UWB-based vehicle-mounted millimeter wave radar installation calibration system

Abstract

The invention discloses a UWB-based vehicle-mounted millimeter wave radar installation calibration system, and relates to the technical field of vehicle-mounted electronics. Through cooperative work of an angle deviation calibration module and a translation deviation correction module, the system firstly compares a UWB echo feature vector set Vu with a millimeter wave echo feature vector set Vm, and calculates an installation angle deviation Px; comparing the angle value with a preset angle threshold value Pth; and when the installation angle deviation Px exceeds a threshold value, the translation deviation value d of the millimeter wave radar is calculated based on the phase difference between the UWB receiving arrays. The two-stage correction strategy of first angle and second translation can comprehensively cover two kinds of errors of angle deviation and position deviation occurring in the installation of the millimeter wave radar, and compared with the mode that only single deviation can be corrected in the prior art, the error compensation capability of the system is remarkably enhanced, and the high consistency of the radar detection direction and space positioning is ensured.

Description

Technical Field

[0001] This invention relates to the field of automotive electronics technology, specifically to a UWB-based automotive millimeter-wave radar installation and calibration system. Background Technology

[0002] With the rapid development of intelligent transportation technology and autonomous driving systems, the role of onboard sensing devices in vehicle safety and environmental perception is becoming increasingly prominent. Millimeter-wave radar, due to its advantages such as all-weather operation and strong resistance to obstruction, has become one of the important sensors for onboard environmental perception. However, in practical applications, millimeter-wave radar requires strict accuracy in its installation position and angle; otherwise, detection accuracy will decrease. Therefore, sensor calibration technology has gradually become an important component of onboard sensing systems. Against this technical background, this invention proposes an onboard millimeter-wave radar installation calibration system based on the synergistic effect of ultra-wideband (UWB) and millimeter-wave radar, to solve the target detection error problem caused by installation angle or position deviations in complex environments.

[0003] Currently, millimeter-wave radars typically correct their installation angle and position through factory calibration or self-calibration algorithms based on open road environments. However, in confined spaces such as narrow tunnels and underground parking garages, millimeter-wave echo signals are easily affected by multipath interference from tunnel walls and surrounding structures, leading to deviations in echo signal strength and azimuth estimation. Existing calibration methods based on single millimeter-wave characteristics often fail to accurately distinguish between environmental reflection interference and true target features, resulting in unstable calibration processes and excessively biased results. Especially during vehicle operation, the environment changes frequently, making traditional calibration methods poorly adaptable to complex environments.

[0004] Millimeter-wave radars typically correct their installation angle and position through factory calibration or self-calibration algorithms based on open road environments. However, in confined spaces such as narrow tunnels and underground parking garages, millimeter-wave echo signals are easily affected by multipath interference from tunnel walls and surrounding structures, leading to deviations in echo signal strength and azimuth estimation. Existing calibration methods based on single millimeter-wave characteristics often fail to accurately distinguish between environmental reflection interference and true target features, resulting in unstable calibration processes and excessively biased results. Especially during vehicle operation, the environment changes frequently, making traditional calibration methods poorly adaptable to complex environments. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a UWB-based vehicle-mounted millimeter-wave radar installation and calibration system, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: including an echo joint acquisition module, an angle deviation calibration module, a translation deviation correction module, and a calibration control module; The echo joint acquisition module synchronously acquires multi-source echo parameters inside the tunnel wall through the vehicle-mounted UWB transmitter, UWB receiver array, and millimeter-wave radar equipment, and transmits the multi-source echo parameters to the vehicle-mounted central control system to form UWB echo feature vector set Vu and millimeter-wave echo feature vector set Vm. The angle deviation calibration module calculates the installation angle deviation △Px by comparing the UWB echo feature vector set with the millimeter wave echo feature vector set, and performs a preliminary comparison and evaluation with the preset angle threshold Pth. Based on the preliminary comparison and evaluation results, it triggers translation correction. The translation deviation correction module calculates the translation deviation Δd of the millimeter-wave radar based on the phase difference between the UWB receiving arrays after triggering translation correction. The calibration control module combines the installation angle deviation △Px and the translation deviation △d to form a comprehensive calibration matrix Mfinal. If the correction result corresponding to the comprehensive calibration matrix Mfinal meets the preset threshold condition, the comprehensive calibration matrix is ​​output; otherwise, re-acquisition and secondary optimization are triggered.

