A one-way transmission ultra-narrow band preamble processing method, data processing center and system for radio frequency positioning

Abstract

The application discloses a one-way transmission ultra-narrow-band preamble processing method for radio frequency positioning, a data processing center and a system, and relates to the technical fields of satellite communication and radio signal processing. The application aims to solve the problems of high-precision ranging, Doppler error compensation and integer ambiguity resolution in an ultra-narrow-band satellite Internet of Things system. The technical solution points of the application are as follows: a three-section preamble composed of a continuous wave carrier acquisition section, a forward and reverse interlaced linear frequency modulation time-frequency coarse estimation section and a continuous wave double-tone fine measurement section is used for terminal blind transmission; a receiving base station extracts a coarse arrival time through geometric feedforward compensation, time-frequency decoupling of the time-frequency coarse estimation section, time-frequency dual averaging and quadratic polynomial fitting; the double-tone fine measurement section adopts a symmetric time-division double-tone structure, and high-precision absolute propagation time delay resolution is realized through two-stage joint deblurring; the processing architecture and the observation system are decoupled, and a positioning mode is adaptively switched according to the number of visible base stations. The application can realize tens of meters of ranging accuracy, reduce the hardware cost of a terminal, has a fuse capacity and is suitable for low-cost terminal positioning based on a medium-low orbit satellite and a high-dynamic aircraft platform.

Description

Technical Field

[0001] This invention relates to the field of satellite communication and radio signal processing technology. Specifically, it relates to a one-way transmission ultra-narrowband preamble processing method, data processing center and system for radio frequency positioning. It is particularly suitable for receiving base stations deployed on medium and low orbit satellites, extraterrestrial orbiters or high dynamic spacecraft platforms, etc., to perform precise timing, ranging and positioning of low-cost terminals in high dynamic scenarios. Background Technology

[0002] With the rapid development of Low Earth Orbit (LEO-IoT) satellite Internet of Things (LEO-IoT) and 5G / 6G non-terrestrial networks (NTN), radio frequency tracking and positioning services for IoT terminals that do not rely on GNSS, are low-cost, low-power, and suitable for large-scale deployment have attracted widespread attention. Its foundation lies in achieving high-precision Time-of-Flight (TOF) measurement and Doppler Frequency Offset (FOA) estimation. In highly dynamic satellite or spacecraft communication scenarios, communication links face extremely high Doppler frequency shifts and their second-order rate of change, leading to severe distortion and aliasing of the signal in both the time and frequency domains.

[0003] Traditional narrowband IoT space-to-ground communication is limited by ultra-long links, low power consumption, and extremely narrow shared spectrum resources, making it extremely difficult to achieve precise time and frequency synchronization in short burst data (SBD) communication. Existing systems often rely excessively on closed-loop bidirectional interaction between the terminal and the base station, or require the terminal to transmit burst signals multiple times over a long arc for multi-point fitting. This results in high power consumption and latency, and the multiple signal positioning method also significantly affects the positioning accuracy of the mobile terminal. In addition, in low-cost terminal deployment scenarios, the temperature drift and frequency drift phase noise introduced by low-end crystal oscillators are severely intertwined with ultra-high Doppler parameters, making it extremely difficult for traditional RF processing systems to achieve high-precision ranging and integer ambiguity resolution in short burst data under such high dynamic and low signal-to-noise ratio environments, severely restricting the accuracy of subsequent positioning calculations. Summary of the Invention The technical problem to be solved by the present invention

[0004] In view of the limitations of existing technologies, this invention aims to solve the following core engineering and technical problems: (1) High-precision ranging challenge under ultra-narrowband communication: Under the constraints of extremely narrow bandwidth and extremely weak signal, we break through the traditional physical understanding that "ultra-narrowband systems cannot perform high-precision ranging" and achieve high-precision absolute propagation time delay (TOF) measurement at the tens of meters level.

[0005] (2) Low-cost unidirectional uplink terminal, difficulty in compensating for Doppler dynamic error: The high-speed operation of low-orbit satellites generates large frequency offset and high Doppler acceleration. Traditional narrowband unidirectional uplink systems have difficulty accurately measuring and compensating for this dynamic error under the condition of low-cost terminal crystal oscillator, resulting in loss of ranging and velocity coherence.

[0006] (3) Carrier phase integer ambiguity resolution relies excessively on external aids: Traditional precision measurement methods require pseudocode ranging or conventional coarse measurements to determine integer ambiguity or narrow down the candidate interval for integer ambiguity. This invention aims to achieve highly robust absolute delay resolution without integer ambiguity jumps in high dynamic and low signal-to-noise ratio environments.

[0007] (4) High terminal hardware cost and high power consumption: Eliminate the terminal's dependence on high-cost GNSS chips, closed-loop handshake communication modules and high-stability crystal oscillators, and realize coherent precision measurement of uplink single burst blind transmission signals. Technical solution summary

[0008] The present invention provides a one-way transmission ultra-narrowband preamble processing method, data processing center, and system for radio frequency positioning, comprising the following core steps: (1) Terminal-side blind transmission: Under conditions of no downlink auxiliary information and no time-frequency synchronization between the terminal and the base station, the IoT terminal transmits a three-segment ultra-narrowband preamble signal in a single, unidirectional, completely blind transmission. The preamble consists of a continuous wave carrier acquisition segment, a coarse time-frequency estimation segment of interleaved linear frequency modulation (LFM Chirp), and a continuous wave dual-tone precision measurement segment cascaded together. The RF front-end is immediately shut down after the single blind transmission is completed.

[0009] (2) Receiving base station side capture and feedforward compensation: The receiving base station deployed on the dynamic platform uses Fast Fourier Transform (FFT) to capture and lock the carrier frequency and establish the baseband digital downconversion center frequency; it uses the known spatial geometric parameters of the receiving base station to perform geometric feedforward Doppler coarse compensation on the full frame signal and shrinks the processing window.

[0010] (3) Coupling and coarse measurement of time-frequency coarse estimation segment: Matched filtering and frequency domain interpolation are performed using the structure of alternating forward and reverse frequency sweeps and coincident time-domain geometric centroids of the time-frequency coarse estimation segment. The extracted equal-interval time-domain phase sequence is fitted with a quadratic polynomial to estimate the residual first-order frequency offset and continuous second-order Doppler acceleration in closed loop. The distance and Doppler coupling are eliminated by the deslant dual averaging operator to calculate the coarse arrival time.

[0011] (4) Symmetrical dual-tone multi-channel feedforward error self-counterweighting: In response to the hardware channel inconsistency and feedforward error introduced by the multi-channel receiving architecture of the receiving base station, each group of time-division dual-tones in the dual-tone precision measurement segment is configured as a symmetrical dual-tone structure mirror-symmetric about the carrier center frequency; by utilizing the phase co-mode characteristics of the symmetrical dual-tone signal in multi-channel transmission and RF front-end, a channel error differential self-counterweighting model is constructed in the data processing center, which cancels and eliminates the multi-channel feedforward phase error without the need to introduce an external calibration source, thereby achieving ultra-precise purification of the original measurement phase of each single-tone signal.

[0012] (5) Global projection and cascaded defuzzification of the precision measurement segment: Based on the calculated Doppler parameters, a continuous phase time-domain projection function is constructed, and the phase observation values ​​of each time-division dual-tone sub-segment of the dual-tone precision measurement segment are uniformly back-projected and normalized to the global time-domain anchor point; the arbitrary-level defuzzification algorithm is called to perform defuzzification search, and the two-level defuzzification system of macroscopic unfuzzy search space and robust Chinese Remainder Theorem (Robust CRT) is preferred to calculate the high-precision absolute propagation delay (TOF) from the terminal to the receiving base station.

[0013] (6) Solution topology and observation system matching solution: According to the solution topology configuration, the system combines the absolute propagation delay extracted above with the time difference of arrival / Doppler frequency offset (TDOA / FDOA) when there are two receiving base stations, and adopts spatial resection system or multi-solution weighted fusion solution when there are multiple receiving base stations. When there is a single station, it combines one or more of the observations in the spatial auxiliary constraints (such as two-dimensional angle of arrival AoA, digital prior map constraints, temporal displacement hard constraints, etc.) to construct a set of positioning solution equations and solve for the geographic spatial coordinates of the output terminal.

[0014] The core essential technical solution of this invention is preamble signal processing and time delay and arrival frequency calculation. Multi-geometric configuration weighting and dynamic allocation of computational tasks are only used for optional post-processing accuracy optimization and are not essential steps for implementing this invention. Conducting error simulations for various high-dynamic scenarios is a conventional engineering method, and those skilled in the art can adjust and adapt it according to different scenarios based on existing engineering experience. Decoupled topology configuration of computing architecture and observation system

[0015] To adapt to different physical hardware platforms, computing resource limitations, and other business requirements, the processing flow of this invention achieves complete decoupling and parallel operation in solving the topology and the space observation system: (1) Solving topology (operational architecture): Supports the following three parallel operational architectures: Pure receiving base station local processing architecture: All fuzzy decomposition and spatial multidimensional equation solving are completed locally on-site at the receiving base station on the low-Earth orbit platform or dynamic aircraft, and the terminal position coordinates are directly output.

