Test method and system for vehicle-mounted antenna

Abstract

The invention belongs to the technical field of communication engineering, and particularly relates to a test method and system for a vehicle-mounted antenna, and the system comprises an environment feature sensing module, a dynamic excitation generation module, a vehicle-mounted antenna interface unit, a multi-channel radio frequency collection module, a state synchronization controller and a closed-loop verification engine. Through the synergistic effect of the environment characteristic sensing module and the dynamic excitation generation module, electromagnetic-physical disturbance in a real road is converted into a programmable test variable instead of a passively suppressed noise source; the physical information neural network model in the closed-loop verification engine can be used to complete performance prediction and deviation determination in a single test period, and self-adaptive convergence of the test process is realized.

Description

A test method and system for vehicle-mounted antennas Technical Field

[0001] This invention belongs to the field of communication engineering technology, specifically a testing method and system for vehicle-mounted antennas. Background Technology

[0002] With the rapid development of intelligent connected vehicle technology, the vehicle communication system, as the core carrier for information interaction between vehicles, roads, and the cloud, directly affects the functional safety and user experience of the entire vehicle. The vehicle antenna, as the physical interface for transmitting and receiving radio frequency signals, determines the stability of the communication link and the quality of data transmission through its radiation characteristics and impedance matching. Therefore, high-precision, high-efficiency, and reproducible testing is a crucial step in vehicle R&D and production verification. Currently, the industry mainstream adopts microwave anechoic chamber far-field / near-field scanning testing, combined with vector network analyzers and automated turntables for offline antenna parameter calibration. Although the laboratory accuracy is high, the testing conditions are idealized and cannot truly reflect the combined effects of complex electromagnetic environments, vehicle body reflections, multipath effects, and dynamic attitudes during actual vehicle operation. In existing technologies, some solutions deploy fixed probe arrays or mobile robotic arms carrying probes in the vehicle state to simulate approximate far-field in-situ testing within a limited space, improving the realism of the scenario to some extent, but still having structural limitations. Such systems simplify antenna performance evaluation to static parameter acquisition, measuring unidirectional signals at preset discrete angles / positions, and then interpolating and fitting to extrapolate omnidirectional performance. However, the actual operation of vehicle-mounted antennas is a dynamic coupling system. During vehicle movement, the vehicle's attitude and surrounding obstacles will modulate the radiation field distribution in real time. Traditional methods lack the ability to model the closed-loop interaction mechanism of test stimulus-environment response-antenna feedback, and cannot capture time-varying coupling effects. Moreover, the existing process treats environmental factors as interference and ignores them, and does not include them in the test variable system for quantitative correlation, resulting in a significant deviation between the test results and the actual vehicle communication performance.

[0003] Current automotive antenna testing technologies suffer from the following problems: an irreconcilable conflict between high-fidelity environment simulation and testing standardization and efficiency. Pursuing environmental fidelity requires the introduction of a large number of dynamic variables, leading to difficulties in testing standardization, poor repeatability, and time consumption. Emphasizing efficiency and repeatability requires sacrificing environmental complexity and reverting to idealized static models, which cannot characterize the antenna performance degradation mechanism under real-world conditions. This contradiction stems from the architectural flaws of existing testing paradigms: treating antennas as isolated devices rather than functional units embedded in dynamic physical-electromagnetic coupling systems, and is not due to insufficient equipment precision or algorithm optimization. Existing systems lack causal modeling of environmental disturbances and antenna responses, making it impossible to support high-order verification requirements such as fault tracing and performance boundary detection, which severely restricts the reliable deployment of automotive communication systems in safety-critical scenarios such as high-level autonomous driving.

[0004] Therefore, the present invention provides a testing method and system for vehicle-mounted antennas. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve its technical problem is as follows: A test system for vehicle-mounted antennas, comprising an environmental feature perception module, a dynamic excitation generation module, a vehicle-mounted antenna interface unit, a multi-channel RF acquisition module, a state synchronization controller, and a closed-loop verification engine. Each module is interconnected via a hard real-time communication bus and driven by a unified clock source, ensuring that the entire system maintains event alignment and state consistency at the microsecond-level time granularity.