[0007] Preferably, the echo joint acquisition module includes a UWB echo acquisition unit, a millimeter-wave echo acquisition unit, and a vector generation unit; The UWB echo acquisition unit transmits UWB signals through a UWB transmitter installed on the vehicle, and the reflected signals are received by a UWB receiving array. The reflected signals are then feature-extracted in the UWB receiving array to obtain UWB echo parameters. The UWB echo parameters include the echo intensity parameter A of the i-th UWB reflection path. u，i The echo delay parameter T of the i-th UWB reflection path u，i The phase parameter Xw of the i-th UWB reflection path u，i .

[0008] Preferably, the millimeter-wave echo acquisition unit acquires millimeter-wave echo signals with the same propagation path as the UWB signal through a millimeter-wave radar device installed on the vehicle, and performs signal processing on the echo signals to obtain millimeter-wave echo parameters; The millimeter-wave echo parameters include the echo intensity parameter A of the k-th path of the millimeter wave. m,k The azimuth parameter O of the k-th path of the millimeter wave m,k And the path identifier parameter L of the k-th path of the millimeter wave. m,k ; The millimeter-wave echo parameters and UWB echo parameters are combined to generate multi-source echo parameters.

[0009] Preferably, the vector generation unit encapsulates the multi-source echo parameters in the form of data frames and transmits them to the vehicle central control system via the vehicle bus, wherein the data frames contain timestamp information; After receiving the multi-source echo parameters, the vehicle central control system performs dimensionless processing, eliminating the dimensions of all parameters in the multi-source echo parameters and normalizing them to the same range. Then, it synchronizes the UWB echo parameters and millimeter-wave echo parameters based on the timestamp information of the dimensionless multi-source echo parameters, and generates the UWB echo feature vector set Vu and the millimeter-wave echo feature vector set Vm.

[0010] Preferably, the angle deviation calibration module includes an angle deviation analysis unit and a deviation angle evaluation unit; The installation angle deviation analysis unit calculates the installation angle deviation △Px based on the UWB echo feature vector set Vu and the millimeter wave echo feature vector set Vm. The installation angle deviation △Px is the difference between the mean values ​​of the millimeter wave echo feature vector set Vm and the UWB echo feature vector set Vu, and is obtained by taking the arctangent value of the ratio and comparing it with the mean value of the UWB echo feature vector set Vu.

[0011] Preferably, after obtaining the installation angle deviation ΔPx, the deviation angle evaluation unit performs a preliminary comparison evaluation between the installation angle deviation ΔPx and a preset angle threshold Pth to determine the degree of deviation of the installation angle, and triggers translation correction based on the preliminary comparison evaluation results. The specific evaluation content is as follows: When the installation angle deviation ΔPx is less than or equal to the angle threshold Pth, the angle deviation is considered to be within an acceptable range, and no calibration is required. When the installation angle deviation △Px is greater than the angle threshold Pth, it is determined that the installation angle is obviously wrong, and translation correction is triggered.

[0012] Preferably, the translation deviation correction module includes a translation deviation analysis unit; The translation deviation analysis unit, after triggering translation correction, sets the phase parameter Xw of the i-th UWB reflection path. u，i The phase parameter Xw of the j-th UWB reflection path with different reflection paths u，j The phase difference ΔXw is obtained by performing difference calculation, and the translation deviation Δd is calculated and output based on the phase difference ΔXw.

[0013] Preferably, the calibration control module includes a comprehensive calibration unit, a calibration evaluation unit, and a secondary optimization unit; The integrated calibration unit, based on the initially set reference angle calibration matrix Mo, introduces the installation angle deviation ΔPx and the translation deviation Δd respectively, and performs linear adjustment through the convergence control coefficient Q to obtain the integrated calibration matrix Mfinal.