[0016] Pure data center centralized processing architecture: The receiving base station is only responsible for the FFT capture and phase sequence extraction of the underlying signal physical layer, and transmits the original observation feature parameters in full to the ground / space-based centralized data center through the backhaul link to complete the final solution.

[0017] Distributed processing architecture: The system dynamically switches task boundaries and task slicing between the receiving base station and the data center based on the current satellite-to-ground link bandwidth resources, latency threshold, and the current computing power configuration of the receiving base station.

[0018] (2) Space observation system: The final positioning equation set will be independently switched to one of the following three systems based on the number of effective receiving base stations that are currently visible at any given time: pure single receiving base station positioning system, pure multi-receiving base station joint positioning system, or hybrid positioning system. Preamble signal structure design Preamble signal structure design

[0019] In view of the reality that low-cost terminals have low crystal oscillator accuracy (10 ~ 50 ppm), large initial Doppler frequency offset and low signal-to-noise ratio (SNR at the receiving antenna port is close to 0dB), this invention adopts a three-segment composite preamble structure with functional layer design.

[0020] (1) Carrier acquisition segment: Single-frequency continuous wave signal (CW segment) is used. The spectral characteristics of its concentrated energy in the frequency domain are used in conjunction with the panoramic search of the receiving base station to lock the estimated carrier frequency, establish the baseband frequency reference, and narrow the subsequent analysis window.

[0021] (2) Time-frequency coarse estimation segment: Linear frequency modulation (LFM Chirp) symbols (such as UDUD structure) with equal intervals and an even number of forward sweep (U) and reverse sweep (D) time domain alternating are used. The physical bandwidth of the sweep is limited to no more than 10 kHz (preferably 1 kHz to 2 kHz). The forward and reverse sweep symbols adopt a symmetrical configuration with completely coincident centroids in the time domain. The time delay (TOA) and frequency offset (FOA) are completely decoupled through dual averaging, and the Doppler acceleration is extracted by fitting a quadratic polynomial of multiple symbols.

[0022] (3) Dual-tone precision measurement segment: A time-division dual-tone (DTT) CW structure is adopted, comprising multiple time-division dual-tone sub-segments. Each discrete single-tone component occupies an extremely narrow signal area. Each sub-segment contains a reference component located at the center reference frequency f0 = 0 Hz, and a component located at the discrete offset frequency +f i or -f i The measurement components at the location are determined using a two-level joint solution fuzzy solution designed in this patent. A Synergistic Explanation of the Intrinsic Attributes of Ultra-Narrowband Dual-Tone Precision Measurement Section and Large-Scale Virtual Synthetic Bandwidth

[0023] The preamble of this invention possesses stringent ultra-narrowband (UNB) characteristics: the actual sweep width of the linear frequency modulated signal in the time-frequency coarse estimation segment is limited to an extremely narrow range; the dual-tone fine measurement segment employs multiple sets of time-division dual tones, forming a sparse discrete frequency point structure. A large frequency difference is set between each set of single tones to satisfy the number theory conditions for deambiguity, but this frequency difference is only a gap between frequency points, and the intrinsic bandwidth occupied by each single tone itself still maintains an ultra-narrowband level. Waveform definition

[0024] (1) Carrier acquisition segment waveform:

[0025] Where T acq f is the carrier acquisition segment width. c This refers to the nominal carrier frequency.

[0026] (2) Time-frequency coarse estimation segment waveform: The expression for the k-th Chirp symbol (k=0,1,…,N−1) is: Where B is the sweep bandwidth, and T s The symbol duration is specified. Positive sweep (U) takes a positive sign, and negative sweep (D) takes a negative sign. The symbol direction is arranged in a UDUD cycle.

[0027] (3) Waveform of the dual-tone precision measurement segment: The expression of the m-th sub-segment is: Where T CW The integration time for each sub-segment, f 0,m =0 is the center reference frequency, f 1,m Let m be the offset frequency of the m-th sub-segment. Geometric feedforward Doppler compensation

[0028] The Doppler acceleration introduced by the high-speed operation of low-Earth orbit satellites causes severe distortion and time-domain broadening of the matched filter peak in linear frequency hopping signals. The onboard receiving base station retrieves the precise orbital position P provided by the onboard GNSS in real time. sat (t) and velocity vector V sat (t), combined with the pointing unit vector u of the currently received beam. beam Calculate the unit vector r in the direction of the beam center. center Solving for the theoretical Doppler frequency shift at the beam center: For f d,center (t) Perform time difference analysis to extract the theoretical Doppler acceleration a d,centerUsing this as a feedforward input to the de-skew processor, most of the dynamic Doppler acceleration is coarsely reduced, compressing the spatial geometric residual acceleration at the beam edge to within the linear estimation range of the signal fitter. Receiver base station FFT frequency sweep acquisition and carrier frequency locking mechanism

[0029] This method introduces a burst carrier acquisition and distribution mechanism based on panoramic frequency domain scanning. The specific implementation process includes the following core steps: (1) Panoramic frequency domain scanning: The receiving base station sends the received burst signal into the Fast Fourier Transform (FFT) operator and performs frequency domain discrete spectrum scanning within the preset processing window total bandwidth to map the weak time domain signal into discrete power spectrum (frequency bin) in the frequency domain.

[0030] (2) Peak capture and decision: Search and extract the global maximum value of the power spectrum within the processing window. When it is greater than the preset detection threshold, the burst signal is determined to be successfully captured, and the frequency point corresponding to the maximum value is synchronously latched as the actual arrival carrier frequency of the carrier capture segment.

[0031] (3) Reference Frequency Locking and Distribution: The actual arriving carrier frequency is established as the initial reference frequency of the receiver, and the baseband digital downconversion center frequency of the digital downconversion (DDC) module is established using this frequency. The initial reference frequency is synchronously distributed to the digital downconversion (DDC) module and the Doppler acceleration compensation module to remove the full-frame Doppler residual, so that the downconverted signal is aligned to the preset center reference frequency in the digital baseband domain to guide subsequent phase purification. Dynamic filtering and purification mechanism based on Doppler acceleration compensation

[0032] This method introduces a dynamic narrowband filtering mechanism based on Doppler acceleration compensation, specifically designed for processing the carrier acquisition and dual-tone precision measurement segments in burst signals. The digital filter's processing window undergoes two levels of dynamic control depending on whether deskewing compensation is introduced: (1) First window (wide window): In the initial stage before compensation is performed, the signal has a large dynamic drift in the frequency domain. At this time, a wider first window is used to wrap the signal so as to fully contain the energy of the signal that shifts rapidly with Doppler acceleration.

[0033] (2) Second window (narrow window): After compensation is performed, the rapid dynamic frequency offset drift trajectory caused by the relative motion of the signal with the satellite and the ground is leveled. The system starts the narrow window purification logic, which dynamically narrows the digital filtering processing window into a narrower second window with a total processing bandwidth of only 10Hz to 200Hz.

[0034] By dynamically intercepting the narrowband filter, most of the out-of-band thermal noise and multipath interference within the total bandwidth can be filtered out, significantly improving the effective input signal-to-noise ratio (SNR) of the signal, effectively suppressing noise spurious peaks and greatly reducing the variance of the measurement phase thermal noise. Time-frequency coarse estimation segment matched filtering and time-frequency decoupling dual averaging

[0035] After geometric feedforward coarse compensation, this invention utilizes an interleaved Chirp structure to achieve a fully decoupled representation of range and Doppler. For the forward sweep (U) symbol, the Doppler frequency offset fd causes a time shift Δt in its matching peak. U =-f d / K, (K=B / T s (This refers to the sweep slope); for the reverse sweep (D) symbol, a time shift Δt is generated. D =+f d / K. Let the actual time delay be τ and the Doppler frequency offset be f. d The measured peak times are t, respectively. U =τ-f d / K and t D =τ+f d / K. Parameter decoupling is achieved through sum-difference duality: Extract the peak complex phase ϕ of the N Chirp symbols matching the output. i Construct an equally spaced time-domain phase observation sequence: Where t i = i·t s At the symbol center moment, The phase sequence is fitted with a quadratic polynomial.

[0036] By performing a quadratic polynomial fitting on the peak complex phase of multiple linear frequency modulation (FM) symbols, the baseband delay phase term φ0 in the fitting parameters is extracted. This allows for high-precision purification of the initial value in the phase domain, outputting a fine delay value that meets the symbol timing window alignment requirements for symbol timing alignment in the dual-tone precision measurement segment. In the aforementioned quadratic polynomial fitting, the first-order Doppler frequency offset f in the parameter vector θ is synchronously and jointly estimated by solving a simultaneous solution to the phase observation sequence. d With second-order Doppler acceleration a d This is used to strip the full-frame Doppler residuals.