[0007] Preferably, the environmental feature perception module is deployed on the boundary of the test site and on the mobile vehicle, and includes a distributed electromagnetic probe array, an inertial measurement unit, a lidar scanning unit, and a meteorological sensing unit. The electromagnetic probe array consists of multiple broadband dipole antennas arranged along a three-dimensional spatial grid to capture the background electromagnetic field intensity, polarization direction, and multipath delay distribution within the test area in real time. The inertial measurement unit is fixed to the chassis of the vehicle under test and is used to continuously output three-axis attitude data of the vehicle's pitch angle, roll angle, and yaw angle. The lidar scanning unit performs point cloud modeling of the surrounding obstacle contours at a preset scanning frequency to generate a dynamic occlusion map. The meteorological sensing unit synchronously collects temperature, humidity, air pressure, and precipitation status information. All perception data is timestamped and transmitted to the state synchronization controller via industrial Ethernet.

[0008] Preferably, the dynamic excitation generation module comprises a multi-channel signal generator, a reconfigurable phased array transmitting unit, and a path simulator. The multi-channel signal generator generates a radio frequency excitation signal with specified carrier frequency, modulation format, power level, and angle of arrival parameters according to instructions issued by the state synchronization controller. The reconfigurable phased array transmitting unit consists of several independently controllable radiating units, whose feed phase and amplitude are adjusted in real time by a field-programmable gate array to synthesize an incident wavefront with arbitrary directionality. The path simulator, based on the obstacle model provided by the lidar and the multipath parameters fed back by the electromagnetic probe, reconstructs the propagation path loss, reflection coefficient, and scattering components online using a ray tracing algorithm, and superimposes them onto the original excitation signal to form a composite excitation waveform with scene semantics.

[0009] Preferably, the vehicle-mounted antenna interface unit is integrated into the original antenna mounting location of the vehicle under test. It contains an impedance matching network, a duplexer, and a radio frequency switch matrix. On the one hand, the unit leads the transmitted signal from the vehicle-mounted communication terminal to the multi-channel radio frequency acquisition module. On the other hand, it injects the test signal output by the dynamic excitation generation module into the antenna port, realizing bidirectional isolation and switching of the transmit and receive links. The radio frequency switch matrix is ​​controlled by a state synchronization controller, and completes the seamless switching between the transmit path, receive path, and calibration path according to a preset timing sequence in each test cycle, ensuring that the measurement process does not interfere with the normal working logic of the system under test.

[0010] Preferably, the multi-channel RF acquisition module consists of a high dynamic range receiver, a digital down-conversion unit, and a high-speed analog-to-digital converter. This module synchronously receives reflected signals, transmitted signals, and environmental coupling signals from the vehicle-mounted antenna interface unit. Under the synchronous drive of a local oscillator, it down-converts each channel's signal to baseband, generates a time-aligned complex baseband data stream after analog-to-digital conversion, and strictly adheres to the Nyquist criterion for data sampling. The data is temporarily stored in a dual-port RAM buffer and then written in batches to the shared memory pool of the central processing unit by the direct memory access controller.

[0011] Preferably, the state synchronization controller, as the core scheduling unit of the system, adopts a hard real-time operating system kernel and has a built-in multi-task state machine. This state machine defines six main states: initialization, environment modeling, stimulus configuration, data acquisition, feedback evaluation, and adaptive adjustment. It responds to external events through an interrupt triggering mechanism. Before the start of each test cycle, the state synchronization controller first reads the latest environment snapshot from the environment feature perception module and, combined with a preset test case library, generates the stimulus parameter set required for the current cycle. Then, it issues configuration instructions to the dynamic stimulus generation module and simultaneously starts the sampling clock of the multi-channel RF acquisition module. After data acquisition is completed, it immediately packages the raw data and environmental context and sends them to the closed-loop verification engine.

[0012] Preferably, the closed-loop verification engine runs within a dedicated secure computing unit, which possesses a trusted execution environment to ensure that sensitive test logic and data are not interfered with by the general-purpose operating system. The closed-loop verification engine first extracts channel state information from the received radio frequency data to obtain the S-parameter matrix, radiation pattern slice, and effective omnidirectional radiated power of the antenna under test under the current environmental excitation. Then, based on the multi-dimensional context provided by the environmental feature perception module (including vehicle posture, obstacle distribution, meteorological conditions, and background electromagnetic spectrum), it constructs an environment-antenna coupling feature vector. Furthermore, through a pre-trained physical information neural network model, this feature vector is mapped to the expected performance index space, and a deviation analysis is performed between it and the measured index. If the deviation exceeds a preset tolerance threshold, an adaptive adjustment process is triggered, generating a new set of excitation parameters and transmitting it back to the state synchronization controller to initiate the next iteration cycle.