[0014] Preferably, after generating the comprehensive calibration matrix Mfinal, the calibration evaluation unit calibrates the installation of the UWB vehicle-mounted millimeter-wave radar based on the comprehensive calibration matrix Mfinal. After calibration, the comprehensive calibration matrix Mfinal is applied to the multi-source echo data of the current UWB vehicle-mounted millimeter-wave radar, and the installation angle deviation ΔPx and translation deviation Δd are recalculated to obtain the corrected detection result set Rfinal. Based on the detection result set Rfinal and the ideal detection result Rideal, residual calculation is performed to obtain the residual value. A preset residual threshold Eth is then compared with the residual value for evaluation to determine the effectiveness of the correction. The specific evaluation content is as follows: When the residual value is less than or equal to the residual threshold Eth, the correction is deemed valid, the calibration is completed, and the comprehensive calibration matrix Mfinal is output. When the residual value is greater than the residual threshold Eth, the correction is deemed invalid, and the process proceeds to the second optimization.

[0015] Preferably, the secondary optimization unit enters secondary optimization after determining that the correction is invalid. The secondary optimization involves re-acquiring multi-source echo parameters to form a new UWB echo feature vector set Vu and a millimeter-wave echo feature vector set Vm. Based on the new UWB echo feature vector set Vu and millimeter-wave echo feature vector set Vm, the installation angle deviation result and translation deviation result are recalculated. The recalculated installation angle deviation ΔPx and translation deviation Δd are then fused to form an updated comprehensive calibration matrix Mfinal. During the fusion process, an iterative convergence mechanism is used to iteratively correct the updated calibration matrix multiple times until the corrected residual value is less than or equal to the residual threshold Eth condition.

[0016] This invention provides a UWB-based vehicle-mounted millimeter-wave radar installation and calibration system. It offers the following advantages: (1) This system uses a combined echo acquisition module to achieve synchronous acquisition with the vehicle-mounted UWB transmitter, UWB receiver array, and millimeter-wave radar equipment. It then generates a UWB echo feature vector set Vu and a millimeter-wave echo feature vector set Vm in the vehicle's central control system, thereby establishing consistency of multi-source echo parameters at the data level. This solution effectively solves the problem of detection result offset caused by multipath reflection from tunnel walls or underground parking garages in narrow, constrained spaces. By leveraging the complementary stability of UWB echoes and the high resolution of millimeter-wave echoes, the system can significantly improve the target detection accuracy and installation calibration precision of vehicles in complex environments.

[0017] (2) Through the coordinated operation of the angle deviation calibration module and the translation deviation correction module, this invention first compares the UWB echo feature vector set Vu with the millimeter-wave echo feature vector set Vm to calculate the installation angle deviation ΔPx, and compares it with the preset angle threshold Pth. When the installation angle deviation ΔPx exceeds the threshold, the translation deviation Δd of the millimeter-wave radar is calculated based on the phase difference between the UWB receiving arrays. This two-level correction strategy of "angle first, then translation" can comprehensively cover the two types of errors that occur during the installation of millimeter-wave radar: angle deviation and position deviation. Compared with the existing technology that can only correct a single deviation, this invention significantly enhances the error compensation capability of the system and ensures a high degree of consistency between the radar detection direction and spatial positioning.

[0018] (3) Through three stages in the calibration control module—generating the comprehensive calibration matrix Mfinal, calibration evaluation, and secondary optimization—this invention not only achieves linear fusion correction of the installation angle deviation ΔPx and translation deviation Δd, but also dynamically evaluates the correction effect through the residual threshold Eth. When the comprehensive calibration matrix Mfinal fails to meet the preset conditions, the system triggers re-acquisition and secondary optimization, using an iterative convergence mechanism to further reduce the residual until the detection results stabilize. This adaptive and secondary optimization mechanism overcomes the limitations of the existing "one-time calibration" technology, endowing the system with continuous self-learning and stability optimization capabilities under long-term operation, thereby ensuring the long-term reliability and accuracy of the vehicle millimeter-wave radar in multiple scenarios. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a vehicle-mounted millimeter-wave radar installation and calibration system module based on UWB according to the present invention.

