Software-defined radio adaptive image transmission method and system for low-altitude unmanned equipment

Abstract

This invention relates to the field of wireless communication and multimedia data transmission technology, and discloses a software-defined radio frequency adaptive image transmission method and system for low-altitude unmanned equipment. The method acquires the UAV video stream, angular velocity, and absolute altitude parameters; calculates the channel degradation index to trigger a reassembly response; extends the cyclic prefix and pilot; generates a physical resource squeeze token and reconfiguration flag; generates a reduced-dimensional motion vector frame using the angular velocity parameter; truncates the frame based on the token and maps it to a key frame for transmission; the receiving end demodulates and reconstructs the video stream based on the flag. The system includes data acquisition, baseband processing, physical layer reconstruction, joint scheduling, and receiver reconstruction modules to implement the above method. This invention utilizes dynamic parameters to predict channel conditions and implements cross-layer data scheduling, ensuring that the amount of data sent from the application layer matches the transient available capacity of the physical layer, avoiding disordered packet loss, and maintaining the continuity of video transmission.

Description

Technical Field

[0001] This invention relates to the field of wireless communication and multimedia data transmission technology, specifically to a software-defined radio frequency adaptive image transmission method and system for low-altitude unmanned equipment. Background Technology

[0002] In low-altitude inspection and reconnaissance operations, UAVs need to transmit real-time video streams to ground equipment. The low-altitude flight environment has complex obstacles, and the multipath reflection effect of electromagnetic waves is more obvious. When UAVs perform large-angle attitude adjustments or rapid trajectory changes, the continuous changes in the body attitude will cause the antenna radiation pattern to deflect. The Doppler frequency shift generated by the maneuvering flight will cause transient fading of the radio frequency physical channel, resulting in a rapid decrease in the available transmission capacity of the underlying channel in a short period of time.

[0003] Most existing video transmission systems use fixed underlying physical parameter configurations and static video compression strategies at the application layer for data transmission. Although some systems have dynamic bitrate adjustment capabilities, their adjustment mechanisms mainly rely on network statistical indicators such as packet loss rate or bit error rate calculated and fed back by the receiver. This passive adjustment method based on receiver indicator feedback has an inherent time lag and is difficult to cope with sudden channel condition deterioration caused by high-maneuverability aircraft movements. During the feedback lag period, the application layer video encoder continues to output video data at the original bitrate, causing the data volume to exceed the actual carrying capacity of the physical link. A large amount of unscheduled data accumulates in the physical layer transmission buffer and causes buffer overflow, resulting in disordered packet loss at the underlying layer. The random loss of video data packets directly damages the integrity of the underlying transmission bitstream protocol structure, causing the receiver's video decoder to stall and report errors due to missing key data. This results in frame skipping or stuttering in the video displayed on the terminal, making it difficult to ensure the continuity and reliability of video monitoring when the UAV is performing high-dynamic flight missions. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a software-defined radio frequency adaptive image transmission method and system for low-altitude unmanned equipment. It solves the problem that when the transient channel fading caused by the maneuvering flight of UAVs leads to transient channel fading, the existing system suffers from packet loss due to the lagging feedback adjustment mechanism, resulting in the stagnation and stuttering of terminal video decoding.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of this invention provides a software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment, the method comprising: Acquire the drone's raw video stream, three-axis angular velocity parameters, and absolute altitude parameters; The kinematic channel degradation index is calculated based on the three-axis angular velocity parameter and the absolute height parameter. When the time derivative value is greater than the maneuver threshold, a reorganization response is triggered. Based on the recombination response, the extended cyclic prefix and pilot subcarriers are used to calculate the payload reduction, generate a physical resource squeeze token, and transmit the reconfiguration flag. The original encoding was stopped, and a reduced-dimensional motion vector frame was generated using the three-axis angular velocity parameters; Based on physical resource squeezing tokens, truncate the reduced-dimensional motion vector frames and map the key image frames and the truncated reduced-dimensional motion vector frames to the service subcarriers for transmission. The receiving end demodulates the service data based on the reconfiguration flag, completes the data syntax structure, and reconstructs the video stream by combining it with the reduced-dimensional motion vector frames.

[0006] Further steps for obtaining the raw video stream, three-axis angular velocity parameters, and absolute altitude parameters of the UAV include: importing the raw video stream output by the payload camera into a circular buffer for cyclic overwriting; extracting the roll rate, pitch rate, yaw rate, and absolute altitude contained in the data packets sent by the UAV flight control system to obtain the three-axis angular velocity parameters and absolute altitude parameters; using the hardware exposure trigger edge of the video frame as the time reference point, retrieving the acquired inertial measurement unit dataset, and selecting data with attached parameters to generate timestamps that meet the principle of minimum absolute difference with the time reference point to perform hard synchronization alignment on the raw video stream.

[0007] This process utilizes hardware timestamps to align heterogeneous data streams, reducing sampling delay deviations between the flight control bus and the video bus, and maintaining time consistency between kinematic parameter calculations and video frame matching.

[0008] Furthermore, the kinematic channel degradation index is calculated based on the three-axis angular velocity parameters and the absolute height parameter. The steps for triggering a reassembly response when the time differential value is greater than the maneuver threshold include: introducing three-axis physical weighting coefficients to perform a linear summation operation on the squared terms of the three-axis angular velocity parameters, and performing a division operation on the summation result with the absolute height parameter and the positive compensation constant to obtain a discrete kinematic channel degradation index; extracting the kinematic channel degradation index corresponding to adjacent sampling points, performing a difference operation, and dividing it by the discrete sampling period constant after mean filtering and smoothing to extract the channel change rate parameter as the time differential value; when the channel change rate parameter is greater than the set maneuver threshold, synchronously broadcasting a low-level hardware interrupt trigger command to the system physical layer and video buffer unit as a reassembly response.

[0009] This process uses the rigid body dynamics parameters of the UAV to predict the channel state and converts the maneuver parameters into physical layer state transition commands to shorten the system's adjustment delay.

[0010] Furthermore, the steps of calculating the payload reduction, generating a physical resource squeeze token, and transmitting a reconfiguration flag based on the reconfiguration response, extended cyclic prefix, and pilot subcarriers include: in response to a low-level hardware interrupt trigger instruction, adjusting the cyclic prefix length of the low-level orthogonal frequency division multiplexing module to an extended ratio, and adjusting the pilot subcarriers from a comb-like spacing distribution mode to a dense block distribution mode; calculating the payload reduction within a single transmission time interval based on the increased value of the cyclic prefix time length and pilot subcarrier density, combined with the modulation order and error correction code rate, to obtain a physical resource squeeze token measured in binary bits; allocating a fixed subcarrier set in the edge boundary region of the operating frequency band to construct an extremely robust control signaling channel, and encapsulating the reconfiguration flag associated with the cyclic prefix length index value and the pilot mode index value into the extremely robust control signaling channel and transmitting it.

