel construction module for constructing a parallel regular service data transmission channel and a physical layer bypass transmission channel, wherein the regular service data transmission channel includes a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module.

[0169] In actual hardware deployment, the channel construction module is located within the system's baseband processing unit (BPU). The regular service data transmission channel handles service flows with high data throughput requirements, such as high-definition video surveillance and hyperspectral remote sensing data. To ensure the robustness of such large-capacity data in millimeter-wave channels, this channel allocates dedicated hardware codecs (such as LDPC codecs) and a deep interleaving storage matrix within the FPGA logic. The physical layer bypass transmission channel, on the other hand, is a dedicated hardwired path for latency-sensitive telemetry and control commands (such as emergency attitude adjustment and obstacle avoidance commands). It is also important to note that this bypass channel directly directs the data stream from the frame synthesis module to the physical layer arbitrator at the FPGA gate level, physically bypassing any channel coding and interleaving logic requiring store-and-forward or complex mathematical iterations, thereby compressing the data processing latency at the hardware link level.

[0170] Specifically, the system further includes a frame structure configuration module for configuring a deterministic frame structure for the physical layer bypass transmission channel. The front of the deterministic frame structure includes an arbitration-specific preamble field that carries air interface arbitration information.

[0171] In practical operation, the frame structure configuration module defines a highly concise and fixed-length bitstream format through the register group of the programmable logic (PL) terminal. This deterministic frame structure does not include dynamically adjusted overhead fields; the relative offsets of its various fields on the time axis are fixed during system initialization. The arbitration-specific preamble field serves as the identifier for each frame, carrying key arbitration elements such as unique node identifiers and priority levels. For example, the frame structure configuration module uses a state machine to ensure that the preamble field is preferentially fed to the RF front-end, enabling the receiver to determine the arbitration attribute of a frame instantly through the preamble detection circuit even before receiving the entire data frame. This deterministic timing design effectively improves the physical layer's agility in perceiving the state of air interface resources.

[0172] Specifically, the system also includes a transmitter arbitration transmission module, which is used to decode the arbitration-specific preamble field of the received signal in monitoring mode, perform distributed priority arbitration according to preset arbitration rules, and preempt the physical layer bypass transmission channel to send high-priority measurement and control commands after the arbitration is successful.

[0173] The transmitter arbitration module consists of a high-speed digital comparator array and an air interface monitoring controller. When the UAV node's flight control system generates a transmission request, this module first drives the RF transceiver into high-speed monitoring mode. The comparator array within this module decodes the arbitration-specific preamble detected by the antenna from other nodes in real time. According to the preset arbitration rules (including three levels of arbitration: priority, node identifier, and random number), the module autonomously determines whether the node possesses the current transmission privilege. Once arbitration is successful, the module triggers a hardware-level interrupt signal, forcibly suspending the data stream of the regular service data transmission channel and enabling the transmission logic of the physical layer bypass transmission channel. This distributed arbitration based on the physical layer preamble eliminates the need for command scheduling by the cluster's central node, significantly reducing conflict avoidance latency during multi-machine concurrent communication.

[0174] Specifically, the system also includes a receiver detection output module, which is used to detect the arriving signal through a multi-parallel preamble detection circuit, latch only the highest priority high-priority telemetry and control command, and output it to the flight control system of the UAV node.

[0175] In actual device interaction, the receiver detection output module directly interfaces with the analog-to-digital converter (ADC) interface of the millimeter-wave RF front-end. This module integrates multiple parallel correlation matchers, each corresponding to a preset priority preamble feature. When the air interface signal arrives, the multi-channel parallel preamble detection circuit performs real-time cross-correlation calculations on the input baseband signal stream. If signals from multiple nodes overlap simultaneously in space, the latching logic quickly selects the signal with the lowest priority value (i.e., the highest priority) based on the arrival time and amplitude of the correlation peaks, and latches the corresponding telemetry and control command data into a high-speed cache. Subsequently, this module directly transmits the command data to the flight control computer of the corresponding UAV node through a low-latency inter-chip interface (such as the AXI-Stream interface), achieving near real-time flow from physical waveforms to control logic.

[0176] Specifically, the system also includes a cross-layer linkage adjustment module, which is used to extract the environmental and link status characteristics of the UAV node, and perform cross-layer linkage based on the status characteristics and the preset arbitration rules to dynamically adjust the arbitration parameters and transmission strategies in the arbitration rules.

[0177] This module connects to the UAV's inertial navigation system (INS), global positioning system (GPS), and physical layer anomaly monitoring unit via an internal bus. The cross-layer linkage adjustment module extracts features including the UAV's real-time 3D velocity, rotor rotation frequency, and the current channel fading index. For example, when this module detects that the UAV is performing a high-G maneuver that could lead to severe baseband Doppler frequency offset, it proactively sends a command to the frame structure configuration module to adjust the preamble sequence length or increase the guard interval. Simultaneously, based on the extracted rotor obstruction prediction information, this module dynamically modifies the backoff algorithm parameters in the transmitter arbitration module, giving this node a higher preemption weight before a link interruption, thereby improving the success rate of telemetry and control command transmission in harsh environments.