[0020] Figure 2 This is a comparison image of echoes in a tunnel environment. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1 Please see Figure 1 and Figure 2This invention provides a vehicle-mounted millimeter-wave radar installation and calibration system based on UWB. To achieve the above objectives, this invention is implemented through the following technical solutions: including an echo joint acquisition module, an angle deviation calibration module, a translation deviation correction module, and a calibration control module; The echo joint acquisition module synchronously acquires multi-source echo parameters inside the tunnel wall through the vehicle-mounted UWB transmitter, UWB receiver array, and millimeter-wave radar equipment, and transmits the multi-source echo parameters to the vehicle-mounted central control system to form UWB echo feature vector set Vu and millimeter-wave echo feature vector set Vm; The angle deviation calibration module calculates the installation angle deviation △Px by comparing the UWB echo feature vector set with the millimeter wave echo feature vector set, and performs a preliminary comparison and evaluation with the preset angle threshold Pth. Based on the preliminary comparison and evaluation results, it triggers translation correction. The translation deviation correction module calculates the translation deviation Δd of the millimeter-wave radar based on the phase difference between the UWB receiving arrays after triggering translation correction. The calibration control module combines the installation angle deviation △Px and the translation deviation △d to form a comprehensive calibration matrix Mfinal. If the correction result corresponding to the comprehensive calibration matrix Mfinal meets the preset threshold condition, the comprehensive calibration matrix is ​​output; otherwise, re-acquisition and secondary optimization are triggered.

[0023] In this embodiment, the system's echo joint acquisition module synchronously acquires reflected echoes within constrained spaces such as tunnel walls using a vehicle-mounted UWB transmitter and UWB receiver array, while simultaneously acquiring echo signals along the same path using millimeter-wave radar. This setup aims to leverage the stronger stability of UWB signals in multipath environments, allowing them to serve as a reference for millimeter-wave echoes. For example, in tunnel scenarios, millimeter-wave echoes can be affected by strong wall reflections, leading to false targets. By introducing the UWB echo feature vector set Vu, the main path signal can be stably identified, avoiding angle misjudgments caused by environmental interference. The angle deviation calibration module compares the UWB echo feature vector set Vu with the millimeter-wave echo feature vector set Vm, calculates the installation angle deviation ΔPx, and evaluates it against a preset threshold Pth. The physical significance of this implementation is that if the radar installation is tilted, the target reflection direction and the actual detection direction will be inconsistent, similar to a deflection of the vehicle's detection range, resulting in misjudgments of obstacle positions. By setting a threshold Pth, the calibration matrix can be updated directly when the deviation is small, and further corrections can be triggered when the deviation is too large, avoiding situations where "minor issues are over-corrected" or "major issues are ignored." After the deviation exceeds the threshold, the translation deviation correction module uses the phase difference between UWB receiving points... The translational deviation Δd of the millimeter-wave radar is calculated. The core physical meaning here is that when the radar position shifts, the path length of the target echo changes, directly causing an overall shift in the ranging result. By converting the phase difference into a path difference, this translational error can be precisely quantified, thus correcting the range dimension of millimeter-wave detection. For example, if the radar's installation position is offset by a few centimeters, it will directly cause a fixed deviation in the detection distance of obstacles ahead, and this module can effectively eliminate this error. Finally, the calibration control module comprehensively corrects the angle deviation ΔPx and the translational deviation Δd, forming a comprehensive calibration matrix Mfinal, and judges the validity of the calibration result through threshold conditions. If valid, the correction is completed; if invalid, re-acquisition and secondary optimization are triggered. This setup aims to ensure the system has adaptability and continuous convergence, enabling dynamic correction under complex operating conditions. Its true physical meaning is to map the dual errors of angle and position caused by installation into a unified matrix calibration model, equivalent to establishing a "virtual standard installation position" for the millimeter-wave radar, ensuring that its detection results are always consistent with the real environment. This implementation method solves the problem of unstable detection caused by installation deviation and environmental interference of millimeter-wave radar in complex environments such as tunnels by combining data acquisition, angle evaluation, translation correction and integrated control. It achieves the beneficial effects of improving calibration accuracy, enhancing target positioning reliability and improving vehicle safety in complex spaces.