[0011] This setup, while expanding the cyclic prefix and pilot density, transmits parameter update information through the signaling channel and uses physical resource squeeze tokens to quantify changes in available resources, providing a processing benchmark for cross-layer data scheduling.

[0012] Furthermore, the steps of suspending video encoding and generating a reduced-dimensional motion vector frame using the three-axis angular velocity parameters include: blocking the transmission path of the original video stream to the standard video encoder and suspending the ongoing discrete cosine transform and quantization calculation pipeline operation of the current image sequence; clearing the residual frequency domain coefficient register associated with the current video frame while retaining the sequence parameter set and image parameter set; and, based on the rigid body kinematics model, calculating the global physical motion vectors of the image in the horizontal and vertical pixel directions according to the yaw and pitch angular velocities included in the three-axis angular velocity parameters, and filling them into the motion vector-specific fields corresponding to the encoding protocol to encapsulate and generate a reduced-dimensional motion vector frame.

[0013] This reduces residual texture coding computation when bandwidth resources are limited, uses flight control motion parameters to replace part of the inter-frame search, and replaces the original structure with basic headers and motion vectors, thereby reducing the amount of data per frame.

[0014] Furthermore, the steps for truncating the reduced-dimensional motion vector frame based on the physical resource squeeze token include: parsing the header of the data packet received by the media access control layer, transferring data packets identified as key image frames to the master transmission queue, and transferring data packets identified as reduced-dimensional motion vector frames to the truncation processing buffer queue; determining whether the actual binary length of the current reduced-dimensional motion vector frame is greater than or equal to the bit value of the physical resource squeeze token identifier; when the actual binary length is greater than or equal to the bit value, truncating the corresponding token length bits bit by bit from the end of the reduced-dimensional motion vector frame data packet towards the head; when the actual binary length is less than the bit value, discarding the current reduced-dimensional motion vector frame and deferred the remaining token value to the next data packet.

[0015] This operation truncates the data to the corresponding length based on the physical resource squeeze token, so that the amount of data sent by the application layer matches the available channel capacity of the physical layer, avoiding disordered packet loss caused by physical buffer overflow.

[0016] Furthermore, the step of mapping key image frames and truncated dimensionality-reduced motion vector frames to service subcarriers for transmission includes: mapping key image frames in the master control transmission queue to the central service subcarrier set corresponding to the center frequency band of the total subcarrier sequence of the system's working bandwidth; mapping the truncated dimensionality-reduced motion vector frame fragments to the edge service subcarrier set located at both ends of the total subcarrier sequence and adjacent to the inner frequency band of the extremely robust control signaling channel; and converting the frequency domain data blocks that have completed asymmetric mapping into time domain transmission signals through inverse fast Fourier transform to perform air interface transmission operations.

[0017] This mapping process allocates key image frames to the flatter central subcarrier region and truncated data to the edge subcarrier regions to maintain the integrity of the core image frames during transmission.

[0018] Furthermore, the steps for demodulating service data at the receiving end based on the reconfiguration flag include: splitting the input discrete time-domain sampled data stream into independent listening window pipelines and service data window pipelines at the receiving end, and configuring a buffer at the front end of the service data window pipeline to perform clock cycle delay dwell; using the listening window pipeline to extract and parse the reconfiguration flag data packet in the extremely robust control signaling channel frequency band; hard-writing the index parameters contained in the reconfiguration flag into the core state machine of the service data window to update the demodulation configuration parameters, and extracting time-domain data from the buffer according to the updated configuration parameters, stripping the corresponding cyclic prefix and performing a fast Fourier transform to obtain the frequency domain service subcarrier data.

[0019] This dual-window processing structure is used to separate the timing of control signaling parsing and service data demodulation, so that the receiving end can switch the underlying demodulation parameters.

[0020] Furthermore, the steps of completing the data syntax structure and reconstructing the video stream by combining the reduced-dimensional motion vector frames include: adding a bit 1 and appending several bits of zero to the end of the truncated data packet binary truncation point in the demodulated output to supplement the original byte sequence payload alignment bit; re-encapsulating the header length identifier and tail delimiter sequence of the network abstraction layer unit; extracting the global physical motion vector contained in the reduced-dimensional motion vector frame and filling it into the standard prediction macroblock, and performing all-zero initialization overlay processing on the associated residual coefficient matrix; performing a picture block translation extrapolation operation on the corresponding block data in the reference frame buffer pool of the receiving end according to the global physical motion vector, and outputting the reconstructed video stream.

[0021] This helps to correct incomplete network protocol syntax caused by truncation operations, and uses extracted physical motion vectors to perform motion compensation, outputting a reconstructed video stream.

[0022] The second aspect of this invention provides a software-defined radio frequency adaptive image transmission system for low-altitude unmanned equipment, using the software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment as described in any one of the first aspects of this invention. The system specifically includes: The data acquisition module is used to acquire the drone's raw video stream, three-axis angular velocity parameters, and absolute altitude parameters; The baseband processing module is used to calculate the kinematic channel degradation index based on the three-axis angular velocity parameters and absolute height parameters, and triggers a reorganization response when the time derivative value is greater than the maneuver threshold. The physical layer reconfiguration module is used to expand the cyclic prefix and pilot subcarriers based on the reconfiguration response, calculate the payload reduction, generate a physical resource squeeze token, and transmit the reconfiguration flag. The joint scheduling module is used to suspend video encoding, generate a reduced-dimensional motion vector frame using the three-axis angular velocity parameter, truncate the reduced-dimensional motion vector frame according to the physical resource squeeze token, and map the key image frame and the truncated reduced-dimensional motion vector frame to the service subcarrier for transmission. The receiving reconstruction module is used to demodulate service data at the receiving end based on the reconfiguration flag, complete the data syntax structure, and reconstruct the video stream by combining it with the reduced-dimensional motion vector frames.

[0023] This invention provides a software-defined radio frequency adaptive image transmission method and system for low-altitude unmanned equipment. It has the following beneficial effects: 1. This invention calculates the kinematic channel degradation index by acquiring the three-axis angular velocity parameters and absolute altitude parameters of the UAV, and triggers a reorganization response in advance when its time derivative value is greater than the maneuver threshold. This design utilizes the rigid body dynamics data of the UAV to construct a priori prediction mechanism for the channel state, changing the traditional image transmission system's lag adjustment method that relies on the receiver's packet loss or bit error feedback. It can proactively complete the expansion of the underlying cyclic prefix and pilot subcarriers before the radio frequency channel degradation is caused by the body maneuver, thereby improving the image transmission link's anti-multipath interference capability and connection stability in dynamic flight environments.