[0013] Preferably, the physical information neural network model integrates first-principles electromagnetic simulation data and real vehicle road test data during the training phase. Its loss function explicitly embeds the constraint terms of Maxwell's equations to ensure that the model output conforms to the basic laws of electromagnetic fields. During the inference phase, the model only relies on the current environmental feature vector as input and does not require historical data backtracking, thereby supporting real-time performance prediction and deviation determination within a single test cycle.

[0014] Preferably, the entire testing process is executed according to the following steps: Step 1, the vehicle under test is driven into the test site, the physical connection between the vehicle antenna interface unit and the multi-channel RF acquisition module is completed, and all sensing and control units are started; Step 2, the state synchronization controller performs system self-test and time synchronization calibration to ensure that all modules share the same time reference; Step 3, the environmental feature perception module continuously collects the initial state of the environment and generates the first frame environmental feature vector; Step 4, the closed-loop verification engine calls the test case template based on the vector to generate the initial stimulus parameter set; Step 5, the state synchronization controller configures the dynamic stimulus generation module and triggers a complete transmit and receive test cycle; Step 6, the multi-channel RF acquisition module synchronously records the antenna port response data and uploads it to the closed-loop verification engine along with the environmental context; Step 7, the closed-loop verification engine performs performance deviation analysis. If it is determined that the current test has not converged, the stimulus parameter set is updated and the process returns to Step 5; if it is determined that the test has converged, the current test case is terminated and a structured test report is output; Step 8, Steps 3 to 7 are repeated to traverse all preset test scenarios until full operating condition coverage is completed.

[0015] Preferably, the test report includes the environmental feature vector, excitation parameter set, measured antenna performance index, predicted performance index and deviation quantification result corresponding to each test cycle, and is stored in an anti-tampering storage medium in the form of encrypted logs to support subsequent fault tracing and performance boundary analysis.

[0016] Preferably, the system supports a multi-vehicle concurrent testing mode. In this mode, multiple vehicle-mounted antenna interface units establish independent control channels with their respective state synchronization controllers via power line carrier communication, while the environmental feature perception module allocates a dedicated perception sub-region to each vehicle through a spatial segmentation multiplexing mechanism, ensuring electromagnetic isolation and data independence between each test instance.

[0017] Preferably, to improve testing efficiency, the system incorporates a scene compression engine. This engine, based on principal component analysis, reduces the dimensionality of environment-performance correlation patterns in historical test data, extracts key influencing factor combinations, and generates a representative set of test scenarios accordingly, avoiding redundant verification of operating conditions.

[0018] Preferably, all RF links use shielded twisted-pair cables and coaxial connectors, and the grounding system conforms to the equipotential bonding standard; the communication bus uses a time-triggered protocol to ensure the determinism and predictability of message transmission; the secure computing unit has a built-in hardware random number generator and a national cryptographic algorithm coprocessor for testing data integrity verification and access control.

[0019] The beneficial effects of this invention are as follows: The testing method and system for vehicle-mounted antennas described in this invention, through the synergistic effect of the environmental feature perception module and the dynamic excitation generation module, transforms electromagnetic-physical disturbances in real roads into programmable test variables, rather than passively suppressed noise sources; it utilizes the physical information neural network model in the closed-loop verification engine to complete performance prediction and deviation determination within a single test cycle, achieving adaptive convergence of the testing process; it realizes in-situ characterization of the radiation characteristics of vehicle-mounted antennas under dynamic vehicle posture and complex obstacle distribution, avoiding performance misjudgments caused by neglecting coupling effects in traditional offline calibration methods; it ensures a hard real-time synchronization mechanism between the multi-channel RF acquisition module and the state synchronization controller, enabling millisecond-level environmental changes to be accurately correlated to the antenna transient response; through the seamless switching of the vehicle-mounted antenna interface unit, it ensures the transparency of the test process to the operating logic of the communication system under test; and through the closed-loop verification logic within the trusted execution environment, it prevents test data from being tampered with or leaked at the general operating system level. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 is a schematic diagram of the overall structure of the test system for vehicle-mounted antennas according to the present invention; Figure 2 is a schematic diagram of the test method for vehicle-mounted antennas according to the present invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] As shown in Figure 1, a test system for vehicle-mounted antennas according to an embodiment of the present invention includes an environmental feature perception module, a dynamic excitation generation module, a vehicle-mounted antenna interface unit, a multi-channel radio frequency acquisition module, a state synchronization controller, and a closed-loop verification engine. The above modules are interconnected through a hard real-time communication bus and driven by a unified clock source to ensure that the entire system maintains event alignment and state consistency at the microsecond time granularity. The entire system is deployed in a dedicated electromagnetic compatibility test site, where the ground is covered with a conductive copper mesh and connected to an equipotential grounding system. The surrounding walls are equipped with absorbing material to suppress external electromagnetic interference and simulate free space propagation conditions.