[0024] Example 2 Please see Figure 1 Specifically: the echo joint acquisition module includes a UWB echo acquisition unit, a millimeter-wave echo acquisition unit, and a vector generation unit; The UWB echo acquisition unit transmits UWB signals through a UWB transmitter installed on the vehicle, and the reflected signals are received by a UWB receiving array. The UWB receiving array extracts features from the reflected signals to obtain UWB echo parameters. UWB echo parameters include the echo intensity parameter A of the i-th UWB reflection path. u，i The echo delay parameter T of the i-th UWB reflection path u，i The phase parameter Xw of the i-th UWB reflection path u，i ; The UWB transmitter is positioned at the centerline of the front of the vehicle and includes a UWB signal transmitter; the UWB receiver array is positioned on both sides of the millimeter-wave radar equipment and includes UWB receiver array antennas. The echo intensity parameter A of the i-th UWB reflection path u，i Envelope detection and extraction of signals received by a UWB receiver array; The echo delay parameter T of the i-th UWB reflection pathu，i By performing a correlation operation between the received reflected signal and the locally stored transmitted template signal, the delay corresponding to the peak value of the correlation function is the propagation delay. The phase parameter Xw of the i-th UWB reflection path u，i The in-phase component I and quadrature component Q are obtained by down-converting the received signal, and the phase angle is obtained by the arctangent function.

[0025] The millimeter-wave echo acquisition unit acquires millimeter-wave echo signals with the same propagation path as UWB signals through a millimeter-wave radar device installed on the vehicle, and performs signal processing on the echo signals to obtain millimeter-wave echo parameters. Millimeter-wave echo parameters include the echo intensity parameter A of the k-th path of the millimeter wave. m,k The azimuth parameter O of the k-th path of the millimeter wave m,k And the path identifier parameter L of the k-th path of the millimeter wave. m,k ; The millimeter-wave echo parameters and UWB echo parameters are combined to generate multi-source echo parameters; The millimeter-wave radar equipment is installed at a standard mounting position at the front of the vehicle to collect millimeter-wave echo signals with the same propagation path as UWB signals. The echo intensity parameter A of the k-th path of the millimeter wave m,k Obtained by the peak value of the FFT spectrum from the front end of the millimeter-wave radar equipment; The azimuth parameter O of the k-th path of the millimeter wave m,k Beamforming via MIMO antenna arrays; The path identifier parameter L of the k-th path of millimeter wave m,k The path number is marked by the time delay gate and angle index of the echo signal; In this context, the letters u and m represent the signal source UWB and millimeter wave, respectively, and the letters i and k represent the path number.

[0026] The vector generation unit encapsulates multi-source echo parameters in the form of data frames and transmits them to the vehicle central control system via the vehicle bus. The data frames contain timestamp information to ensure that the UWB echo parameters and millimeter-wave echo parameters correspond in the time domain. After receiving the multi-source echo parameters, the vehicle central control system performs dimensionless processing, eliminating the dimensions of all parameters in the multi-echo parameters and normalizing them to the same range. Then, it synchronizes the UWB echo parameters and millimeter-wave echo parameters based on the timestamp information of the dimensionless multi-echo parameters and generates the UWB echo feature vector set Vu and the millimeter-wave echo feature vector set Vm. The specific form of the UWB echo feature vector set Vu is: Where e represents an exponential function, n represents the number of reflection paths in the tunnel wall, and j represents the imaginary unit, satisfying j 2 =-1. In electromagnetic waves and signal processing, signals are often represented in complex form (amplitude + phase), which allows for the simultaneous representation of signal strength and phase changes. Therefore, in the formula... This means that a phase is For a rotating signal, 2π represents the angular frequency conversion factor, and the phase angle needs to be expressed in radians. Converting the frequency to radians requires multiplying by 2π; therefore, 2πfuT u,i In reality, it represents the phase rotation angle corresponding to the time delay, and its value is dimensionless. fu represents the carrier frequency of UWB. To facilitate signal processing and calculation, a center frequency is usually chosen to represent the carrier attribute of the signal. This center frequency is the so-called UWB carrier frequency fu. During hardware configuration, the operating frequency band is specified, such as the 3.1-10.6GHz range. Once the operating frequency band is selected, its center frequency can be used as the UWB carrier frequency fu. For example, if operating at 6.5–7.0GHz, its center frequency is approximately 6.75GHz. This is the origin of the UWB carrier frequency fu, and its value is dimensionless. At the receiving end, the UWB signal includes amplitude (energy level), delay (propagation time), and phase (carrier information). A single amplitude or time delay cannot fully characterize a path; therefore, the effect of time delay on phase is embedded in a complex exponential form, with dimensions: the echo intensity parameter A of the i-th UWB reflection path. u，i The echo delay parameter T of the i-th UWB reflection path u，i All are dimensionless parameters after dimensionless processing. The pure phase factor is essentially a rotation on the complex plane and has no dimension; therefore, the UWB echo eigenvector set Vu is dimensionless. Physics Source: Electromagnetic wave propagation follows the Fourier transform principle and is modeled using the Channel Impulse Response (CIR). Each term in the CIR represents the amplitude and phase delay of a propagation path. This formula abstracts the CIR into a "set of echo eigenvectors," directly expressing a single path as a set. This vector not only preserves the amplitude but also contains phase information caused by path propagation, which can be used for angle and position calibration. The specific form of the millimeter-wave echo feature vector set Vm is: Where max represents the maximum value operation, cos represents the cosine function, and p represents the number of resolvable multipath reflection paths in the millimeter-wave received signal; Physics source: The echo modeling of millimeter-wave radar typically uses beamforming and array pattern output power of the power angle spectrum; the intensity and angle of each path are combined, and the main reflection path is obtained by taking the maximum value; Dimensionality Consistency Explanation: Echo intensity parameter A of the k-th path of millimeter waves m,k The azimuth parameter O of the k-th path of the millimeter wave m,k All of these are dimensionless parameters after dimensionless processing; therefore, the millimeter-wave echo eigenvector set Vm represents dimensionless numerical values. Derivation process: The beamforming output formula ignores weights and interference terms, considers only the principal component of a single path, and uses cosθ. m，k The intensity is adjusted for different azimuth angles, and the strongest path is extracted using max(...) to avoid noise.