[0024] 2. This invention generates a physical resource squeeze token based on the effective payload reduction calculated after the extended cyclic prefix and pilot. When the channel is limited, conventional video coding is stopped, and a reduced-dimensional motion vector frame is generated using physical motion parameters. The reduced-dimensional motion vector frame is then strictly truncated bit by bit according to the value of the physical resource squeeze token. This cross-layer scheduling strategy forces the video data volume sent from the application layer to be aligned with the transient available channel capacity of the physical layer, avoiding disordered packet loss caused by physical layer buffer overflow. Combined with asymmetric frequency domain scheduling that maps key image frames to the central service subcarrier, it ensures the complete transmission of the core video sequence under limited bandwidth.

[0025] 3. At the receiving end, for service data that has undergone truncation, this invention adds bits to the end of the binary truncation point to supplement the byte sequence payload alignment bits and re-encapsulates the header and tail delimiter sequences of the network abstraction layer unit. This processing repairs the missing protocol syntax structure caused by forced truncation, preventing standard video decoders from experiencing decoding stall errors due to incomplete bitstreams. At the same time, it extracts the global physical motion vectors within the reduced-dimensional motion vector frame and performs translation extrapolation operations on the reference frame block, enabling image prediction compensation even in the absence of residual texture coefficients, maintaining the visual continuity of the video stream displayed at the receiving terminal. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the module architecture of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention; Figure 3 This is a schematic diagram illustrating the logic principle of the feedforward triggering mechanism of the present invention; Figure 4 This is a schematic diagram illustrating the principle of cross-layer reverse processing and dimensionality reduction frame generation in this invention; Figure 5 This is a schematic diagram illustrating the data truncation and asymmetric mapping principle of the present invention; Figure 6 This is a comparison chart of application layer network packet loss rates under different flight conditions according to the present invention; Figure 7 This is a comparison chart of the average peak signal-to-noise ratio of the receiver under different flight conditions according to the present invention.

[0027] Among them, 100 is the data acquisition module; 110 is the video buffer unit; 120 is the parameter extraction unit; 200 is the baseband processing module; 210 is the index calculation unit; 220 is the trigger analysis unit; 300 is the physical layer reconstruction module; 310 is the parameter reconfiguration unit; 320 is the signaling encapsulation unit; 400 is the joint scheduling module; 410 is the reverse processing unit; 420 is the data allocation unit; 500 is the receiver reconstruction module; 510 is the dual-window demodulation unit; and 520 is the proxy reconstruction unit. Detailed Implementation

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] See attached document Figure 1 The present invention provides a software-defined radio frequency adaptive image transmission system for low-altitude unmanned equipment, which may include: a data acquisition module 100, a baseband processing module 200, a physical layer reconstruction module 300, a joint scheduling module 400, and a receiving reconstruction module 500.

[0030] The data acquisition module 100 includes a video buffer unit 110 and a parameter extraction unit 120. The video buffer unit 110 is used to receive the raw video stream output by the camera on the UAV payload and perform data buffering operations. The parameter extraction unit 120 is used to extract the inertial measurement unit data and global navigation satellite system data of the flight control system through the system bus, and obtain the three-axis angular velocity parameters and absolute altitude parameters of the UAV at the current time point.

[0031] The baseband processing module 200 includes an index calculation unit 210 and a trigger analysis unit 220. The index calculation unit 210 is used to quantize and calculate the kinematic channel degradation index according to the triaxial angular velocity parameter and the absolute height parameter with a preset weight coefficient. The trigger analysis unit 220 is used to perform time differentiation operation on the kinematic channel degradation index and issue a reconstruction trigger command when the differentiation calculation result exceeds a set threshold.

[0032] The physical layer reconstruction module 300 includes a parameter reconfiguration unit 310 and a signaling encapsulation unit 320. The parameter reconfiguration unit 310 is used to respond to the reconstruction trigger command, extend the cyclic prefix length of the orthogonal frequency division multiplexing symbol, increase the pilot subcarrier density, and calculate the reduction of effective physical payload bits in a single transmission time interval due to parameter adjustment to generate a physical resource squeeze token. The signaling encapsulation unit 320 is used to allocate an extremely robust control signaling channel in the set out-of-band frequency band and encapsulate the reconfiguration flag bit containing the cyclic prefix length and pilot mode index into the extremely robust control signaling channel.

[0033] The joint scheduling module 400 includes a reverse processing unit 410 and a data allocation unit 420. The reverse processing unit 410 is used to send a hardware interrupt signal to the video buffer unit 110 when a reconstruction trigger instruction is received, blocking the transmission path of the video stream to the standard encoder, extracting the motion vector of the picture based on the absolute height parameter and generating a dimension-reduced motion vector frame. The data allocation unit 420 is used to read the network abstraction layer unit data, perform end data truncation on the dimension-reduced motion vector frame according to the bit value of the physical resource squeeze token, map the key image frame data to the central service subcarrier set, and map the truncated dimension-reduced motion vector frame to the edge service subcarrier set.

[0034] The receiving reconstruction module 500 includes a dual-window demodulation unit 510 and a proxy reconstruction unit 520. The dual-window demodulation unit 510 is used to synchronously run dual-channel fast Fourier transform processing windows at the receiving end, parse the data of the extremely robust control signaling channel through the listening window, and update the demodulation configuration parameters of the service window according to the reconfiguration flag obtained by parsing. The proxy reconstruction unit 520 is used to complete the truncated data stream output by the service window with standard protocol syntax, encapsulate the motion vector of the picture into macroblocks, reset the image residual frequency domain coefficient matrix to zero, and output a restored video stream that meets the standard video decoder syntax.

[0035] See attached document Figure 2 This invention provides a software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment, comprising the following steps: S1, the data acquisition module 100 performs data acquisition operations in parallel, the video buffer unit 110 stores the raw video stream output by the payload camera, and the parameter extraction unit 120 simultaneously extracts the three-axis angular velocity parameters and absolute altitude parameters of the UAV at the current time node. S2, the index calculation unit 210 in the baseband processing module 200 calculates the kinematic channel degradation index based on the three-axis angular velocity parameter, absolute height parameter and set physical weighting coefficient, triggers the judgment unit 220 to perform time differentiation operation on the kinematic channel degradation index, and triggers the underlying protocol architecture reorganization response when the time differentiation calculation result is greater than the set maneuver threshold. S3, the parameter reconfiguration unit 310 in the physical layer reconstruction module 300 adjusts the cyclic prefix length of the underlying orthogonal frequency division multiplexing framing module to an extended value and increases the number of pilot subcarriers. Based on the adjusted parameter configuration, it calculates the physical resource squeeze token caused by the reduction of effective payload during the transmission time interval. The signaling encapsulation unit 320 establishes an extremely robust control signaling channel using ultra-long cyclic prefix and binary phase shift keying modulation. It encapsulates the reconfiguration flag containing parameter configuration information in the extremely robust control signaling channel and transmits it. S4, the reverse processing unit 410 in the joint scheduling module 400 sends a hardware interrupt instruction to the video buffer unit 110 through the reverse bus to stop the current discrete cosine transform calculation process, extracts the flight control attitude change quantity and converts it into the physical motion vector of the screen, and generates a dimensionality-reduced motion vector frame data packet containing only the sequence parameter set header and the physical motion vector of the screen. S5, the data allocation unit 420 identifies the current data attributes based on the network abstraction layer packet header, reads the bit value carried by the physical resource squeezing token when the data is identified as a dimensionality reduction motion vector frame, performs a truncation operation at the end of the dimensionality reduction motion vector frame packet body with a length equal to the bit value, maps the standard keyframe in the video stream to the center frequency band service subcarrier, and maps the truncated dimensionality reduction motion vector frame fragments to the edge frequency band service subcarrier. S6, the dual-window demodulation unit 510 in the receiving reconstruction module 500 continuously operates in the extremely robust control signaling channel frequency band using the listening window. After listening to the reconfiguration flag, it updates the cyclic prefix synchronization point and channel estimation equalization parameter matrix of the corresponding service demodulation window to complete the service data demodulation. The proxy reconstruction unit 520 fills the truncated bits of the demodulated truncated packet with the original byte sequence payload alignment bits and the network abstraction layer terminator, reads the physical motion vector of the picture in the reduced motion vector frame, fills it into the standard macroblock structure, and performs all-zero forced initialization processing on the associated residual frequency domain coefficient matrix. It outputs the reconstructed video stream to the display terminal to perform motion compensation picture prediction.