[0024] The environmental feature perception module is deployed at the boundary of the test site and on the mobile vehicle, and includes a distributed electromagnetic probe array, an inertial measurement unit, a lidar scanning unit, and a meteorological sensing unit. The electromagnetic probe array consists of 48 broadband dipole antennas, arranged along a three-dimensional spatial grid on eight vertical pillars in the test area. Six probes are installed on each pillar, spaced 1.5 meters apart, forming an 8×6×1 spatial sampling structure. Each probe is equipped with a low-noise amplifier and an analog-to-digital converter front-end to capture the background electromagnetic field strength, polarization direction, and multipath delay distribution within the test area in real time. The inertial measurement unit uses industrial-grade MEMS devices, fixed at the center of the vehicle chassis, and outputs a 1kHz frequency, continuously providing the vehicle's pitch angle. The system provides three-axis attitude data for roll and yaw angles with an angular resolution better than 0.01°. The lidar scanning unit is a 16-line mechanical rotating LiDAR, installed at the center of the test site top, with a horizontal field of view of 360° and a vertical field of view of ±15°. The point cloud density is no less than 30 points / square meter at a distance of 10 meters, used for real-time modeling of surrounding obstacle contours and generating dynamic occlusion maps. The meteorological sensing unit integrates temperature and humidity sensors, a barometer, and a rain gauge, with a sampling period of 1 second, synchronously collecting environmental meteorological parameters. All sensing data is timestamped in the local FPGA, with the time reference derived from the Precision Time Protocol (PTP) master clock, and then transmitted to the state synchronization controller via gigabit industrial Ethernet.

[0025] In some implementations, the dynamic excitation generation module consists of a multi-channel signal generator, a reconfigurable phased array transmitter unit, and a path simulator. The multi-channel signal generator is built on the PXIe platform and includes eight independent RF channels, each supporting a frequency range of 9kHz to 6GHz, an output power range of -120dBm to +10dBm, a modulation bandwidth of up to 200MHz, and supporting multiple modulation formats such as AM / FM / PM, QPSK, 16QAM, and OFDM. The multi-channel signal generator generates signals with a specified carrier frequency according to instructions issued by the state synchronization controller. Modulation format Power level and Angle of Arrival (AoA) parameter θ,ϕ The radio frequency excitation signal; the reconfigurable phased array transmitter unit consists of 64 independently controllable Vivaldi antenna elements arranged in an 8×8 planar array. Each element is connected to a digital phase shifter and a digitally controlled attenuator at the rear end. The phase adjustment step is 1°, and the feed phase and amplitude are controlled in real time by a field-programmable gate array (FPGA) at a refresh rate of 100kHz to synthesize an incident wavefront with arbitrary directionality; the path simulator runs on an embedded GPU server, based on the obstacle point cloud model provided by the lidar and the multipath parameters (including the number of paths) fed back by the electromagnetic probe. Delay of each path Reflection coefficient and scattering intensity The propagation channel characteristics are reconstructed online using an improved ray tracing algorithm. This algorithm considers Fresnel reflection, diffraction, and scattering effects from rough surfaces, and the calculation formula is as follows: ;in, Indicates the composite channel impulse response. Let be the complex gain of the i-th path. and Here, λ represents the gain of the transmitting and receiving antennas, respectively, and λ is the wavelength. The path length; the path simulator will... The original baseband signal s(t) is convolved to generate a composite excitation waveform with scene semantics. After being up-converted, the signal is sent to the phased array transmitting unit.