[0037] Parameter vector φ=[φ 0, f d, a d ] TThe closed Cramer-Rao lower bound (CRLB) corresponding to the inverse of the Fischer information matrix (FIM) is: (1) Baseband delay phase term: (Corresponding to coarse delay standard deviation) ): (2) First-order Doppler frequency offset term: ; (3) Second-order Doppler acceleration term: ; Symmetrical double tone difference and error offset elimination

[0038] This invention eliminates terminal homogeneity error and reception multi-channel inconsistency error respectively through two-tone differential and symmetrical two-tone design: Terminal co-source error elimination: The two subcarriers of the two-tone signal share the terminal's transmit path and local oscillator, and their temperature drift, phase noise, and transmit delay are highly co-sourced. The receiver performs phase subtraction (differential) operation on the two subcarriers to basically offset the above-mentioned terminal co-source error, thereby extracting delay information with high accuracy.

[0039] Multi-channel error cancellation: Phase deviations caused by inconsistencies in the reception of multiple channels are eliminated through a symmetrical frequency domain structure of symmetrical two-tone signals. Utilizing the physical characteristic of the symmetrical distribution of subcarriers about the center frequency, the phase deviations introduced by the multiple channels are mutually canceled during the solution process, thereby preserving the phase relationship of the baseband signal.

[0040] This invention designs a three-tone reference waveform using a center frequency paired with symmetrical offset frequencies. To reduce peak-to-average power ratio (PAPR) and terminal hardware overhead, the physical layer effectively splits it into two equal-length, time-division-alternating symmetrical two-tone transmissions, preferably with a single segment duration of a few milliseconds to tens of milliseconds. Within this extremely short time, channel temperature drift and crystal oscillator drift can be approximated as constants. Based on this, the system implements hierarchical differential: intra-segment two-tone differential eliminates common-mode error of the same crystal oscillator; symmetrical two-tone secondary differential eliminates asymmetric group delay and phase disturbance caused by channel temperature drift, ensuring high-precision absolute ranging.

[0041] (1) Multi-channel antenna scheme (such as phased array, antenna array, MIMO, etc.): Use intra-group positive and negative symmetrical dual-tone differential technology. Extract the center reference frequency f0 and the measurement offset frequency +f in each sub-segment. i and -f i Calculate the single-sided differential phase based on the received phase: Assuming the initial phase difference of the terminal transmission is 0, the common non-ideal phase physical error introduced by the onboard multi-channel to adjacent frequency points is φ. chan The absolute time delay of spatial propagation is τ. Expanding the receiving phase relationship: The differential phase is converted to the time delay domain, and a dual average of the two sets of one-sided time delays is performed: Mathematical proofs show that, through symmetrical two-tone differential modulation, the common common-mode phase error φ of multi-channel hardware can be reduced. chan It is completely offset and eliminated, and compared with the three-tone, the variance of thermal noise is significantly reduced.

[0042] (2) Single-channel antenna scheme: If the system uses a single-channel antenna, the adaptive degradation adopts ordinary two-tone technology, directly calculates the differential phase between the single-side baseband offset frequency and the center reference frequency, and outputs the time delay remainder sequence Δφ=φ(f0+f i )-φ(f0)=2πf i τ+δφ eliminates the computational overhead of positive and negative symmetrical hedging. Fuzzy solution of two-level joint solution The present invention preferably adopts a two-level joint defuzzification architecture based on geometric boundary constraints to completely avoid the error cascading amplification caused by traditional hierarchical defuzzification. Its core processing logic is as follows:

[0043] (1) Establish a macroscopic, unfuzzy search space: The first-level search space described in this invention does not rely on the coarsely measured Time of Arrival (TOA), but rather on an absolute macroscopic search window W constructed based on the satellite-to-ground geometric distance constraint. virt establish.

[0044] The distance between satellites and the Earth is constrained by satellite orbits, and there exists a maximum line-of-sight distance D. max , zenith closest sight distance D min Two boundaries. The difference between the two represents the maximum range of distance fluctuation between the ground and the satellite. A macroscopic, unambiguous search window W is set. virt If the difference is greater than this value, the entire actual propagation distance range can be covered.

[0045] Configure the first set of frequency differences Δf1 and the second set of frequency differences Δf2 for the dual-tone precision measurement segments. Utilize the greatest common divisor (gcd(Δf1, Δf2)) of these two frequencies in number theory to synthesize an outer macroscopic unambiguous virtual solution window that satisfies the above conditions. .

[0046] Due to W virt > [D min D max Therefore, any actual distance falls within [D]. min Dmax Within this absolute physical space, the phase correspondence of the first-level precision measurement signal is physically unique, and there will absolutely be no full-cycle overlap (i.e., no phase ambiguity).

[0047] Therefore, the algorithm can directly solve the problem of size W without performing any integer-cycle deblurring operations. virt Within the macroscopic window, precise deblurring of subsequent stages (high frequency difference, short equivalent virtual wavelength) establishes a unique coherent reference.

[0048] (2) Multi-frequency maximum likelihood one-time solution: In the W virt Within the unambiguous spatial range, the integrated simultaneous introduction includes the highest frequency difference Δf m Multiple sets of differential phase measurements φ, including those corresponding to a precision ranging narrow window. i =wrap(2πΔf i τ+n i ), where n i Let be the measurement noise of the i-th differential phase. Construct an objective function that minimizes the weighted sum of squared circular residuals, weighted by the reciprocal of the variance of the measurement noise: By performing an integrated grid search within a global window using maximum likelihood estimation (MLE), the precise absolute propagation time delay (TOF) is calculated in a single mapping, skipping intermediate levels.

[0049] (3) Signal abnormality dynamic fuse: A signal abnormality dynamic fuse mechanism is introduced in the process of defuzzification search. If the calculated generalized phase residual exceeds the preset closed-loop anti-multipath fuse phase residual threshold (limited to 0.25 cycles), it is determined that the current observation is subject to strong multipath interference. The system will immediately and automatically trigger the fuse and switch to the pure time delay coarse ranging result obtained by fitting the time-frequency coarse estimation segment to ensure survivability.

[0050] 3.2.10 Transmitter and Receiver Compensation Mechanism Based on Preset Multi-tone Phase Difference

[0051] The preamble signal has radio frequency protection intervals between the coarse time-frequency estimation segment and the fine two-tone measurement segment, as well as between each two-tone sub-segment. To reduce terminal control complexity and ensure phase resolvability, the numerically controlled oscillator (NCO) and continuous wave signal inside the terminal are allowed to use differentiated operating mechanisms within the protection intervals: Full-segment reset mode: Allows all NCO and continuous wave signals to perform shutdown and restart operations within the protection interval; Phase-correlated interruption mode: The first numerically controlled oscillator corresponding to the center frequency f0 is allowed to continue to operate continuously, while the second numerically controlled oscillator corresponding to the offset frequency f1 is interrupted during the protection interval; wherein, at the start time of the second numerically controlled oscillator resuming transmission, its initial phase is forcibly reset to an initial state with a known fixed phase offset relationship with the first numerically controlled oscillator.

[0052] Mechanisms that allow shutdown and restart operations and require the system to reunify the phase reference across sub-segments include: (1) Terminal segmented transmission: At the start of transmission of each two-tone segment, the terminal controls the relevant NCO to perform initialization or phase reset, forcing the controlled frequency point to maintain a preset absolute initial phase or a known fixed relative phase offset Δψ with the continuous operating frequency point. i Launch.

[0053] (2) Receiver differential inverse compensation: The receiving base station or ground data processing center extracts the original measured phase φ of each subcarrier. rx,i Call the locally synchronized pre-stored corresponding sub-segment fixed initial phase offset Δψ i Perform reverse subtraction hedging compensation: Note: No phase compensation is required for frequencies that operate continuously without phase adjustment.

[0054] (3) Reference alignment: φ' after compensation rx,i They are realigned to the same virtual coherent time axis, thereby eliminating phase discontinuities and solution ambiguities across time periods caused by NCO interruptions, resets, or jumps, and are directly input into the subsequently selected Robust CRT operator. Joint positioning and observation system of receiving base stations

[0055] The data processing center independently switches or adaptively matches the corresponding geometric positioning equations to solve for the final location based on the number of currently visible effective receiving base stations (platforms): (1) Multi-receiving base station cooperative positioning mode (two or more): When there are two visible base stations, the corresponding set of equations is established for coordinate calculation by either the one-way absolute propagation delay (TOF) spatial spherical intersection system or the time difference of arrival / Doppler frequency offset (TDOA / FDOA) time-frequency joint intersection system.