[0036] In specific implementations, step S1 provided by this invention may include the following steps: S101, the video buffer unit 110 establishes a bypass interception and storage link for the video stream output by the payload camera. Before the original uncompressed video stream output by the payload camera is transmitted to the system hardware encoder, the video buffer unit 110 uses the direct memory access controller configured inside the field programmable gate array to import the original video stream into a preset circular buffer. The minimum depth parameter of the circular buffer is configured as the product of the system's preset decision time window duration and the payload camera's highest output frame rate and is deployed in the system's double data rate synchronous dynamic random access memory. In this embodiment, the value range of the system's preset decision time window duration is 0.5 seconds to 2.0 seconds.

[0037] The video stream is written in a circular overlay manner in units of image groups. The input enable pin of the system's standard hardware video encoder is controlled by an internal system bus trigger. In the absence of a subsequent cross-layer hardware interrupt signal, the standard hardware video encoder retrieves the original image frame data from the circular buffer according to a fixed read step size and performs routine encoding operations. The specific physical connections of the payload camera video acquisition interface protocol and the routine format conversion operations of the image data color space can be configured according to the standard hardware chip manual.

[0038] S102, the parameter extraction unit 120 performs high-frequency polling and extraction of kinematic data based on timestamp matching at the current time node. The daemon process running on the baseband processor inside the parameter extraction unit 120 establishes a serial communication connection with the UAV flight control motherboard through the universal asynchronous transceiver bus or controller area network bus. The daemon process reads the fused data packets sent by the flight control system at a fixed sampling frequency, extracts the parameter generation timestamp generated by the internal hardware timer of the flight control system by parsing the protocol field of the data packet header, and directly extracts the roll rate of the UAV from the bus register. Pitch angular velocity yaw rate And the absolute altitude of the UAV relative to the ground obtained by combining barometer parameters and satellite positioning calculations. .

[0039] S103, the data acquisition module 100 executes the timestamp hard synchronization alignment mechanism of the heterogeneous data stream. The system uses the edge of the hardware exposure trigger pulse of the video frame at the sensor as the time reference point. When each original video frame is stored in the video buffer unit 110, the parameter extraction unit 120 retrieves the currently acquired inertial measurement unit dataset and selects the kinematic parameter with attached parameter timestamp that satisfies the principle of minimum absolute difference with the time reference point as the synchronization parameter corresponding to the video frame. The calculation method of this time alignment filtering logic is executed in real time by the underlying logic operation unit and outputs the corresponding matching pair containing the video frame and the kinematic parameter.

[0040] ; In the formula, The kinematic parameter timestamps that are finally selected and match the currently stored original video frames. This serves as the hardware exposure trigger time reference point at the underlying level of the video frame. The parameter extraction unit 120 acquires the first parameter via the bus. Each discrete kinematic parameter sample point has its own parameter generation timestamp at the top, which is a discrete identifier variable. It is a positive integer.

[0041] See attached document Figure 3 In specific implementations, step S2 provided by the present invention may include the following steps: S201, the exponent calculation unit 210 in the baseband processing module 200 receives the synchronized three-axis angular velocity parameters and absolute altitude parameters. The deflection of the antenna radiation pattern caused by the UAV's maneuvering flight, superimposed with the low-altitude multipath reflection effect, constitutes the general physical cause of channel degradation. The exponent calculation unit 210 uses an internal hardware multiplier-adder array to construct a physical calculation model. This model introduces pre-set three-axis physical weighting coefficients associated with the aircraft body and performs a linear summation operation on the square terms of roll angular velocity, pitch angular velocity, and yaw angular velocity. The exponent calculation unit 210 performs a division operation with the summation result, the absolute altitude parameter, and the positive compensation constant to obtain a discrete kinematic channel degradation exponent sequence. The specific values ​​of the three-axis physical weighting coefficients can be obtained by performing anechoic chamber measurements based on the aerodynamic shape of the aircraft and the antenna offset angle. In this embodiment, the values ​​of the three-axis physical weighting coefficients corresponding to the roll dimension, pitch dimension, and yaw dimension are all in the range of 0.1 to 1.0, and the sum of the values ​​of the three physical weighting coefficients is equal to the constant 1.0.

[0042] ; In the formula, To calculate the kinematic channel degradation index at the current moment; , and These are the three-axis physical weighting coefficients corresponding to the roll, pitch, and yaw dimensions, respectively. This is the absolute height relative to the ground after the bottom-level amplitude limiting processing; A positive compensation constant preset for the system, with values ​​ranging from 0.001 to 0.01; It is a continuous-time variable parameter.

[0043] S202, the trigger judgment unit 220 opens a sliding window memory array within the digital signal processing core. This memory array receives and caches the discrete kinematic channel degradation index sequence output by the exponent calculation unit 210 in real time. The arithmetic logic unit inside the trigger judgment unit 220 extracts the degradation index sample value corresponding to the current time node and the degradation index sample value corresponding to the previous adjacent time node from the sliding window and performs difference calculation. Before performing the difference operation, the digital signal processing core performs mean filtering smoothing processing on the exponent sequence data entering the sliding window based on the smoothing window composed of 3 to 5 consecutive sampling points configured by the system. The underlying arithmetic logic unit divides the smoothed difference by the discrete sampling period constant configured by the system to extract the channel rate of change parameter. In this embodiment, the value range of the discrete sampling period constant configured by the system is 10 milliseconds to 50 milliseconds.