[0178] In actual operation, the above modules work together. For example, when two UAVs are flying towards each other at high speed and there is an obstacle avoidance requirement, the transmitting arbitration module bypasses the encoding time of the regular service channel through the physical layer bypass transmission channel, pushing the obstacle avoidance command to the air interface with a time overhead of microseconds. At the same time, the cross-layer linkage adjustment module uses the kinematic data of the flight control feedforward to adjust the arbitration parameters in the arbitration rules to ensure that the high-priority command can quickly suppress other non-urgent services. The receiving detection output module uses a multi-channel parallel preamble detection circuit to accurately extract the command from the complex interference noise and send it to the flight control system actuator.

[0179] As is known from existing technologies, traditional UAV communication systems often cannot simultaneously achieve broadband service transmission and extremely low latency for telemetry and control commands, especially as swarm density increases, where channel arbitration latency grows exponentially. The system disclosed in this embodiment effectively mitigates the conflict between communication processing latency and air interface preemption reliability by constructing a decoupled dual-mode transmission path at the FPGA level and supplementing it with hardware-based distributed arbitration and cross-layer state linkage. This system not only enhances the link self-healing capability of UAVs in extreme maneuvering scenarios but also provides a robust underlying physical layer technical guarantee for collaborative operation of UAV swarms in complex, highly dynamic environments. Compared to traditional software-defined radio architectures, this system solution offers significant performance improvements in end-to-end command latency and Doppler distortion resistance, making it highly valuable for engineering implementation.

[0180] Example 5:

[0181] Corresponding to the above embodiments, the present invention also proposes an electronic device.

[0182] like Figure 9 The diagram shows a structural schematic of an electronic device according to the present invention. The electronic device 100 includes a processor 101 and a memory 103. The processor 101 and the memory 103 are connected, for example, via a bus 102. Optionally, the electronic device 100 may further include a transceiver 104. It should be noted that in practical applications, the transceiver 104 is not limited to one unit, and the structure of this electronic device 100 does not constitute a limitation on the embodiments of the present invention.

[0183] Processor 101 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 101 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0184] Bus 102 may include a pathway for transmitting information between the aforementioned components. Bus 102 may be a PCI bus or an EISA bus, etc. Bus 102 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0185] The memory 103 stores a computer program corresponding to the low-latency millimeter-wave transmission method for UAV swarm communication and telemetry in the above embodiments of the present invention. This computer program is executed by the processor 101. The processor 101 executes the computer program stored in the memory 103 to implement the content shown in the aforementioned method embodiments.

[0186] Among them, electronic devices 100 include, but are not limited to: mobile terminals such as laptops and PADs (tablet computers) and fixed terminals such as desktop computers. Figure 9 The electronic device 100 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0187] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A low-latency millimeter-wave transmission method for communication and control of unmanned aerial vehicle (UAV) swarms, characterized in that, Includes the following steps: Construct a parallel regular service data transmission channel and a physical layer bypass transmission channel, wherein the regular service data transmission channel includes a channel coding and interleaving module, and the physical layer bypass transmission channel bypasses the channel coding and interleaving module; A deterministic frame structure is configured for the physical layer bypass transmission channel, wherein the front of the deterministic frame structure includes an arbitration-specific preamble field carrying air interface arbitration information; In listening mode, the transmitter decodes the arbitration preamble field of the received signal, performs distributed priority arbitration according to the preset arbitration rules, and after the arbitration is successful, it preempts the physical layer bypass transmission channel to send high-priority measurement and control commands. The receiving end detects the arriving signal through a multi-channel parallel preamble detection circuit, latches only the highest priority high-priority telemetry and control command, and outputs it to the flight control system of the UAV node. Extract the environmental and link status features of the UAV node, and perform cross-layer linkage based on the status features and the preset arbitration rules to dynamically adjust the arbitration parameters and transmission strategies in the arbitration rules; The process involves extracting the environmental and link state features of the UAV node, and then performing cross-layer linkage based on these state features and the preset arbitration rules, including linkage with the rotor jamming prediction module. Based on the rotor parameters, relative position, and attitude angle of the UAV node, the link blocking status within the first preset time period is predicted. When it is predicted that the link will be blocked by the rotor, the priority of all commands to be sent by this node will be increased by a preset level, and the length of the air interface listening window will be shortened. Switch to a current and bias voltage mapping table pre-stored in the transmitter and suitable for non-line-of-sight environments, and increase the transmitter's transmission power.