[0056] When the number of visible base stations is greater than or equal to three, an analytical solution is obtained by simultaneously solving the system of equations for the multi-faceted intersecting spherical space using a pure absolute propagation delay spatial reintersection method; alternatively, all visible base stations are paired up or independently divided and solved separately to obtain multiple spatial candidate location solutions X.k Subsequently, the error variance σ of each solution was quantitatively evaluated based on the real-time extracted signal-to-noise ratio (SNR) value and the spatial geometric configuration. 2 k Information weight W k =1 / σ 2 k Perform a weighted normalized average.

[0057] (2) Single receiving base station positioning mode (single visible degradation condition): When the number of visible base stations degrades to only one, the system introduces spatial auxiliary constraints. These spatial auxiliary constraints include: radio spatial measurement constraints (based on array-measured 2D angle of arrival (AoA), spatial spectrum estimation, etc.), antenna physical beam constraints (beam projection footprint boundaries on the ground, etc.), digital prior map constraints (2D / 3D high-precision road network maps, DEM surface height constraints, etc.), or temporal behavior characteristic constraints (dynamic displacement hard constraints based on terminal limit speed and message time intervals, etc.). The absolute propagation delay of a single base station is combined with one or more of the above auxiliary constraints to perform spatial intersection calculations to solve for the terminal's geographical location. Kalman filtering can then be used for smoothing optimization in conjunction with historical temporal trajectories.

[0058] 3.2.12 Data Processing Center Distance Aberration Unified Normalization

[0059] The coarse estimation segment of the preamble and multiple sets of two-tone segments are transmitted at different times. The receiving platform is in a high-speed motion state, and the time delay varies at different times. If the phase of each time period is directly used to solve the problem, it will introduce a significant distance offset error.

[0060] (1) Establishment of global anchor point in time domain: The receiving base station captures the preamble signal at the frame synchronization time and sets the time domain geometric centroid of the time-frequency coarse estimation segment as the global time domain anchor point t0. The time of this anchor point is transmitted down along with the observation value.

[0061] (2) Distance aliasing back projection normalization: The data processing center uses the first-order Doppler frequency offset f extracted from the time-frequency coarse estimation segment. d and the second-order Doppler acceleration a constructed by cascaded refinement d Construct a time-domain projection function for continuous phase: For the i-th in t i Phase observation φ of segmented two-tone precision measurement segment at different times i Perform reverse spatiotemporal translation correction and normalize to the anchor time t0: (3) Consistency solution: φ that eliminates spatiotemporal aliasing and satisfies one-dimensional spatiotemporal consistency.i,proj The carrier phase is substituted into the optimized global grid search operator to calculate the precise Time of Flight (TOF) and complete the final geographic coordinate output.

[0062] Beneficial effects

[0063] Compared with the prior art, the present invention has the following significant advantages: (1) High-precision absolute ranging capability at tens of meters: It breaks through the limitation that ultra-narrowband systems cannot measure distances precisely. Under the harsh conditions of extremely narrow physical bandwidth (such as 1.6kHz), the optimized global grid search algorithm is used to break the phase integer ambiguity, and high-precision absolute propagation time delay (TOF) measurement at tens of meters is realized.

[0064] (2) Extremely low terminal hardware cost and power consumption: The complex synchronization, purification and calculation algorithms are all moved to the base station and ground processing center, so that the terminal is completely free from the dependence on high-cost GNSS chips, downlink signal receiving circuits, high-stability crystal oscillators and terminal-side Doppler compensation. Time and distance measurement can be completed with a single burst blind transmission.

[0065] (3) Powerful hardware non-ideal phase self-calibration and offsetting capability: The unique symmetrical time-division multi-tone waveform topology of the preamble, combined with the ground reverse phase difference compensation mechanism, not only allows the terminal transmitter NCO to operate intermittently to save power, but also completely adaptively offsets the common hardware channel group delay error introduced by the onboard multi-channel phased array antenna due to amplitude and phase inconsistency, temperature drift, and slight mechanical surface changes, eliminating the need for on-orbit calibration and maintenance.

[0066] (4) Two-level Doppler full-link suppression under high dynamic environment: Through the two-level constraint of "spatial geometric feedforward coarse compensation + equal interval phase sampling quadratic polynomial fitting fine estimation", the Doppler acceleration and first-order linear distance aliasing caused by high spatial dynamics are completely stripped and canceled, ensuring the long integral coherence purification of the signal.

[0067] (5) Excellent concurrent access flexibility and cross-constellation generalization: It adopts a completely unidirectional, single-transmission blind transmission system with zero downlink interaction overhead, which greatly releases burst communication capacity. Its preamble processing framework is completely decoupled from the observation system, and can seamlessly and adaptively adapt to radio frequency positioning systems at different altitudes in low and medium orbits, which has extremely high engineering universality value. Attached Figure Description

[0068] Figure 1 Schematic diagram of a three-segment burst preamble signal structure.

[0069] Figure 2 Flowchart for signal processing and extraction of time and distance measurement feature parameters.

[0070] Figure 3: Schematic diagram of the mechanism for establishing a macroscopic unfuzzy search space.

[0071] Figure 4 Adaptive geometric solution system discrimination and coordinate solution positioning flowchart.

[0072] Figure 5 : Overall architecture diagram of the radio frequency positioning system. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the following embodiments are merely illustrative of one implementation method of this invention and should not be construed as limiting the scope of patent protection claimed by this invention. Example 1: Precise Timing and Ranging Measurement Using Ultra-Narrowband Preamble Based on Low-to-Medium Orbit Dual-Satellite System

[0074] This embodiment uses a two-satellite observation scenario in a global low-Earth orbit constellation as an example to detail the execution process of the preamble processing method of the present invention in obtaining the precise absolute propagation time delay (TOF) and first-order Doppler frequency offset (FOA) characteristic parameters. The timing and ranging system used in this embodiment utilizes the principle of common-mode cancellation of common-source errors, which places extremely low demands on the hardware and software of the ground terminal. System Architecture

[0075] The system architecture of this embodiment is shown in Figure 5. The system consists of three main parts: a label terminal transmitter 500, a receiving base station 510, and a data processing center 520.

[0076] The terminal transmitter 500 includes a memory 5001, a processor 5002, a numerically controlled oscillator 5003, and a radio frequency transmission link 5004. The memory 5001 stores the terminal-side control software, and the processor 5002 reads and runs the software to synchronously control the numerically controlled oscillator 5003 to generate a three-segment ultra-narrowband preamble baseband waveform. After radio frequency modulation and power amplification are completed by the radio frequency transmission link 5004, the signal is transmitted unidirectionally to the receiving base station in the form of wireless radio frequency, thus fully realizing the asynchronous uplink unidirectional transmission process.

[0077] The receiving base station 510 is equipped with an antenna / antenna array 5101, an RF receiving link 5102, a baseband processor 5103, and a memory 5104. After the air RF signal is received by the antenna / antenna array 5101, it is sent to the RF receiving link 5102 to complete RF front-end processing such as low noise amplification and down-conversion, and outputs the baseband signal to the baseband processor 5103. The memory 5104 stores the base station receiving processing software. The baseband processor calls the software to sequentially execute baseband processing operations such as FFT panoramic carrier acquisition, geometric feedforward Doppler compensation, Chirp time-frequency coarse estimation, and symmetrical dual-tone phase offset. The processed observation data is sent to the data processing center 520 through the digital link.

[0078] The data processing center 520 is deployed on a ground-based or space-based server 530. The data processing center can work with the data processing unit 5105 within the receiving base station to perform dynamic distributed computation tasks. The hardware host 5201 serves as the physical carrier, on which the operating system 5202 is deployed, along with the positioning and computation application software 5203. After loading, the software executes corresponding data processing and coordinate computation logic, including observation data reception and phase back-projection normalization, robust CRT deblurring, multi-base station spherical intersection / single-station spatial constraint fusion, and optional Kalman temporal smoothing, ultimately outputting the terminal's accurate geospatial coordinates. System parameter configuration

[0079] The main core physical parameters of the system are configured as follows: Radio frequency carrier frequency: 920 MHz Total available bandwidth: 2 MHz Signal occupied channel physical bandwidth: 1.6 kHz (random access single-channel ultra-narrowband state) Antenna port input signal-to-noise ratio (SNR): 0 - 18 dB Receiving base station platform orbital altitude: 1200 km Typical physical distance between binary stars: 2400 km Satellite phased array antenna: 7 x 7 array (due to temperature drift and inconsistent group delay in multi-channel arrays, noise reduction is required using the symmetric time-division differential algorithm of this system). Receiver base station noise figure: 2 dB, Receiver sensitivity: -139 dBm Terminal local crystal oscillator accuracy: 10 ppm. Ordinary crystal oscillator (producing a maximum inherent initial frequency deviation of ±9.2 kHz at 920 MHz). Terminal transmit power: 24 dBm, without complex temperature control and closed-loop two-way interactive hardware. Preamble signal structure design

[0080] The physical topology of the three-segment burst preamble designed in this embodiment is as follows: (1) Carrier acquisition segment 101: continuous wave single frequency single tone signal, duration 16 ms.