[0027] In this embodiment, the system acquires echo intensity, delay, and phase parameters for different paths through a UWB echo acquisition unit via a transmitter and receiver array. This design aims to provide high-resolution path information in complex multipath environments, such as tunnels where multiple reflection paths overlap, making a single intensity parameter insufficient to accurately characterize path differences. Introducing phase and delay parameters effectively avoids target aliasing, resulting in higher path discrimination. The millimeter-wave echo acquisition unit shares the propagation path with the UWB signal and extracts echo intensity, azimuth, and path identification parameters. The physical significance of this deployment lies in the fact that while millimeter-wave radar has advantages in angle resolution, it is prone to strong reflection interference under multipath conditions; for example, tunnel walls can cause strong lateral echo interference in target localization. By binding the millimeter-wave path characteristics with the UWB path characteristics, angle deviations caused by path mismatches can be avoided, improving the reliability of angle information. The vector generation unit packages the multi-source echo parameters into a unified data frame and transmits it to the central control system via the vehicle bus. The purpose of adding timestamps is to ensure the physical correspondence between UWB and millimeter-wave signals at the same moment. For example, if a vehicle is traveling at a certain speed and is not synchronized, the same target may be identified as two different results, leading to positioning drift. Timestamps ensure synchronization, followed by dimensionless processing to eliminate dimensional differences. For instance, amplitude, delay, and phase values ​​may not be in the same dimension range; direct calculation would lead to distortion. Unifying them to dimensionless allows for vector comparison, ensuring consistency in the calculated physical meaning. Ultimately, the UWB echo feature vector set Vu represents the combined effect of amplitude, delay, and phase of a single path in complex exponential form. Its physical meaning is to establish a complete description of a path in the complex plane, avoiding deviations caused by missing phase or delay. The millimeter-wave echo feature vector set Vm extracts main path features through power angle spectrum, ensuring stable angle detection results free from interference by weak paths. Combining these two methods allows for simultaneous consideration of path resolution and angle accuracy in constrained scenarios such as tunnels and underground parking garages, avoiding misjudgments caused by single signal processing.