[0044] ; In the formula, The first-order time discrete derivative of the channel rate of change parameter obtained by the calculation is the kinematic channel degradation index. This represents the sample value of the kinematic channel degradation index corresponding to the previous sampling period within the sliding window; The discrete sampling period constant of the exponential sequence configured internally by the system; For continuous time variables The kinematic channel degradation index is calculated one sampling period prior to the time step. It is a continuous-time variable parameter.

[0045] S203, the trigger analysis unit 220 is configured with a hardware comparator circuit including a numerical input port. This circuit continuously receives the channel change rate parameter input from the arithmetic logic unit and compares it with the maneuver threshold constant stored in the built-in register. When the hardware comparator circuit determines that the channel change rate parameter value is less than or equal to the maneuver threshold constant, the system maintains the current protocol stack parameter state unchanged. When the hardware comparator circuit determines that the current channel change rate parameter value is greater than the maneuver threshold constant, the general-purpose input / output port of the trigger analysis unit 220 generates a rising edge level transition signal. The internal bus controller converts the level transition signal into a low-level hardware interrupt trigger instruction and broadcasts it synchronously to the physical layer protocol stack state machine and the video buffer unit 110 through the system high-speed bus. The maneuver threshold constant is set according to the statistical average value of the channel change rate critical value recorded by the prototype in previous link failure tests and is solidified and written into the register during the system power-on initialization phase. In this embodiment, the value range of the maneuver threshold constant is 1.5 to 3.5.

[0046] In specific implementations, step S3 provided by this invention may include the following steps: S301, after receiving the underlying hardware interrupt trigger instruction broadcast by the bus, the parameter reconfiguration unit 310 in the physical layer reconstruction module 300 suspends the current protocol stack state machine. The parameter reconfiguration unit 310 extracts the preset configuration matrix and expands the cyclic prefix length of the underlying orthogonal frequency division multiplexing framing module from the default one-sixteenth symbol period ratio to a one-quarter symbol period ratio. The parameter reconfiguration unit 310 synchronously adjusts the frequency domain insertion rule of the pilot subcarrier from the comb-shaped spacing distribution mode to the dense block distribution mode. The register address overwrite operation of the relevant underlying timing logic is completed within the current transmission time interval. When the kinematic channel degradation index calculated by the exponent calculation unit 210 is lower than the set recovery threshold for multiple consecutive sampling periods, the parameter reconfiguration unit 310 controls the relevant registers to restore the above radio frequency parameters to the default state. In this embodiment, the value range of the set recovery threshold is 0.1 to 0.5.

[0047] S302, increasing the cyclic prefix time length and pilot subcarrier density will encroach on the original data transmission time-frequency space. The parameter reconfiguration unit 310 calculates the reduction in effective payload within a single transmission time interval based on the changes in the cyclic prefix length and the number of pilot subcarriers. The internally configured hardware arithmetic unit reads the total number of subcarriers in the current system and the modulation and coding strategy table to perform quantitative numerical calculations of the physical resource squeeze token. The calculated result is measured in binary bits and output to the cross-layer interaction register address set in the shared memory block. The specific calculation logic of the physical resource squeeze token is determined by the following calculation model.

[0048] ; In the formula, The specific bit reduction value contained in the physical resource squeeze token generated for accounting purposes; This is the reference bit base value carried by the link under standard parameter conditions; This is a constant representing the total number of subcarriers in an orthogonal frequency division multiplexing system. This represents the total number of newly inserted pilot subcarriers in the frequency domain after reconfiguration. The number of guard empty subcarriers reserved at both ends of the spectrum for which no energy is transmitted; The time-domain duration parameter for the data symbol; The actual time length of the expanded maximum proportional cyclic prefix; This refers to the modulation order parameter used by the current baseband modulator; The error correction code rate ratio used in the channel coding mechanism; The underlying transmission time interval is fixed.

[0049] S303, the signaling encapsulation unit 320 allocates a fixed set of subcarriers ranging from 2% to 5% of the total number of subcarriers in the boundary region near the guard space subcarriers on both sides of the frequency domain edge of the system's operating frequency band to construct an independent extremely robust control signaling channel. The extremely robust control signaling channel maintains constant internal configuration parameters when the underlying protocol reconfiguration state changes. The signaling encapsulation unit 320 configures it with a binary phase shift keying modulation mode and a fixed quarter-proportion ultra-long cyclic prefix. The signaling encapsulation unit 320 concatenates the cyclic prefix length index value associated with the physical state of the current service subcarrier region and the pilot mode index value into a reconfiguration flag data packet through cyclic redundancy check encoding. The signaling encapsulation unit 320 maps and encapsulates the reconfiguration flag data packet into the subcarrier sequence of the extremely robust control signaling channel and delivers it to the radio frequency front-end module through the baseband interface along with the service data subcarrier for air interface concurrent transmission.

[0050] See attached document Figure 4 In specific implementations, step S4 provided by the present invention may include the following steps: S401, the reverse processing unit 410 in the joint scheduling module 400 continuously monitors the hardware state transition signal of the underlying physical layer through the reverse control bus inside the system. After receiving the underlying hardware interrupt trigger instruction broadcast by the baseband processing module 200, the reverse processing unit 410 sends a state machine intervention instruction to the video buffer unit 110 and the standard video encoder. The hardware logic control circuit physically blocks the normal transmission path of the original video stream to the standard video encoder by pulling down the input enable pin level of the data output channel of the video buffer unit. The hardware bus controller then diverts the original image frame data stream currently residing in the video buffer unit 110 to a custom bypass processing logic pipeline configured inside the field programmable gate array. This pipeline silently discards the currently buffered original image frame data and releases the corresponding memory address space. When the physical layer hardware interrupt trigger instruction is released, the hardware logic control circuit pulls the input enable pin level high again to restore the normal transmission path of the original video stream.

[0051] S402, for video frame data already in the processing state within the standard video encoder, the standard video encoder is preferably a video coding IP soft core deployed in a field-programmable gate array, whose internal pipeline state machine and register addressing interface are open to the internal system bus. The reverse processing unit 410 stops the ongoing discrete cosine transform and quantization calculation pipeline operation of the current image sequence through the register configuration interface. The system's underlying state machine clears the associated residual frequency domain coefficient register and directly discards the generated residual data. The reverse processing unit 410 retains the basic header information such as the sequence parameter set and image parameter set of the frame and stops performing regular inter-frame matching prediction and spatial residual calculation on the current frame. The relevant timing control logic is triggered by the internal timer locked at the microsecond-level macroblock boundary of the current processing to prevent the underlying memory access address from going out of bounds or the bus from malfunctioning.