[0081] (2) Time-frequency rough estimation segment 102: 12 Chirp symbols, using a time-domain interleaved UDUDUD... structure, with a single symbol duration of 4 ms and a total duration of 48 ms.

[0082] (3) Dual-tone precision measurement segment 103: Time-division dual-tone continuous wave structure, consisting of 3 sets of cascaded dual-tone time segments, each set containing 2 strictly symmetrical sub-segments, for a total of 6 sub-segments. The duration of a single sub-segment is 16 ms, and a 2 ms protection interval is left between sub-segments.

[0083] Each sub-segment transmits two frequencies simultaneously: one is the center reference component f0 = 0 Hz, which is always located at the baseband relative zero frequency point, and the other is the offset frequency f0 = 0 Hz. i When the terminal transmits each two-tone segment, the baseband processor controls the NCO to force transmission at the start point of each segment according to a pre-set fixed initial phase offset.

[0084] Table 1: Frequency and Number Theory Parameter Configuration of Each Subsegment in the Dual-Tone Precision Measurement Segment

[0085] Figure 1 The physical topology of the preamble frame 100 in the time-frequency two-dimensional space is specifically demonstrated. The signal as a whole exhibits extremely stringent ultra-narrowband characteristics. The first segment transmitted is a 16ms carrier acquisition segment 101, which has a flat spectrum and is used for fast frequency locking at the receiver. This is followed by a time-frequency coarse estimation segment 102, which consists of 12 forward sweep chirp symbols 1021 and reverse sweep chirp symbols 1022, each with a duration of 4ms, cascaded in an UDUD interleaved structure. This symmetrical configuration, through the design of the coincidence of the centroids in the time domain geometry, allows the receiver to perfectly decouple the delay and frequency offset through dual averaging. Finally, there is a dual-tone precision measurement segment 103, which consists of 3 groups (6 in total) of sub-segments. The dual-tone precision measurement segment 103 contains multiple segments of time-division alternating continuous wave ranging signals. Each segment contains a center carrier component located at the center frequency f0, and discrete frequency components located at ±Δf1, ±Δf2 and ±Δf3 respectively. Unambiguous precision measurement is performed using large-scale sparse frequency differences under large-scale virtual synthetic bandwidth. Workflow step-by-step explanation and accuracy estimation

[0086] Step 1: Terminal unidirectional burst blind transmission Without acquiring any prior time-frequency auxiliary information, the terminal transmits the aforementioned three-segment preamble signal to the satellite in a single, unidirectional manner within a total bandwidth of 2 MHz and at random times and frequencies. The minimum nominal input signal-to-noise ratio is 0 dB when the signal arrives at the satellite phased array receiving base station antenna port.

[0087] Step 2: Spatial geometric feedforward compensation The spaceborne receiving base station retrieves the precise orbital position and velocity vector provided by the spaceborne GNSS in real time. Combined with the known predetermined pointing and beamwidth of the received beam (in this example, the beamwidth is 20° and the minimum ground elevation angle is 20°), it directly calculates the Doppler acceleration at the center of each beam as the feedforward input for Doppler acceleration compensation. After feedforward compensation, the residual acceleration within the beam caused by the spatial geometric difference between the edge and center is compressed to a maximum of less than 45 Hz / s.

[0088] Step 3: Base Station FFT Panoramic Blind Acquisition and Carrier Frequency Locking The receiving base station uses a 2.5 MHz instantaneous bandwidth for panoramic data acquisition, covering the randomly occurring areas caused by crystal oscillator drift and the superposition of large dynamic Doppler effects between the satellite and ground. The receiver performs frequency-domain discrete spectral scanning through the carrier acquisition segment, capturing and latching the signal spectral peaks in real time as a coarse estimate of the carrier frequency to establish the initial reference frequency for the entire frame. After signal acquisition and latching, the system configures a narrow window of 500 Hz (total bandwidth 1 kHz) on one side in the carrier acquisition segment to perform high signal-to-noise ratio purification of the carrier.

[0089] Step 4: Time-frequency coarse estimation segment Chirp matched filtering and time delay coarse measurement When switching to the time-frequency coarse estimation segment, the filter window is widened to 900 Hz or 1000 Hz on one side (total bandwidth 1.8 kHz or 2.0 kHz) to completely encapsulate the 1.6 kHz chirp signal and ensure phase integrity while preventing spectral clipping. After completing the geometric feedforward coarse compensation, the receiving base station performs de-chirp processing and frequency domain interpolation on 12 chirp symbols in 6 alternating forward and reverse (6U6D) groups; extracts the complex phase of each matched peak, generates an equally spaced time-domain phase sequence, and then performs quadratic polynomial fitting using the weighted least squares method.

[0090] The fine frequency offset FOA is demodulated using the first-order coefficient c1, and the residual continuous second-order Doppler acceleration is extracted using the second-order coefficient c2, achieving two-stage Doppler closed-loop suppression. Simultaneously, to address the dynamic first-order linear range aliasing generated throughout the total duration, complete descrambling is achieved through geometric feedforward time-varying phase resampling and a deskewing dual averaging operator with a 6U6D symmetric structure, eliminating range-Doppler coupling and calculating the coarse time of arrival (TOA).

[0091] Closed-form derivation and verification of the accuracy of time and distance measurement coarse estimation: (1) Unify known physical parameters: LFM Chirp sweep bandwidth B = 1600 Hz, single symbol duration Ts = 4 ms, total number of symbols N = 12, and total fitting time span T for equally spaced sampling. span =11xT s =0.044 s.

[0092] (2) Peak signal-to-noise ratio of matched single symbol : The product of standard LFM time-width-bandwidth BT s =1600x0,004=6.4, Peak Signal-to-Noise Ratio (SNR) after matching p =SNR·BT s =6.4 (linear value, corresponding to SNR=0 dB).

[0093] (3) Single symbol phase estimation 1 σ Standard deviation : .

[0094] (4) Closed-form derivation of the inverse matrix of a quadratic fitting at N points with equal intervals : The phase of the constant term (corresponding to the baseband delay term) 1 σ standard deviation: LFM baseband delay standard deviation: First-order Doppler frequency shift f d The theoretical lower bound: Second-order Doppler acceleration a d The theoretical lower bound: When the SNR is 6 dB, the calculation process is the same, and the following summary table is obtained: Table 2: Summary of rough performance estimates under different SNRs

[0095] Figure 2This demonstrates the signal feature purification process at the base station physical layer, as well as the processing flow for the time-frequency coarse measurement segment. After the raw signal is input, step 201 performs a panoramic frequency domain scan. Step 202 captures the carrier acquisition segment peak and locks the coarsely estimated carrier frequency. Then, step 203 retrieves the onboard geometric parameters for geometric feedforward Doppler coarse compensation, removing large-head dynamics, enabling step 204 to switch the digital filter to a 50Hz narrow window to intercept most noise. Entering the time-frequency coarse measurement segment processing, step 205 performs matched filtering and de-skewing dual averaging. Step 206 performs quadratic polynomial fitting on the extracted phase sequence, thereby extracting the second-order Doppler acceleration in step 207 and decoupling the output coarse time of arrival (TOA).

[0096] Step 5: Dual-tone precision measurement segment multi-frequency phase measurement and two-level joint fuzzy solution The system performs symbol timing window alignment using the coarse time delay extracted in step 4. With feedforward Doppler acceleration compensation already performed, after entering the dual-tone precision measurement segment, the system adaptively and dynamically narrows the filter window to the second window of 100 Hz (50 Hz on one side) to deeply intercept out-of-band noise and adjacent channel interference, ensuring the accuracy of subsequent phase extraction and absolute ranging.

[0097] In the dual-tone precision measurement segment, the original measured phase of the single-tone signal in each time-division observation segment is extracted. First, reverse subtraction offset compensation is performed using the fixed phase difference data pre-stored on the ground to reunify the phase reference for the entire segment. Then, positive and negative symmetrical dual-tone differential is performed within the group, and secondary differential dual averaging is performed across time segments. Common-mode offset completely eliminates the non-ideal phase error φ introduced by the multi-channel antenna hardware. chan It outputs a high-precision, pollution-free time delay remainder observation sequence.

[0098] Subsequently, the data processing center performed a two-level joint defuzzification on the sequence: like Figure 3 As shown, the spatial distance between the ground terminal 303 deployed on the Earth's surface 302 and the low-orbit satellite 301 in the air is limited by geometric boundaries: the ranging value is minimized when the satellite is directly above the nadir point on the vertical path 305 (D). min = 1200km); the ranging value is maximum when the satellite is on the maximum visible slant range path with the lowest available elevation angle (D). max = 3000 km); the difference between the two constitutes the dynamic range of slant distance variation of 1800 km 306. In the system design phase, an absolute macroscopic search window W with a window size of 2000 km is pre-established. virtSince 2000 km can fully cover and exceed the actual variation range of 1800 km, this window can be directly used as a global fixed boundary constraint for subsequent multi-frequency joint defuzzification, without any additional calculation processing in the actual solution process of the terminal.