[0026] In some implementations, the vehicle-mounted antenna interface unit is integrated into the original antenna mounting location of the vehicle under test. Internally, it includes an impedance matching network, a duplexer, and an RF switch matrix. This unit is encapsulated in an IP67-rated metal housing with an 8-layer internal PCB stack-up, and the RF trace impedance is strictly controlled to 50Ω ± 1Ω. The impedance matching network consists of adjustable inductors and varactor diodes, supporting automatic tuning within the 700MHz to 5.9GHz frequency band, with a VSWR optimized to below 1.5:1. The duplexer is a cavity filter structure with isolation better than 50dB, used to separate the uplink and downlink frequency bands. The RF switch matrix is ​​composed of GaAs PIN diode switches with a switching time of less than 100ns, an insertion loss of less than 0.3dB, and an isolation of more than 60dB. This unit, on the one hand, leads the transmit signal from the vehicle communication terminal (such as the 5GC-V2X module) to the multi-channel RF acquisition module through the SMA interface, and on the other hand, injects the test signal output by the dynamic excitation generation module into the antenna port, realizing bidirectional isolation and switching of the transmit and receive links. The RF switch matrix is ​​controlled by a state synchronization controller, which completes the seamless switching between the transmit path, receive path, and calibration path according to a preset timing sequence in each test cycle. For example, in the 5GNR test scenario, the switching timing is configured as follows: 0-2ms for the calibration path (connected to the built-in 50Ω load), 2-10ms for the receive path (connected to the acquisition module), and 10-12ms for the transmit path (connected to the vehicle terminal), ensuring that the measurement process does not interfere with the normal operating logic of the system under test.

[0027] The multi-channel RF acquisition module consists of a high dynamic range receiver, a digital down-conversion unit, and a high-speed analog-to-digital converter. Based on an 8-channel synchronous receiver architecture, each channel includes a low-noise amplifier (noise figure <2dB), a mixer, an anti-aliasing filter, and a 16-bit ADC. The receiver input frequency range is 100MHz to 6GHz, with a maximum input power of +10dBm, a third-order cross-interval adjustment point (IP3) better than +25dBm, and a dynamic range of 100dB. The digital down-conversion unit is implemented using a Xilinx Zynq UltraScale+MPSoC, and the local oscillator (LO) is driven by the same 10MHz temperature-compensated crystal oscillator, ensuring that the phase consistency error of each channel is less than 1°. Under the synchronous drive of the LO, the signals from each channel are down-converted to baseband, filtered by an anti-aliasing filter, and then converted to digital by the ADC at a sampling rate of 500MS / s to generate a time-aligned complex baseband data stream. Where k=1,8 represents the channel index; data sampling for all channels strictly follows the Nyquist criterion, with sampling clock jitter less than 50fs. The converted data is temporarily stored in a dual-port RAM cache with a cache depth of 64MB / channel, and is written in batches to the central processing unit's shared memory pool by the PCIeGen3x8 interface direct memory access (DMA) controller at a bandwidth of 8GB / s.

[0028] In some implementations, the state synchronization controller, as the core scheduling unit of the system, adopts a heterogeneous architecture based on a Xilinx Kintex Ultrascale FPGA and an ARM Cortex-A53 dual-core processor, running the Xenomai hard real-time operating system kernel, with task scheduling jitter below 1μs. This controller has a built-in multi-task state machine, defining six main states: initialization, environment modeling, stimulus configuration, data acquisition, feedback evaluation, and adaptive adjustment. It responds to external events through a hardware interrupt triggering mechanism. In some implementations, before the start of each test cycle, the state synchronization controller first reads the latest environmental snapshot (containing a 128-dimensional feature vector) from the environmental feature perception module, and combines it with preset test parameters. The use case library (stored in NVMe SSD, 2TB capacity) generates the set of excitation parameters required for the current cycle through table lookup and interpolation algorithms. This set of parameters includes 32 configuration items such as carrier frequency, modulation type, transmit power, AoA angle, and multipath model parameters. The controller sends configuration commands to the dynamic excitation generation module through the AXI4-Lite bus and synchronously starts the sampling clock of the multi-channel RF acquisition module through the LVDS differential signal, with timing alignment error controlled within ±20ns. After data acquisition is completed (typical cycle is 12ms), the controller immediately packages the raw RF data (approximately 512MB) and the environmental context (approximately 4KB) into a JSON format message and sends it to the closed-loop verification engine via Gigabit Ethernet.