[0052] S403, the dramatic change in the drone's attitude is manifested as an overall translation trend in the image at the physical optical mapping level. The reverse processing unit 410 calls the drone's three-axis angular velocity parameters acquired synchronously by the parameter extraction unit 120 to construct an image motion vector conversion model based on rigid body kinematics. The built-in digital signal processor calculates the global physical motion vector of the image in the horizontal and vertical pixel directions based on the drone's yaw and pitch angular velocity parameters at the current moment. Since the motion compensation mechanism of the standard video coding macroblock only supports two-dimensional translation models, the rotation component caused by roll angular velocity is ignored in this calculation process. The system uses a projection conversion formula based on camera optical parameters to extract pixel-level motion deviations from the image. The extracted global physical motion vector data is written into the motion vector dedicated field of the corresponding macroblock of the encoding protocol. A preset length of redundant placeholder bits is added to the end of the packet. Finally, it is combined with the reserved sequence parameter set and image parameter set to encapsulate a dimension-reduced motion vector frame data packet.

[0053] ; ; In the formula, and These are the horizontal and vertical physical motion vectors, respectively, calculated and extracted. Here is the yaw rate parameter of the UAV at the current time point. This is the pitch angular velocity parameter of the UAV at the current time point. This is the sampling time interval parameter between two adjacent video frames. The physical equivalent focal length parameter of the camera lens is used to measure the load. For the single physical pixel size constant of the payload camera image sensor, including and The optical constant information is obtained by reading the peripheral configuration descriptor during the system initialization phase and stored in the relevant read-only registers; It is a continuous-time variable parameter.

[0054] See attached document Figure 5 In specific implementations, step S5 provided by the present invention may include the following steps: S501, the data allocation unit 420 in the joint scheduling module 400 receives video stream data packets sent by the application layer video encoder at the media access control layer. The network data packet sniffing logic built into the data allocation unit 420 parses the network abstraction layer unit message header of each video stream data packet. The data allocation unit 420 identifies whether the current data packet is a key image frame (i.e., an I-frame) containing full spatial information and intra-frame prediction or a reduced-dimensional motion vector frame based on the type identifier field in the message header. The data allocation unit 420 transfers the data packets identified as key image frames to the master control sending queue and transfers the data packets identified as reduced-dimensional motion vector frames to the subordinate truncation processing buffer queue.

[0055] S502, the data allocation unit 420 reads the physical resource squeeze token value stored in the shared memory through the internal system bus. The data allocation unit 420 calls the underlying bit operation instruction set to perform end-of-line truncation processing on the dimensionality reduction motion vector frame data packets in the subordinate truncation processing buffer queue. Before performing the truncation operation, the data allocation unit 420 obtains the actual binary length of the current dimensionality reduction motion vector frame and compares it with the bit value of the physical resource squeeze token. When the actual binary length of the dimensionality reduction motion vector frame is greater than or equal to the bit value, the data allocation unit 420 truncates the corresponding token length of binary bits bit by bit from the end of the data packet towards the head. Since the tail of the dimensionality reduction motion vector frame has been pre-filled with redundant placeholder bits, this end-of-line truncation operation will not destroy the sequence parameter set and physical motion vector and other core data in the front of the data packet. When the actual binary length of the reduced-dimensional motion vector frame is less than the bit value, the data allocation unit 420 discards the current data packet as a whole and deferred the remaining token value to the next reduced-dimensional motion vector frame to continue the truncation operation. After completing the bit field truncation, the data allocation unit 420 recalculates the actual payload length of the current truncated data packet and overwrites the length field in the media access control layer header. At the same time, it triggers the underlying logic gate circuit to recalculate the cyclic redundancy check code of the data packet and append it to the end of the truncated data packet.

[0056] S503, the data allocation unit 420 executes an asymmetric spatial mapping rule based on the anti-channel fading characteristics of subcarrier frequency domain position. The central service subcarrier set of the orthogonal frequency division multiplexing system avoids the influence of local oscillator leakage on the central DC subcarrier. The central DC subcarrier is less affected by out-of-band attenuation caused by hardware analog filter and Doppler frequency shift, and the channel quality is better than that of edge subcarriers. The data allocation unit 420 maps the key image frame data in the master control transmission queue to the central service subcarrier set in the middle third interval of the corresponding total subcarrier sequence index within the system working bandwidth. The data allocation unit 420 maps the truncated motion vector frame fragment data generated by the truncation processing buffer queue to the edge service subcarrier set located at both ends of the total subcarrier sequence and adjacent to the inner side of the aforementioned extremely robust control signaling channel, each occupying one-third of the interval. After the mapping is completed, the frequency domain data block is converted into a time domain transmission signal through inverse fast Fourier transform. Conventional physical layer processing processes such as interleaving of orthogonal frequency division multiplexing symbols and inverse Fourier transform can be implemented using a general wireless communication baseband chip processing architecture.

[0057] In specific implementations, step S6 provided by the present invention may include the following steps: S601, the dual-window demodulation unit 510 in the receiver reconstruction module 500 deploys parallel operation logic of dual fast Fourier transform windows at the digital baseband input node of the receiver. The dual-window demodulation unit 510 copies and splits the discrete time-domain sampled data stream input from the front end into two independent hardware processing pipelines. The first listening window pipeline is fixedly configured as a fixed ultra-long cyclic prefix mode with a quarter symbol period ratio. The listening window pipeline is dedicated to extracting extremely robust control signaling channel subcarrier data distributed in the edge boundary region of the working frequency band. The listening window pipeline continuously performs demodulation and channel decoding operations of binary phase shift keying signals to extract reconfiguration flag data packets associated with the physical state of the current service subcarrier.

[0058] In S602, due to the hardware pipeline clock consumption in the demodulation and parameter parsing operations of the extremely robust control signaling channel, the dual-window demodulation unit 510 configures a first-in-first-out circular buffer at the front end of the second service data window pipeline to perform a delay dwell on the service time domain data stream for the corresponding clock cycle. After completing the control signaling parsing of the current orthogonal frequency division multiplexing symbol, the monitoring window pipeline sends a parameter update instruction to the configuration register of the service data window pipeline through the underlying system high-speed bus. Within a nanosecond clock cycle, the service data window pipeline hard writes the cyclic prefix length index value and pilot mode index value contained in the reconfiguration flag bit into the internal core state machine. Based on the updated index parameters, the service data window pipeline extracts the time domain data from the first-in-first-out circular buffer and strips the corresponding cyclic prefix to perform a fast Fourier transform to obtain the frequency domain service subcarrier data.

[0059] S603, after the frequency domain service subcarrier data is merged and decoded, it is reported to the media access control layer of the receiving end. The proxy reconstruction unit 520 in the receiving reconstruction module 500 performs the repair operation of the closed syntax of the network abstraction layer for the truncated data packet with the internal identifier of the reduced-dimensional motion vector frame. Since the precise truncation of the bit field at the sending end changes the original byte alignment rule structure of the application layer video frame, the proxy reconstruction unit 520 supplements the original byte sequence payload alignment bit at the binary end of the truncated data packet, that is, adds a bit 1 at the end immediately adjacent to the truncation point and continues to add several bits 0 until the length requirement of the standard eight-bit byte boundary is met. The proxy reconstruction unit 520 re-encapsulates the header length identifier and the tail delimiter sequence of the network abstraction layer unit to complete the restoration of the protocol syntax structure of the video data stream between the transport layer and the application layer.