[0099] (1) Establishment of the unfuzzy search space (first-level window locking): The system retrieves the first set of frequency differences Δf1=1050Hz and the second set of frequency differences Δf2=1200 Hz, whose greatest common divisor gcd=150 Hz. Based on this, an outer macroscopic unfuzzy virtual solution window is synthesized in number theory. This window is not only valid in number theory, but also perfectly encompasses the 2000 km macroscopic physical-geometric constraint uncertainty boundary determined by the difference between the low-Earth orbit satellite's orbital altitude and the maximum slant distance from the ground at the physical-geometric level. This 2000 km unfuzzy search space is established during system design, providing an initial search window for global robust Chinese Remainder Theorem joint fuzzy resolution, requiring no additional processing during the solution process.

[0100] (2) Robust CRT Solution (Second-Level Subdivision Hit): Within the aforementioned 2000 km macroscopic search space, the system directly utilizes Robust CRT defuzzification for solution, simultaneously substituting multiple sets of differential phase measurements, including the highest frequency difference Δf3 = 750 kHz (corresponding to a 400 m ranging narrow window), into the objective function that minimizes the weighted circular residual sum of squares. A global integrated grid search is performed using the maximum likelihood estimation (MLE) operator. The algorithm skips the intermediate-level step-by-step derivation, directly and accurately mapping and locking the optimal estimate of the precise absolute propagation delay that minimizes the global residual within the large macroscopic window.

[0101] (3) Closed-loop anti-multipath fusing: During the maximum likelihood grid search, the system synchronously monitors the global generalized phase residual. If the current channel is affected by strong multipath interference due to reflections from the surface buildings, causing the calculated generalized phase residual to exceed the monotonicity boundary threshold of 0.25 cycles (i.e., 90°), the system immediately determines that the coherent phase has been contaminated, triggers the dynamic fusing mechanism, closes the current coherent solution, and adaptively downgrades to the pure coarse time delay ranging result fitted by the time-frequency coarse estimation segment in step 4 for survivability fallback.

[0102] First-order Doppler (FOA) precision estimation and theoretical derivation of ranging accuracy: Under normal, unobstructed communication conditions (taking SNR=0 dB, receiving base station noise figure 2 dB, and dual-tone power evenly distributed in each sub-segment resulting in an actual carrier-to-noise ratio C / N0=29 dB-Hz per tone as an example), after 16 ms baseband coherent integration, the output signal-to-noise ratio SNR is... out=794.33 x 0.0016s = 12.71. The single-tone native phase measurement error is... .

[0103] After undergoing a first-order difference analysis of the two tones and a second-order difference analysis of the symmetrical segment (a superposition of the phases of four independent single tones), the final measured standard deviation of the phase noise is: .

[0104] (1) Precision ranging standard deviation (high frequency band physical interval Δf=750 kHz) (2) Precision frequency measurement standard deviation: The total time domain span of the dual-tone precision measurement segment is 106 ms. Symmetrical subtraction yields three independent time difference values, which are then fitted using a three-point linear regression model passing through the origin. Equivalent total time leverage The standard deviation of the frequency measurement after fitting based on long-baseline full absolute phase linear regression is: When the SNR is 6 dB, the calculation process is the same, and the following summary table is obtained: Table 3: Summary of Precision Ranging and Frequency Measurement Performance under Different SNRs

[0105] Figure 2 The process flow for the dual-tone precision measurement segment is also demonstrated. Step 208 calls a fixed phase difference to perform reverse offset compensation and splices the cross-segment reference. Steps 209 and 210 perform positive and negative symmetrical dual-tone difference and dual averaging within the group to offset multi-channel hardware same-mode error and multi-channel error. Step 211, based on the previously fitted parameters, back-projects all phases to the global time domain anchor point to eliminate distance aliasing. Finally, the data is fed into the Robust CRT grid search in step 212, and the generalized phase residual is judged in step 213. If it does not exceed 0.25 cycles, high-precision absolute TOF and FOA are output in step 215; if it exceeds the threshold, step 214 is triggered to shut down the circuit and output the coarse distance measurement result of the time-frequency coarse estimation segment in an emergency.

[0106] Step 6: Solve the localization equation for extracting feature parameters The data processing center obtains the absolute propagation delay (TOF1, TOF2) and arrival frequency, and other characteristic observation parameters, precisely extracted from the preamble processing flow described above. When both base stations are instantaneously visible, adaptive matching and distribution of the solution stream are performed. The first solution branch (single-way distance TOF spatial intersection): When the terminal is in a favorable geometric region on the flank of the axis of the line connecting the two receiving base stations, a system of spatial spherical equations is constructed with the spatial coordinates of the two receiving base stations as the center and their corresponding absolute slant distances (R=c·TOF) as the radius. This system is then combined with the WGS-84 surface ellipsoid and the ground DEM elevation model, and the Gauss-Newton iterative method is used to efficiently and analytically solve for the terminal's two-dimensional geographic coordinates. Within this optimal lateral intersection region, and with a signal-to-noise ratio (SNR) of 6 dB (single-way distance standard deviation σ...), the terminal's two-dimensional geographic coordinates are determined. ref Under the condition of 12.6 m, the surface comprehensive horizontal position error (RMSE) converges to 11.1 m laterally and 18.2 m axially at an elevation angle of 20°.

[0107] The second solution branch (TDOA / FDOA joint time-frequency constraint solution): When the terminal approaches the baseline blind zone of the ground projection line connecting the two receiving base stations and faces geometric degradation due to the approximate tangency of the equidistant circles, the data processing center constructs observations of Time Difference of Arrival (TDOA) and Frequency Difference of Arrival (FDOA). Utilizing the strong positive interactive complementarity of the spatial geometric geometry between the spatial hyperboloid gradient of the time difference of arrival and the spatial iso-Doppler cone gradient generated by high dynamic Doppler, the observation input is substituted into the TDOA / FDOA time-frequency joint Jacobian matrix for joint weighted iterative solution. This forcibly suppresses the error divergence caused by geometric degradation, achieving balanced high-precision position calculation with no blind spots in the entire spatial domain.

[0108] As long as the positioning equation converges iteratively and is physically reliable, the system determines that the current sudden signal positioning is successful. Simultaneously, it delivers the geographic coordinates to the user and provides the equivalent measurement error variance σ calculated in real-time for the current location. 2 k Delivered as metadata in a synchronized package. 5.2 Example 2: Adaptive geometric solution system and vehicle tracking under visible degradation conditions with a single receiving base station

[0109] This embodiment shares the same low-cost IoT terminal and three-segment burst preamble structure as Embodiment 1 (the frequency and number theory parameter configuration of each sub-segment are completely the same as in Table 1). It focuses on explaining how the data processing center adopts an adaptive geometric connection solution system and introduces spatial domain auxiliary constraints to solve the fusion of terminal locations under the extreme degraded topology conditions visible to a single receiving base station. Adaptive geometric solution system

[0110] Figure 4 The adaptive topology flow of the backend positioning equation in the data processing center is demonstrated. After obtaining the frontend parameters in step 401, step 402 intelligently determines the number of valid base stations. If there are three or more satellites, proceed to branch 403, pure absolute time delay rendezvous system, step 404 to perform pairwise analytical solution of the spherical equation system, and step 405 to perform information weighted average to fuse multi-satellite data.

[0111] If there are two units, step 406 determines whether it is close to the connection blind zone: if it is in the good geometry zone, proceed to branch 407 to construct the one-way distance spherical intersection equation system; if the geometry is degraded, proceed to branch 408 to construct the TDOA / FDOA time-frequency joint constraint solution equation system, and use the spatial Doppler cone and hyperboloid gradient orthogonal to break the degradation.

[0112] If it degenerates to single-station visibility, proceed to branch 409. Step 410 assesses the geometric error of the sub-satellite point: if it is in the singular region, step 411 forces the TOF weights to zero, and step 412 performs spatial area intersection calculation with the strip geometric constraint and the AoA projection error surface; if it is in the normal region, step 413 performs joint multidimensional area intersection calculation with the TOF equidistant strip, the AoA ellipse, and the strip geometric constraint.

[0113] After the preliminary geographic coordinates calculated by all branches are summarized in step 414, step 415 checks whether the solution is an empty set or divergent: if so, step 416 rejects the output by multipath circuit breaking; otherwise, step 417 performs Kalman time-series dynamic smoothing, and finally delivers the precise geospatial coordinates to the terminal in step 414.

[0114] The following provides a detailed explanation of vehicle tracking scenarios under single-station visibility conditions. System scenario optimization parameters

[0115] This embodiment shares the same scenario parameters as Embodiment 1, with only the following differences: (1) Optimization of satellite-borne receiver array: The receiving base station adopts a 23 x 23 square uniform planar phased array antenna with an equivalent physical aperture of L = 3.74 m, providing high-resolution spatial measurement.