2. The method according to claim 1, characterized in that, The total length of the deterministic frame structure is a first preset length, the length of the arbitration-specific preamble field is a second preset length, and the second preset length is less than the first preset length; The deterministic frame structure includes, in sequence: Bit synchronization header and frame synchronization header used for signal synchronization; A priority identifier used to define instruction priority; A unique node identifier used to globally and uniquely identify the sending node; Collision avoidance random numbers generated in real time by hardware; Command length identifier and command data carrying the actual measurement and control content; Hardware checksums for parallel verification, and a response request identifier to indicate the feedback from the receiving end.

3. The method according to claim 1, characterized in that, The pre-defined arbitration rules include a three-tiered arbitration mechanism executed sequentially: Level 1 Priority Comparison: When a signal with a higher priority than the instruction to be sent by this node is detected in the air, this node performs a backoff operation; Level 2 node identifier comparison: When a signal with the same priority as the instruction to be sent by this node is detected in the air, the unique node identifiers of the two are compared, and the node with the smaller value obtains the right to send; Level 3 random number comparison: When the priority and the node's unique identifier are the same, the random number is compared to avoid conflict, and the node with the smaller value gets the right to send.

4. The method according to claim 1, characterized in that, The process of preempting the physical layer bypass transmission channel to send high-priority measurement and control commands specifically includes: Interrupt the current data transmission of the regular business data transmission channel and save the address of the current data transmission breakpoint; Insert a breakpoint marker into the current business data stream and send the high-priority measurement and control command to the Serdes interface of the transmitter for transmission; After the high-priority measurement and control command is transmitted, the interrupted service data transmission is resumed according to the transmission breakpoint address.

5. The method according to claim 1, characterized in that, While sending high-priority measurement and control commands, the transmitter performs radio frequency direct-drive compensation, specifically including: The drain current of the power amplifier in the transmitter is collected in real time as the sampling current value; Based on the current link blocking status, match the target mapping table with several pre-stored current and bias voltage mapping tables; The sampled current value is input into the target mapping table to obtain the bias voltage correction amount, and the bias voltage correction amount is applied to the bias terminal of the RF driver amplifier in the transmitter through digital-to-analog conversion.

6. The method according to claim 1, characterized in that, The receiving end detects the arriving signal through a multi-channel parallel preamble detection circuit, specifically including: The received signal is deserialized into a baseband bit stream, and the baseband bit stream is synchronously input into multiple independent preamble detection circuits, each of which corresponds to a priority level. Calculate the correlation values ​​between the baseband bitstream and the preamble corresponding to each priority level independently; When the correlation value exceeds the first correlation threshold, the latch signal of the corresponding channel is triggered, and the priority level of each channel is compared in real time. Only the latch state of the highest priority channel is maintained to extract the instruction, while all low priority channels are closed.

7. The method according to claim 6, characterized in that, After latching the highest priority measurement and control command, the receiving end further includes: The checksum of the latched received frame is calculated in parallel by the hardware logic and compared with the hardware checksum of the sending end. When the comparison matches, the command data is directly output to the flight control system; If the comparison does not match, the received frame is discarded. If the high-priority telemetry and control command contains a response request identifier and the comparison is consistent, a hardware-level response frame is automatically generated and fed back to the transmitter through the physical layer bypass transmission channel.

8. The method according to claim 1, characterized in that, The step of extracting the environmental and link state features of the UAV node, and performing cross-layer linkage based on the state features and the preset arbitration rules, also includes linkage with the physical layer anomaly detection module: The link synchronization status, abrupt changes in received signal strength, and instruction error rate of the Serdes interface in the receiving end are monitored in real time, and corresponding levels of abnormal status are generated. When the attenuation of the received signal strength reaches the first attenuation threshold and is determined to be a level 1 anomaly, the priority of the instruction to be sent is increased and the transmission power of the transmitter is increased. When the instruction error rate reaches the first error threshold and is determined to be a level 2 anomaly, the backup bypass transmission channel is activated to execute the primary and backup dual transmission mode. When a link synchronization failure is detected and determined to be a Level 3 anomaly, the main channel transmission is stopped, the commands to be sent are completely switched to the backup bypass transmission channel, and re-node pairing and beam allocation are triggered.

9. The method according to claim 1, characterized in that, The step of extracting the environmental and link state features of the UAV node, and performing cross-layer linkage based on the state features and the preset arbitration rules, also includes linkage with the node beam joint optimization module: The default priority of each node is dynamically adjusted periodically based on the historical collision probability and the frequency of emergency command transmission. Antenna beams with corresponding beamwidths are assigned to instructions of different priorities, with the highest priority instructions assigned the first beamwidth, the medium priority instructions assigned the second beamwidth, and the lowest priority instructions assigned the third beamwidth. The first beamwidth is smaller than the second beamwidth, and the second beamwidth is smaller than the third beamwidth; The switching of the antenna beam is directly triggered and controlled by the physical layer hardware based on the priority level of the instruction.

Images