[0029] The closed-loop verification engine runs within a dedicated secure computing unit, which builds a Trusted Execution Environment (TEE) based on Intel SGX (SoftwareGuardExtensions) technology to ensure that sensitive test logic and data are not interfered with by the general-purpose operating system. In some implementations, the closed-loop verification engine first extracts Channel State Information (CSI) from the received radio frequency data. Specifically, this involves extracting complex baseband data for each channel. Perform a Fast Fourier Transform (FFT) to obtain the frequency domain response. Then calculate the elements of the S-parameter matrix. ,in Given the excitation signal spectrum; simultaneously, by inverting the radiation pattern slice using a beamforming algorithm, the effective isotropic radiated power (EIRP) is calculated using the following formula: ,in, For input power, The antenna gain direction function is obtained by least-squares fitting of measured far-field data; in some implementations, the engine is based on multi-dimensional context (including three-axis attitude angles) provided by the environmental feature perception module. Obstacle point cloud density Rainfall intensity R (mm / h), background electromagnetic noise power spectral density (etc.), construct a 128-dimensional environment-antenna coupling feature vector. ; Through a pre-trained Physical Information Neural Network (PINN) model The feature vector is mapped to the expected performance metric space (such as EIRP, cross-polarization discrimination rate, multipath delay spread, etc.), and the predicted value is output. In some implementations, the model integrates first-principles electromagnetic simulation data (generated using CSTStudioSuite, covering 2000 different vehicle postures and obstacle configurations) with real-vehicle road test data (from OTA tests of 100 mass-produced vehicles in urban, highway, and tunnel scenarios) during the training phase. Its loss function explicitly embeds Maxwell's equations constraint terms, in the following form: The first term represents the data fitting loss, the second term represents the physical constraint regularization term, λ=0.1 is the weighting coefficient, and E, B, and D are the electromagnetic field components of the model's implicit output. During the inference phase, the model relies solely on the current environment feature vector as input, requiring no historical data backtracking, and a single inference iteration takes less than 5ms. The engine will then predict the values... Deviation analysis was performed between the measured index y and the normalized mean square error (NMSE) was calculated: If NMSE exceeds the preset tolerance threshold (e.g., 0.05), the current test is determined to have not converged, triggering the adaptive adjustment process to generate a new set of excitation parameters (e.g., increasing the number of multipath components or changing the AoA angle), and sending it back to the state synchronization controller to start the next iteration cycle; if NMSE ≤ 0.05, the test is determined to have converged, and the current test case is terminated.

[0030] The entire testing process is executed according to the following steps: Step 1, the vehicle under test is driven into the test site, and the physical connection between the vehicle-mounted antenna interface unit and the multi-channel RF acquisition module is completed (using N-type coaxial cable, length ≤ 2 meters, shielding effectiveness > 100dB), and all sensing and control units are started; Step 2, the state synchronization controller performs system self-test (including clock synchronization, link connectivity, and RF calibration) and time synchronization calibration, aligning with the site master clock via the PTP protocol to ensure that all modules share the same time reference, with a synchronization error < 100ns; Step 3, the environmental feature sensing module continuously acquires the initial environmental state and generates the first frame environmental feature vector; Step 4, the closed-loop verification engine calls the test application based on this vector. Example template (e.g., urban canyon scenario, rainfall intensity 5mm / h, vehicle speed 40km / h) generates initial excitation parameter set; Step 5, the state synchronization controller configures the dynamic excitation generation module and triggers a complete transmit / receive test cycle; Step 6, the multi-channel RF acquisition module synchronously records antenna port response data and uploads it along with the environmental context to the closed-loop verification engine; Step 7, the closed-loop verification engine performs performance deviation analysis. If it determines that the current test has not converged, it updates the excitation parameter set and returns to Step 5; if it determines that it has converged, it terminates the current test case and outputs a structured test report; Step 8, repeat steps 3 to 7, traversing all preset test scenarios (128 in total) until full operating condition coverage is completed.

[0031] In some implementations, the test report includes the environmental feature vector, excitation parameter set, measured antenna performance indicators (EIRP, S11, 3dB beamwidth of the radiation pattern, etc.), predicted performance indicators, and deviation quantification results (NMSE, maximum absolute error, etc.) for each test cycle. These are stored in an tamper-proof storage medium (HSM hardware security module based on the national cryptographic SM4 algorithm) in the form of AES-256 encrypted logs, supporting subsequent fault tracing and performance boundary analysis.