[0028] Example 3 Please see Figure 1 Specifically: the angle deviation calibration module includes an installation angle deviation analysis unit and a deviation angle evaluation unit; The installation angle deviation analysis unit calculates the installation angle deviation △Px based on the UWB echo feature vector set Vu and the millimeter wave echo feature vector set Vm. The installation angle deviation △Px is the difference between the mean values ​​of the millimeter wave echo feature vector set Vm and the UWB echo feature vector set Vu, and is calculated by taking the arctangent value of the ratio and then the difference between them. The significance of the difference lies in representing the "deviation" between the millimeter wave primary reflection and the overall average path characteristics of UWB. If the installation is normal, the two should be close. If there is an installation deviation, this deviation will increase significantly. The significance of the ratio lies in using the mean of UWB as a normalized benchmark, avoiding misjudgment due to the overall difference in echo intensity in different scenarios, and ensuring that the result only reflects the "relative deviation" rather than the "absolute signal difference". The significance of arctangent lies in the fact that angle deviation is essentially a "combined angle problem"; in trigonometric functions, the most direct relationship between the deviation ratio and the actual deviation angle is the arctangent, so the dimensionless ratio is mapped to an actual angle using the arctangent, which is the installation angle deviation △Px; Physical meaning: The millimeter wave echo feature vector set Vm can be understood as the "millimeter wave detection direction", and the mean value of the UWB echo feature vector set Vu can be understood as the "UWB reference direction"; the difference divided by the reference is exactly the "ratio of opposite side to adjacent side" of a right triangle; taking the arctangent, we naturally get an included angle, which is the physical meaning of the installation angle deviation.

[0029] After obtaining the installation angle deviation △Px, the deviation angle assessment unit performs a preliminary comparison and assessment with the preset angle threshold Pth to determine the degree of deviation in the installation angle. Based on the preliminary comparison and assessment results, it triggers translation correction. The specific assessment content is as follows: When the installation angle deviation ΔPx is less than or equal to the angle threshold Pth, the angle deviation is considered to be within an acceptable range, and no calibration is required. When the installation angle deviation △Px is greater than the angle threshold Pth, it is determined that the installation angle is obviously wrong and the millimeter wave result can no longer match UWB. Then, translation correction is triggered to further compensate and correct the installation position. The angle threshold Pth is obtained by collecting echo signals from millimeter-wave radar and UWB system respectively under standard tunnel test scenario, gradually adjusting the installation angle of millimeter-wave radar, measuring the change curve of system detection accuracy under different angle deviations, and statistically analyzing the maximum installation deviation angle when the detection performance is still within an acceptable range.

[0030] In this embodiment, the system's installation angle deviation analysis unit calculates the installation angle deviation ΔPx by comparing the mean difference between the millimeter-wave echo feature vector set Vm and the UWB echo feature vector set Vu. The reason for using a normalization method of dividing the difference by the mean is to eliminate the influence of overall echo strength differences under different environments. For example, in a narrow tunnel, the energy of all paths will be enhanced overall; if the original difference is directly compared, it is easy to misjudge it as an installation deviation. Normalization ensures that the calculation result only reflects the directional deviation and is not affected by the signal energy level, thereby improving the reliability of the angle deviation determination. During the calculation process, the arctangent function is used to map the normalized ratio to an angle value. Its physical meaning is that the deviation problem is essentially an angular relationship. If the millimeter-wave detection direction is completely consistent with the UWB reference direction, the two should be nearly coincident; once there is a deviation in the radar installation angle, the deviation amount and the reference mean form a right-angled triangle relationship between opposite and adjacent sides, which can be accurately converted into the actual angle using the arctangent function. The advantage of this setting is that it avoids the erroneous understanding of directly substituting the intensity difference for the angle, truly achieving geometric alignment in direction. The deviation angle evaluation unit judges the installation angle deviation ΔPx by setting an angle threshold Pth. Pth is not set arbitrarily, but is obtained by gradually adjusting the millimeter-wave radar installation angle under a standard tunnel test scenario and recording the decrease curve of the system's detection accuracy. For example, when the radar's forward detection range is 100 meters, a deviation of less than ±1° will not significantly affect the detection; however, if the deviation exceeds ±3°, the target will show obvious deviation or even be lost. Based on this threshold obtained from experimental statistics, it can be ensured that the evaluation standard is consistent with the actual physical detection capability. When the installation angle deviation ΔPx is less than or equal to Pth, the system considers the installation error to be within an acceptable range and no further calibration is required, thus avoiding the additional noise that may be introduced by overcompensation; when ΔPx exceeds Pth, it indicates that the radar installation direction has seriously affected the matching between millimeter wave and UWB. At this time, translation correction is triggered to compensate for the installation error in a timely manner and avoid unreliable detection results caused by angle misalignment. The practical effect of this implementation method is that by combining a normalized ratio with an arctangent mapping, the installation angle deviation ΔPx is accurately obtained, and a reasonable judgment is made using a threshold Pth obtained from actual measurements and statistics. This ensures that the installation deviation can be stably evaluated under different environments and signal strength conditions. Its beneficial effect is a significant improvement in the installation robustness of vehicle-mounted millimeter-wave radar in complex scenarios such as tunnels and underground parking garages, avoiding problems such as misjudgment of direction and reduced detection capability caused by installation deviations, thereby ensuring the stability and accuracy of the system in actual operation.