[0116] (2) Spatial Auxiliary Constraint Injection in Vehicle Tracking Application Layer: Digital prior map constraints: The absolute error of the road centerline in the high-precision electronic road map is less than 1 m (considered as an ideal boundary), and the maximum physical width of the road is 30 m (i.e., the lateral width of the road restricting vehicle travel is 15 m).

[0117] Timing behavior constraint: The terminal burst blind reporting interval is a fixed 60 s. The maximum driving speed of ground vehicles is set to 120 km / h. Based on this, the maximum longitudinal displacement boundary of the vehicle within a single message cycle is recursively derived to be 2000 m. Workflow and spatial constraints are integrated for solution

[0118] Step 1: Phased array two-dimensional angle of arrival (AoA) measurement The receiving base station uses a multi-channel phased array receiving antenna to receive the unidirectional burst ultra-narrowband preamble signal. Using a spatial coherence spatial spectrum estimation algorithm, the simulated value of the two-dimensional angle of arrival (AoA) of the current frame signal relative to the base station is directly calculated using an industry-standard algorithm in situ. (1) At a signal-to-noise ratio of 0dB: nadir point 3σ error 4716m; circumferential error at 20° elevation angle 9694m, radial error 28347m; (2) At a signal-to-noise ratio of 6dB: nadir point 3σ error 3072m; circumferential error at 20° elevation angle 6314m, radial error 18462m; Step 2: Precise Absolute Time-of-Flight Extraction and Adaptive Weight Switching The receiver extracts the carrier phase measurement values ​​of the symmetrical two-tone signal, calls the fixed phase difference data of the corresponding sub-segment for inverse subtraction offset compensation, and splices and projects them onto the same virtual coherent time axis to unify the phase reference of the entire segment. The two-stage joint defuzzification solution (macroscopic unfuzzy search space and Robust CRT solution) is exactly the same as in Example 1, and the precise absolute propagation time delay (TOF) is calculated.

[0119] (1) The boundary of the ranging uncertainty at a signal-to-noise ratio of 0 dB is 3σ ≈ 75.71 m.

[0120] (2) The boundary of the ranging uncertainty is 3σ ≈ 37.95 m at a signal-to-noise ratio of 6 dB.

[0121] Before performing the solution, the data processing center introduces an adaptive weight switching logic for nadir / low elevation angle: real-time determination of the geometric error ΔD of the ground projection radial (nadir point radial direction) of the absolute TOF ranging. 3σ (TOF equidistant band width ΔD) 3σ =3σ R / cos(α), where α is the instantaneous observation elevation angle) is greater than the radial ground error of AoA angle measurement.

[0122] (1) Singular region of the sub-satellite point: If the distance projection error is greater than the angle measurement surface error, it is determined that it is in the tangent singular region of the spherical surface of the sub-satellite point. At this time, the TOF weight in the positioning equation set is forced to zero, and the severely degraded absolute time delay constraint is completely cut off.

[0123] (2) Non-star-ground point regular area: If it is not greater than, maintain full parameter weight and fully integrate the measurement.

[0124] Step 3: Construct application spatial constraints and multidimensional area intersection calculation The data processing center combines the acquired feature parameters with prior knowledge injected from the application layer to construct highly constrained boundaries: (1) Construction of a strip-shaped highly constrained drivable area: Taking the centerline of the road on the high-precision map as the axis, a horizontal absolute physical wall is formed by extending 15 m outward on both the left and right sides; taking the perpendicular foot of the previous frame's historical precise positioning point on the road centerline as the center, a longitudinal motion boundary is formed by extending 2000 m outward along the road direction for each vehicle's maximum temporal displacement. The horizontal and vertical boundaries overlap and close, constructing a strip-shaped geometrically constrained area with a length of 4000 m and a width of 30 m in space.

[0125] (2) Multidimensional space intersection solution: Sub-satellite singularity region: Perform spatial area intersection calculation between the constructed strip-shaped geometric constraint region and the ground projection error circle / ellipse generated by the two-dimensional AoA native measurement.

[0126] Non-nadir point conventional region: Perform joint multidimensional area intersection operation in space on the ground projection equidistant wideband generated by absolute TOF ranging, the two-dimensional AoA native measurement error ellipse, and the above-mentioned strip geometric constraint region.

[0127] The geometric center of the final spatial physical area intersection region is used as the final planar positioning coordinate output, and the intersection range is used as the uncertainty error distribution output.

[0128] Simulation results show that, under conditions of visibility degradation at a single receiving base station, the positioning accuracy can reach 15m laterally along the road after introducing spatial auxiliary constraints. In the longitudinal direction, the positioning accuracy is 37.95 / cos(20°) = 40.38m (SNR = 6dB) at a 20° elevation angle and 2000m at the nadir (determined by the longitudinal motion boundary). The AoA angular measurement error is relatively large and is not a primary constraint in most cases.

Claims

1. A method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning, applied to receiving base stations deployed on low-Earth orbit platforms, extraterrestrial orbiters, and high-dynamic spacecraft; characterized in that, The method includes the following steps: Step 1: The terminal blindly transmits a preamble signal unidirectionally without receiving any downlink auxiliary information from the receiving base station or performing any coordination. The preamble signal sequentially includes a continuous wave carrier acquisition segment, a time-frequency coarse estimation segment with alternating positive and negative linear sweep symbols and coincident time-domain geometric centroids, and a dual-tone fine estimation segment composed of multiple sets of time-division dual-tones. The physical bandwidth of the linear sweep of the time-frequency coarse estimation segment is limited to no more than 10 kHz. Step 2: The receiving base station performs frequency domain scanning on the received preamble signal to capture the carrier frequency of the signal, locks the actual arrival carrier frequency of the carrier capture segment, establishes the baseband digital downconversion center frequency, distributes the captured signal to the corresponding baseband processing channel to obtain the baseband signal to be processed, and then closes the subsequent processing window for the baseband signal to be processed into a preset processing window. Step 3: The receiving base station uses its known spatial location and spatial geometric parameters to perform geometric feedforward Doppler coarse compensation on the full frame signal. Based on this, phase extraction is performed on each linear frequency sweep symbol of the time-frequency coarse estimation segment. The residual first-order Doppler frequency offset and continuous second-order Doppler acceleration are estimated by using a quadratic polynomial fitting closed loop. The time delay and frequency offset coupling are eliminated by using a linear frequency modulated signal deskewing dual averaging operator. The coarse arrival time is calculated and used as the global time domain anchor point. Step 4: The receiving base station constructs a continuous phase time-domain projection function based on the Doppler parameters calculated from the time-frequency coarse estimation segment, and uniformly back-projects and normalizes the phase observation values ​​of each time-division dual-tone sub-segment of the dual-tone fine measurement segment to the global time-domain anchor point; it calls the arbitrary cascaded defuzzification algorithm to perform defuzzification search and calculates the absolute propagation delay from the terminal to the receiving base station; Step 5: Based on the propagation delay, and combined with one or more observations from the arrival frequency demodulated by the first-order Doppler frequency offset or the spatial auxiliary constraints, solve for the geospatial coordinates of the output terminal.

2. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that: The physical bandwidth of the linear sweep frequency of the time-frequency coarse estimation segment and the bandwidth occupied by each discrete single-tone component are further uniformly limited to 500 Hz to 3 kHz.

3. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 2, characterized in that: The physical bandwidth of the linear sweep frequency of the time-frequency coarse estimation segment and the bandwidth occupied by each discrete single-tone component are further uniformly limited to 1 kHz to 2 kHz.

4. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning as described in claim 1, characterized in that, The frequency domain scanning specifically involves performing a Fast Fourier Transform (FFT) on the received burst signal and then performing a frequency domain discrete spectral line scan within the preset total bandwidth of the processing window.

5. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The receiving base station retrieves its own spatial position and velocity vector in real time, and calculates the theoretical Doppler acceleration at the center of the beam by combining it with the direction of the current receiving beam. This acceleration is then used as a feedforward input to the Doppler compensator to perform Doppler acceleration compensation on the baseband signal to be processed.

6. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 5, characterized in that, A switchable digital filtering architecture is configured to purify the signals of the carrier acquisition segment and the dual-tone precision measurement segment, and the filtering window of the switchable digital filtering architecture is dynamically switched according to whether Doppler compensation is completed; when Doppler compensation is not completed, a wide first window is used to accommodate the frequency shift signal. After completing Doppler compensation, switch to the narrow second window.

7. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 6, characterized in that, The total processing bandwidth of the narrower second window is 10 Hz to 200 Hz.

8. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The phase extraction process specifically includes performing matched filtering and frequency domain interpolation on each linear sweep symbol to extract an equally spaced time-domain phase sequence.

9. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The method of using a quadratic polynomial fitting closed-loop estimation to determine the residual first-order Doppler frequency offset and the continuous second-order Doppler acceleration specifically includes: performing a quadratic polynomial fitting on the time-domain phase sequence to obtain polynomial coefficients; calculating the residual first-order Doppler frequency offset based on the first-order coefficients of the quadratic polynomial; and calculating the continuous second-order Doppler acceleration based on the second-order coefficients of the quadratic polynomial.

10. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The elimination of time delay and frequency offset coupling by using a linear frequency modulated signal deskewing dual averaging operator specifically includes: performing dual averaging based on the frequency domain peak point positions of the alternating positive and negative linear sweep frequency configuration to offset the linear distance aliasing error.

11. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The phase reference unification mechanism of the dual-tone precision measurement segment includes: allowing the numerically controlled oscillator inside the terminal to be turned off or interrupted and reset during signal transmission switching; controlling the numerically controlled oscillator to transmit according to a preset fixed initial phase offset at the transmission start point of each sub-segment of the dual-tone precision measurement segment; the fixed initial phase offset is pre-stored by the data processing center, and during reception and calculation, the original measurement phase of each single-tone signal is extracted and corrected by subtracting the corresponding fixed initial phase offset through the inverse offset model, so as to eliminate phase jumps across time periods and unify the phase of discrete sub-segments to the same virtual coherent time axis.

12. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, Based on the global time-domain anchor point, the processing interval of the dual-tone signal is determined. Within the current processing window, the dual-tone precision measurement segment contains multiple time-division dual-tone structures. Each group consists of two time-domain strictly symmetrical sub-segments, and all sub-segments contain the same center reference frequency and a discrete offset frequency on one side. The receiving base station performs channel error self-counterweighting processing on each sub-segment, specifically including: when the system uses a multi-channel antenna array, the receiving base station extracts the one-sided differential phase of the positive and negative offset frequency points relative to the center reference frequency within the group; converts the one-sided differential phase to the time delay domain and performs a second differential dual averaging across the time segments; when the system uses a single-channel antenna, it directly calculates the differential phase between the one-sided baseband offset frequency point and the center reference frequency, and outputs the time delay remainder sequence.

13. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The cascaded unfuzzing algorithm is specifically configured as the robust Chinese Remainder Theorem algorithm. When performing the two-level joint unfuzzing search, the mechanism for establishing the first-level macroscopic unfuzzy search space is as follows: configure the first and second sets of frequency differences of the dual-tone precision measurement segments, and use their greatest common divisor to synthesize the outer macroscopic unfuzzy virtual solution window to envelop the relative dynamic range of the receiving base station and the terminal; wherein, the scale of the outer macroscopic unfuzzy virtual solution window is determined by the difference between the maximum and minimum distances between the receiving end and the terminal.

14. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 13, characterized in that, The robust Chinese Remainder Theorem algorithm performs a second-level grid unfuzzy search within the outer macroscopic unfuzzy virtual solution window: it simultaneously introduces multiple sets of differential phase measurements, including the highest frequency difference, to construct a weighted objective function that minimizes the sum of squared circular residuals with the inverse of the measurement noise variance as the weight. It then uses maximum likelihood estimation to search for the optimal precise time delay solution that minimizes this objective function, and directly maps and solves the propagation time delay determined by the highest offset frequency difference in one go.

15. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 13, characterized in that, The robust Chinese residual theorem defuzzification algorithm introduces a dynamic circuit breaker mechanism: during the defuzzification search process, a closed-loop circuit breaker phase residual threshold is preset; if the generalized phase residual calculated in the progressive defuzzification exceeds the threshold, it is determined that the current observation is subject to strong multipath interference or phase distortion, and the system immediately triggers the circuit breaker mechanism, automatically switching to the pure time delay coarse ranging result obtained by fitting the time-frequency coarse estimation segment to ensure survivability.

16. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 15, characterized in that, Based on the phase monotonicity boundary of the robust Chinese Remainder Theorem algorithm, the preset closed-loop fuse phase residual threshold is specifically limited to 0.25 cycles.

17. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The geospatial coordinate solution method specifically includes: when there are two visible receiving base stations, using a one-way propagation delay (TOF) positioning system or a time difference of arrival / Doppler frequency offset (TDOA / FDOA) joint positioning system to establish a corresponding positioning solution equation set for coordinate calculation.

18. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, When the number of visible receiving base stations is greater than or equal to three, the geospatial coordinate solution method includes: using a pure propagation delay spatial reintersection system to solve the simultaneous spatial multi-faceted intersecting spherical equations analytically; or, pairing all visible receiving base stations in pairs or dividing them independently to solve them separately, obtaining multiple candidate location solutions and performing weighted normalized averaging.

19. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, When the number of visible receiving base stations is a single unit, the geospatial coordinate calculation method includes: performing a fusion intersection operation on the absolute propagation delay of a single base station, combined with the two-dimensional angle of arrival (AoA), digital prior map, and temporal behavioral feature constraints, to obtain the terminal's geographical location.

20. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that, The spatial auxiliary constraints include one or more of the following: radio spatial measurement constraints, antenna physical beam constraints, digital prior map constraints, and temporal behavior characteristic constraints. The radio spatial measurement constraints include at least one of the following: two-dimensional angle of arrival (AoA), spatial direction of arrival (DoA), coherent multi-channel spatial spectrum estimation distribution, and relative receiving phase difference of the array's multi-channels, measured by the receiving base station antenna array. The antenna physical beam constraints include at least one of the following: gain footprint boundary formed by the projection of the current receiving beam of the receiving base station onto the space or ground surface where the terminal is located, the elevation angle limit of the visible horizon, and geometric blind zone constraints caused by terrain physical occlusion. The digital prior map constraints include at least one of the following: two-dimensional or three-dimensional digital electronic maps, high-precision road network maps, ground height constraints provided by digital elevation models (DEMs), and three-dimensional urban building cluster spatial white model constraints. The temporal behavior characteristic constraints include at least one of the following: motion trajectory envelope constraints derived from the terminal's historical location point trajectory, legal speed limits, terminal physical speed limits, and message time intervals, and preset geofencing constraints.

21. The method for processing unidirectional ultra-narrowband preamble codes for radio frequency positioning according to claim 1, characterized in that: The unidirectional transmission is an asynchronous uplink unidirectional transmission. The terminal can initiate transmission at any time and complete it independently without relying on the radio frequency receiving circuit. The center frequency of the terminal can be arbitrarily selected within the available bandwidth of the system, and the entire frame signal spectrum, including the dual-tone precision measurement segment, falls within the available bandwidth of the system.

22. A unidirectional ultra-narrowband preamble data processing center for radio frequency positioning, characterized in that, The data processing center is configured to coordinate with at least one receiving base station to execute the method as described in any one of claims 1 to 21. The data processing center includes: a spatial constraint injection module, used to acquire or generate spatial auxiliary constraints corresponding to the terminal, and introduce the spatial auxiliary constraints as time delay or geometric boundary into the system; a distributed collaborative defuzzification module, which, in coordination with the receiving base station, converts the observation data captured and feedforward compensated by the receiving base station into propagation delay or time difference of arrival based on a cascaded defuzzification algorithm and utilizing a cascaded constraint verification and phase residual fuse mechanism; and a multi-mode spatial geometric positioning module, used to call the geospatial coordinate calculation steps as described in claim 1 to calculate the location data of the terminal when the observation data of the effective receiving base station receiving the same signal meets the conditions of a single receiving base station or multiple receiving base stations.

23. The unidirectional ultra-narrowband preamble data processing center for radio frequency positioning according to claim 22, characterized in that: The computing architecture of the processing center is configurable and includes any one of the following: a purely local processing architecture executed independently by the receiving base station; a purely centralized processing architecture executed independently by the data processing center; and a distributed collaborative processing architecture in which computing tasks are dynamically allocated between the receiving base station and the data processing center according to a configurable distribution strategy and in combination with computing resources.

24. A unidirectional transmission ultra-narrowband preamble processing system for radio frequency positioning, characterized in that, A method for implementing any one of claims 1 to 23 comprises: a terminal transmitter, including a numerically controlled oscillator, a radio frequency transmission link, a processor, and a memory, wherein the memory stores terminal-side control software; the processor invokes the terminal-side control software to control the numerically controlled oscillator and the radio frequency transmission link to perform the asynchronous uplink unidirectional transmission action as described in claim 21; a receiving base station, including an antenna array, a radio frequency receiving link, a baseband processor, and a memory, wherein the memory stores base station receiving processing software; the baseband processor invokes the base station receiving processing software to perform baseband processing and Doppler compensation on the signal received by the radio frequency receiving link, and outputs the processed signal data to a data processing center; a data processing center, including server hardware, an operating system deployed on the server hardware, and positioning calculation application software; the server hardware runs the operating system and the positioning calculation application software, receives data sent by the receiving base station, and executes the data processing and geospatial coordinate calculation steps corresponding to claims 1 to 23; wherein the data processing center is deployed locally at the receiving base station, in independent ground or space-based server hardware, or in a distributed collaborative network of both.

Images