[0032] The system supports multi-vehicle concurrent testing; in this mode, up to four vehicles can be tested simultaneously. Each vehicle-mounted antenna interface unit establishes an independent control channel with its respective status synchronization controller via power line carrier communication (PLC, frequency band 2-30MHz), with a communication rate of 1Mbps and a bit error rate of [missing information]. The environmental feature perception module uses a spatial segmentation and reuse mechanism to assign a dedicated perception sub-region to each vehicle: the lidar point cloud is spatially clustered according to the vehicle ID, and the electromagnetic probe array is divided into perception sectors according to the nearest neighbor principle to ensure electromagnetic isolation (isolation degree > 40dB) and data independence between each test instance.

[0033] In some implementations, to improve testing efficiency, the system incorporates a scenario compression engine. This engine, based on principal component analysis (PCA), reduces the dimensionality of environment-performance correlation patterns in historical test data. Specifically, it performs covariance matrix decomposition on the 128-dimensional environmental features of 100,000 historical samples, extracting the top 20 principal components (cumulative variance contribution rate > 95%) to form a combination of key influencing factors. K-means clustering is used to generate 32 representative test scenario sets, covering 99% of the performance variation space, avoiding redundant verification of conditions, and reducing testing time by 68%.

[0034] It should be noted that all RF links use RG-214 / U shielded twisted-pair coaxial cables and N-type connectors, with 360° termination of the shielding layer. The grounding system complies with the IEC61000-5-2 equipotential bonding standard, with a grounding resistance of <0.1Ω. The communication bus uses the Time Triggered Protocol (TTP / C), with a message transmission cycle deterministic error of <1μs, ensuring the predictability of control commands and timing events. The secure computing unit has a built-in TRNG (True Random Number Generator, compliant with NISTSP800-90B standard) and a national cryptographic SM2 / SM3 / SM4 algorithm coprocessor for testing data integrity verification (HMAC-SM3) and access control (SM2-based digital certificate authentication).

[0035] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A testing system for vehicle-mounted antennas, characterized in that, include: The system comprises an environmental feature perception module, a dynamic excitation generation module, a vehicle-mounted antenna interface unit, a multi-channel RF acquisition module, a state synchronization controller, and a closed-loop verification engine. The environmental feature perception module, dynamic excitation generation module, vehicle-mounted antenna interface unit, and multi-channel RF acquisition module are each connected to the state synchronization controller, which is communicatively connected to the closed-loop verification engine. All modules are interconnected via a hard real-time communication bus and driven by a unified clock source to maintain event alignment and state consistency. The environmental feature perception module collects background electromagnetic field distribution, three-axis attitude of the vehicle under test, contours of surrounding obstacles, and meteorological parameters within the test area, and generates a timestamped environmental feature vector. The dynamic excitation generation module generates an RF excitation signal with a specified carrier frequency, modulation format, power level, and angle of arrival based on the excitation parameter set issued by the state synchronization controller, and reconstructs the propagation path characteristics online based on the obstacle model and multipath parameters to form a composite excitation waveform. The vehicle-mounted antenna interface unit, dynamic excitation generation module, vehicle-mounted antenna interface unit, and multi-channel RF acquisition module are connected to the state synchronization controller, which is communicatively connected to the closed-loop verification engine. The interface unit is integrated into the original antenna mounting location of the vehicle under test, used for seamless switching between the transmit path, receive path, and calibration path, and to achieve bidirectional isolation between test signal injection and vehicle terminal signal output; the multi-channel RF acquisition module is used to synchronously receive reflected signals, transmitted signals, and environmental coupling signals from the vehicle antenna interface unit, and generate time-aligned complex baseband data streams after down-conversion and analog-to-digital conversion; the state synchronization controller is used to read environmental feature vectors, generate excitation parameter sets, configure dynamic excitation generation modules, trigger RF acquisition, and package the acquired data with the environmental context to send to the closed-loop verification engine; the closed-loop verification engine runs in a trusted execution environment, used to construct coupling feature vectors based on the environmental feature vectors, map them to the expected performance index space through a pre-trained physical information neural network model, and perform deviation analysis with the measured performance indexes. If the deviation exceeds a preset tolerance threshold, an adaptive adjustment process is triggered and the updated excitation parameter set is sent back to the state synchronization controller.