[0060] S604, the recovery of the dimension-reduced motion vector frame at the decoding end is essentially a process of replacing the residual block with inter-frame motion compensation due to missing high-frequency texture information. The proxy reconstruction unit 520 extracts the physical motion vector data retained in the dimension-reduced motion vector frame, encapsulates it into the underlying syntax field of the standard prediction macroblock, clears the coding block mode status bit, and performs all-zero initialization overwrite on the residual coefficient matrix. When the reference frame reading coordinate indicated by the physical motion vector exceeds the inherent physical resolution boundary of the video frame, the video decoder limits the out-of-bounds coordinates to the nearest edge pixel based on the boundary clipping rule and extracts the corresponding block data. The video decoder directly performs the picture block translation extrapolation operation based on the block data of the reference frame buffer pool. Its pixel reconstruction logic is calculated by the following derivation model.

[0061] ; In the formula, The spatial coordinates of the current video frame Reconstructed pixel values ​​at the location; The horizontal pixel index coordinates in the two-dimensional matrix spatial domain of the image; The vertical pixel index coordinates are given in the spatial domain of the image's two-dimensional matrix. The corresponding pixel value in the buffer pool of the previously decoded reference frame; and The horizontal and vertical physical motion vector parameters carried in the reduced-dimensional motion vector frame; variables The pixel domain error value corresponding to the residual coefficient matrix is ​​forcibly assigned a constant value of zero by the proxy reconstruction unit 520 in this step.

[0062] Application Examples: To aid in understanding the technical solution of this invention, the following is an application example of a drone performing video patrol operations in an urban building complex environment.

[0063] The system sets the maneuver threshold constant to 2.0 and the recovery threshold constant to 0.2. The absolute altitude parameter of the UAV relative to the ground is 50.0m. The discrete sampling period constant configured by the system is 0.02s. At the first moment of the flight phase, the roll rate, pitch rate and yaw rate obtained by the parameter extraction unit 120 are 0.1rad / s, 0.1rad / s and 0.2rad / s, respectively. The calculation is performed based on the three-axis physical weighting coefficients corresponding to the roll dimension, pitch dimension and yaw dimension, which are 0.2, 0.3 and 0.5, respectively, and the positive compensation constant is 0.01. At this time, the kinematic channel degradation index output by the index calculation unit 210 is 0.0005.

[0064] At the second moment, the parameter extraction unit 120 extracts a roll rate of 0.5 rad / s, a pitch rate of 0.8 rad / s, and a yaw rate of 1.5 rad / s. The sum of the numerators obtained at this time is equal to 1.367. The sum of the numerators is divided by the sum of the absolute altitude parameter and the compensation constant, which is 50.01, to obtain the current kinematic channel degradation index of 0.0273. The judgment unit 220 is then triggered to perform differential analysis on the kinematic channel degradation index of two adjacent sampling points and divide it by the sampling period of 0.02s to extract the channel change rate parameter of 1.34.

[0065] The system determines that the protocol stack reassembly is triggered and performs underlying orthogonal frequency division multiplexing parameter reconfiguration. The total number of subcarriers in the system is 1024 and the reference bit base value is 12000 bits under standard conditions. After reconfiguration, the pilot subcarriers increase to 256 and the guard empty subcarriers are set to 100. There are 668 single-symbol subcarriers remaining. Combined with the parameters that the modulation order is orthogonal phase shift keying and each symbol carries 2 bits and half of the error correction code rate, the parameter reconfiguration unit 310 calculates that the physical resource squeeze token corresponding to the effective payload reduction in a single transmission time interval is 4096 bits. The data allocation unit 420 reads the physical resource squeeze token and performs a pruning process on the reduced-dimensional motion vector frames in the subordinate buffer queue, starting from the end and equal to 4096 binary bits in length. The pruned fragmented data and key image frames are mapped to the edge service subcarrier set and the center service subcarrier set of the frequency band, respectively, for transmission.

[0066] To verify the working data of the technical solutions in the embodiments of the present invention, the following provides a comparison of experimental results combined with simulation.

[0067] Under a signal-to-noise ratio of 15dB, the traditional scheme has a bit error rate on the order of 10^-2. The scheme of this invention reduces the transmission bit error rate of core key frames to 10^-2 through a highly robust control signaling channel scheduling and video frame truncation mechanism. -4 Less than orders of magnitude.

[0068] Traditional systems experience a peak signal-to-noise ratio (PSNR) that drops below 15dB within the time window due to network packet loss, accompanied by a 2-second decoding stagnation phase. The system of this invention triggers physical layer reconstruction using angular velocity parameters and applies reduced-dimensional motion vector frames to replace non-critical frame data. At the receiving end, macroblock motion extrapolation compensation is used to complete image reconstruction. The system of this invention achieves a video PSNR greater than 28dB within the time window.

[0069] Table 1 below shows the average data of three operating indicators of the system of the present invention and the traditional control system under different flight conditions of the UAV.

[0070] Table 1: Comparison of System Operation Indicators under Different Flight Conditions According to Table 1 and Figure 6 and Figure 7 It can be seen that the radio frequency spectrum utilization efficiency of the system of the present invention is the same as that of the traditional system under the first operating condition. Under the third operating condition, the system of the present invention reduces the radio frequency spectrum utilization efficiency and controls the application layer network packet loss rate to 1.25% through the cross-layer macroblock parameter compensation mechanism. The average peak signal-to-noise ratio of the receiver of the system of the present invention is not less than 29.5dB under the three flight conditions.

[0071] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment, characterized in that the method include: Acquire the drone's raw video stream, three-axis angular velocity parameters, and absolute altitude parameters; The kinematic channel degradation index is calculated based on the triaxial angular velocity parameter and the absolute height parameter. When the time derivative value is greater than the maneuver threshold, a reorganization response is triggered. Based on the recombination response extended cyclic prefix and pilot subcarrier, the effective payload reduction is calculated, a physical resource squeeze token is generated, and a reconfiguration flag is transmitted. The original encoding is terminated, and a reduced-dimensional motion vector frame is generated using the three-axis angular velocity parameters. Based on the physical resource squeeze token, the reduced-dimensional motion vector frame is truncated, and the key image frame and the truncated reduced-dimensional motion vector frame are mapped to the service subcarrier for transmission. The receiving end demodulates the service data according to the reconfiguration flag, completes the data syntax structure, and reconstructs the video stream by combining it with the reduced-dimensional motion vector frames.

2. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The steps for obtaining the raw video stream, three-axis angular velocity parameters, and absolute altitude parameters of the UAV include: The original video stream output by the payload camera is imported into a circular buffer for cyclic overwrite writing; Extract the roll rate, pitch rate, yaw rate and absolute altitude contained in the data packets sent by the UAV flight control system to obtain the three-axis angular velocity parameters and the absolute altitude parameters; Using the hardware exposure trigger edge of the video frame as the time reference point, the acquired inertial measurement unit dataset is retrieved, and data with attached parameters that generate timestamps and satisfy the principle of minimum absolute difference with the time reference point are selected to perform hard synchronization alignment on the original video stream.

3. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The step of calculating the kinematic channel degradation index based on the triaxial angular velocity parameters and the absolute height parameters, and triggering a reorganization response when the time derivative value is greater than the maneuver threshold, includes: A three-axis physical weighting coefficient is introduced to perform a linear summation operation on the squared terms of the three-axis angular velocity parameters, and the summation result is divided by the absolute height parameter and the positive compensation constant to obtain the discrete kinematic channel degradation index; The kinematic channel degradation index corresponding to adjacent sampling points is extracted, differential operation is performed, and after mean filtering and smoothing, it is divided by the discrete sampling period constant to extract the channel change rate parameter as the time differential value. When the channel change rate parameter is greater than the set maneuver threshold, a low-level hardware interrupt trigger instruction is synchronously broadcast to the system physical layer and video buffer unit as the reassembly response.

4. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The step of calculating the payload reduction amount, generating a physical resource squeeze token, and transmitting a reconfiguration flag bit based on the recombined response extended cyclic prefix and pilot subcarriers includes: In response to the underlying hardware interrupt trigger instruction, the cyclic prefix length of the underlying orthogonal frequency division multiplexing module is adjusted to an extended ratio, and the pilot subcarriers are adjusted from a comb-like spacing distribution mode to a dense block distribution mode. Based on the increase in cyclic prefix time length and pilot subcarrier density, combined with modulation order and error correction code rate, the effective payload reduction within a single transmission time interval is calculated to obtain the physical resource squeeze token measured in binary bits. A fixed set of subcarriers is allocated in the edge boundary region of the operating frequency band to construct an extremely robust control signaling channel, and the reconfiguration flag bit of the associated cyclic prefix length index value and pilot mode index value is encapsulated into the extremely robust control signaling channel and transmitted.

5. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The step of suspending video encoding and generating a dimension-reduced motion vector frame using the triaxial angular velocity parameters includes: The transmission path of the original video stream to the standard video encoder is blocked, and the ongoing discrete cosine transform and quantization calculation pipeline operation of the current image sequence is terminated. Clear the residual frequency domain coefficient register associated with the current video frame and retain the sequence parameter set and image parameter set; Based on the rigid body kinematics model, the global physical motion vectors of the image in the horizontal and vertical pixel directions are calculated according to the yaw and pitch angular velocities included in the three-axis angular velocity parameters, and then filled into the motion vector special field corresponding to the encoding protocol to generate the dimensionality-reduced motion vector frame.

6. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The step of truncating the reduced motion vector frame based on the physical resource squeezing token includes: Parse the header of the data packet received by the media access control layer, transfer the data packet identified as the key image frame to the master transmission queue, and transfer the data packet identified as the dimension-reduced motion vector frame to the truncation processing buffer queue. Determine whether the actual binary length of the current reduced-dimensional motion vector frame is greater than or equal to the bit value of the physical resource squeeze token identifier; When the actual binary length is greater than or equal to the bit value, the binary bits corresponding to the token length are truncated bit by bit from the end of the reduced-dimensional motion vector frame data packet toward the head; when the actual binary length is less than the bit value, the current reduced-dimensional motion vector frame is discarded and the remaining token value is deferred to the next data packet.

7. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The steps of mapping key image frames and truncated, reduced-dimensional motion vector frames to service subcarriers for transmission include: Map the key image frames in the master control transmission queue to the set of central service subcarriers in the center frequency band corresponding to the total subcarrier sequence of the system's working bandwidth; The truncated motion vector frame fragment data is mapped to the set of edge service subcarriers located at both ends of the total subcarrier sequence and adjacent to the inner frequency band of the extremely robust control signaling channel; The frequency domain data block that has completed asymmetric mapping is converted into a time domain transmit signal through inverse fast Fourier transform and then used for air interface transmission.

8. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The step of demodulating service data at the receiving end based on the reconfiguration flag bit includes: At the receiving end, the input discrete time-domain sampled data stream is split into independent listening window pipelines and service data window pipelines, and a buffer is configured at the front end of the service data window pipeline to perform clock cycle delay dwell; The reconfiguration flag data packet is extracted and parsed in the extremely robust control signaling channel frequency band using the aforementioned listening window pipeline; The index parameters contained in the reconfiguration flag are hard-written into the core state machine of the service data window to update the demodulation configuration parameters. Based on the updated configuration parameters, time-domain data is extracted from the buffer, the corresponding cyclic prefix is ​​stripped, and a fast Fourier transform is performed to obtain the service subcarrier data in the frequency domain.

9. The software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment according to claim 1, characterized in that, The steps of completing the data syntax structure and reconstructing the video stream by combining the reduced-dimensional motion vector frames include: Add a 1 bit and append several zero bits to the end of the truncated data packet binary truncation point in the demodulated output to supplement the original byte sequence payload alignment bit, and re-encapsulate the header length identifier and tail delimiter sequence of the network abstraction layer unit. The global physical motion vector contained in the reduced motion vector frame is extracted and filled into the standard prediction macroblock, and the associated residual coefficient matrix is ​​initialized with all zeros and covered. Based on the global physical motion vector, a picture block translation and extrapolation operation is performed on the corresponding block data in the reference frame buffer pool of the receiving end, and the reconstructed video stream is output.

10. A software-defined radio frequency adaptive image transmission system for low-altitude unmanned equipment, using the software-defined radio frequency adaptive image transmission method for low-altitude unmanned equipment as described in any one of claims 1 to 9, characterized in that, The system specifically includes: The data acquisition module is used to acquire the drone's raw video stream, three-axis angular velocity parameters, and absolute altitude parameters; The baseband processing module is used to calculate the kinematic channel degradation index based on the triaxial angular velocity parameters and the absolute height parameters, and to trigger a reorganization response when the time derivative value is greater than the maneuver threshold. The physical layer reconfiguration module is used to expand the cyclic prefix and pilot subcarrier based on the reconfiguration response, calculate the payload reduction amount, generate a physical resource squeeze token, and transmit a reconfiguration flag bit; The joint scheduling module is used to suspend video encoding, generate a reduced-dimensional motion vector frame using the three-axis angular velocity parameter, truncate the reduced-dimensional motion vector frame according to the physical resource squeeze token, and map the key image frame and the truncated reduced-dimensional motion vector frame to the service subcarrier for transmission. The receiving reconstruction module is used to demodulate service data at the receiving end according to the reconfiguration flag, complete the data syntax structure, and reconstruct the video stream by combining the reduced-dimensional motion vector frames.

Images