2. The testing system for vehicle-mounted antennas according to claim 1, characterized in that, The environmental feature perception module includes: a distributed electromagnetic probe array, consisting of multiple broadband dipole antennas arranged along a three-dimensional spatial grid, used to capture background electromagnetic field strength, polarization direction, and multipath delay distribution; an inertial measurement unit, fixed to the chassis of the vehicle under test, used to output the vehicle's pitch angle, roll angle, and yaw angle; a lidar scanning unit, used to perform point cloud modeling of surrounding obstacles to generate a dynamic occlusion map; and a meteorological sensing unit, used to collect temperature, humidity, air pressure, and precipitation status information; all sensing data are timestamped and transmitted to the state synchronization controller via industrial Ethernet.

3. The testing system for vehicle-mounted antennas according to claim 1, characterized in that, The dynamic excitation generation module includes: a multi-channel signal generator for generating radio frequency excitation signals with specified carrier frequency, modulation format, power level, and angle of arrival; a reconfigurable phased array transmitting unit, composed of multiple independently controllable radiating units, whose feed phase and amplitude are adjusted in real time by a field-programmable gate array to synthesize incident wavefronts with arbitrary directionality; and a path simulator for reconstructing propagation path loss, reflection coefficient, and scattering components online based on obstacle point cloud models and multipath parameters using ray tracing algorithms, and superimposing them onto the original excitation signal to form a composite excitation waveform.

4. The testing system for vehicle-mounted antennas according to claim 1, characterized in that, The vehicle-mounted antenna interface unit internally includes an impedance matching network, a duplexer, and an RF switch matrix; the impedance matching network supports automatic tuning; the duplexer is used to separate the uplink and downlink frequency bands; the RF switch matrix is ​​composed of diode switches and is controlled by the state synchronization controller to complete the path switching according to a preset timing sequence.

5. A testing system for vehicle-mounted antennas according to claim 1, characterized in that, The multi-channel RF acquisition module includes a high dynamic range receiver, a digital down-conversion unit, and a high-speed analog-to-digital converter; each channel's local oscillator is driven by the same temperature-compensated crystal oscillator; the converted complex baseband data stream is cached through a dual-port RAM and written in batches to the shared memory pool of the central processing unit by a direct memory access controller.

6. The testing system for vehicle-mounted antennas according to claim 1, characterized in that, The state synchronization controller reads an environmental snapshot from the environmental feature perception module, generates an excitation parameter set containing carrier frequency, modulation type, transmit power, angle of arrival, and multipath model parameters, and synchronously starts the sampling clock of the multi-channel radio frequency acquisition module through differential signals.

7. A testing system for vehicle-mounted antennas according to claim 1, characterized in that, The closed-loop verification engine runs in a trusted execution environment; the physical information neural network model integrates first-principles electromagnetic simulation data and real vehicle road test data during the training phase, and its loss function explicitly embeds the constraints of Maxwell's equations; during the inference phase, the model only relies on the current coupled feature vector as input.

8. A testing system for vehicle-mounted antennas according to claim 1, characterized in that, Supports multi-vehicle concurrent testing mode: Multiple vehicle antenna interface units establish independent control channels with their respective state synchronization controllers through power line carrier communication; The environmental feature perception module assigns a dedicated perception sub-region to each vehicle through a spatial segmentation and reuse mechanism.

9. A testing method for vehicle-mounted antennas, applicable to the testing system for vehicle-mounted antennas as described in any one of claims 1-8, characterized in that, Includes the following steps: S1: Complete the physical connection between the vehicle under test and the test system and start each functional module; S2: Perform system self-test and clock synchronization calibration to ensure consistent time base across the entire system; S3: The environmental feature perception module collects the initial environmental state and generates the first frame environmental feature vector; S4: The closed-loop verification engine calls the test case template based on the vector to generate the initial stimulus parameter set; S5: The state synchronization controller configures the dynamic stimulus generation module and triggers a complete transmit and receive test cycle. S6: The multi-channel RF acquisition module synchronously records the antenna port response data and uploads it to the closed-loop verification engine; S7: The closed-loop verification engine performs performance deviation analysis. If it fails to converge, it updates the stimulus parameter set and returns to step S5. If it has converged, it terminates the current test case. S8: Repeat S3 to S7, traversing all preset test scenarios to complete full-condition coverage.

10. A testing method for vehicle-mounted antennas according to claim 9, characterized in that, The test report includes the environmental feature vector, excitation parameter set, measured antenna performance indicators, predicted performance indicators, and deviation quantification results for each test cycle.

Images