[0031] Example 4 Please see Figure 1 Specifically: the translation deviation correction module includes a translation deviation analysis unit; After triggering translation correction, the translation deviation analysis unit adjusts the phase parameter Xw of the i-th UWB reflection path. u，i The phase parameter Xw of the j-th UWB reflection path with different reflection paths u，j The phase difference ΔXw is obtained by performing difference calculation, and the translation deviation Δd is calculated and output based on the phase difference ΔXw. The translation deviation Δd is calculated and output using the following algorithm formula: In the formula, c represents the propagation speed of electromagnetic waves, which is dimensionless; This formula originates from the electromagnetic wave propagation and phase difference ranging formula in electromagnetic wave propagation theory. It is based on the phase propagation law of waves in physics, that is, when electromagnetic waves generate a propagation path difference between adjacent receiving points, a phase difference will be caused. The path difference and the phase difference satisfy a linear proportional relationship. Based on this, this formula directly defines the calculation result of the path difference as the installation translation deviation of the millimeter-wave radar, thereby realizing the quantification and correction of installation error. The physical dimensions on both sides of the formula are consistent and are dimensionless.

[0032] In this embodiment, the system's translational deviation analysis unit calculates the phase difference ΔXw by comparing the phase parameters Xwu,i and Xwu,j of different UWB reflection paths, and further converts it into a translational deviation Δd. The reason for choosing phase difference as the correction basis is that when the millimeter-wave radar undergoes lateral or longitudinal translation during installation, even a deviation of only a few centimeters will create a stable phase difference on the propagation path of the UWB echo. If judgment is made directly based on distance or energy, it is easily affected by fluctuations in the reflection intensity of the tunnel walls. Calculation based on phase difference can identify positional shifts with subwavelength accuracy, ensuring the sensitivity and reliability of the correction. Its physical meaning is that during electromagnetic wave propagation, the path difference and phase difference always maintain a linear proportional relationship. When the millimeter-wave radar at the front of the vehicle translates relative to the reference installation position, the propagation path difference between the receiving points is converted into a phase difference ΔXw. This module directly maps this difference to a translational deviation Δd using the phase difference formula, achieving a quantitative conversion from "phase error" to "position error." The core purpose of this setup is to ensure that translational errors remain detectable regardless of fluctuations in external signal strength, effectively avoiding detection blind spots caused by relying solely on amplitude information. For example, in a tunnel environment, if the millimeter-wave radar's installation position is offset 2 cm to the left, although the overall detection energy attenuation is not significant, the target echo will experience a path deviation, causing the detection result to shift relative to the UWB reference direction, ultimately affecting the accuracy of vehicle lateral distance measurement. This implementation method, based on phase difference correction, can identify this shift in real time and convert it into a translational deviation Δd for compensation, ensuring that the detection result remains consistent with the vehicle's actual geometric position. Therefore, the beneficial effect of this implementation method is that by mapping phase difference to translational deviation, it overcomes the insensitivity of traditional energy or time delay methods to small-scale shifts, achieving millimeter-level installation error identification and correction. This significantly improves the detection accuracy and stability of millimeter-wave radar in confined environments, avoiding vehicle lateral positioning deviations caused by installation translational errors.

[0033] Example 5 Please see Figure 1 and Figure 2 Specifically: the calibration control module includes a comprehensive calibration unit, a calibration evaluation unit, and a secondary optimization unit; The integrated calibration unit, based on the initially set reference angle calibration matrix Mo, introduces the installation angle deviation ΔPx and the translation deviation Δd respectively, and performs linear adjustment through the convergence control coefficient Q to obtain the integrated calibration matrix Mfinal. The comprehensive calibration matrix Mfinal is calculated and output using the following algorithm: Mfinal=Mo+Q △Px ·△Px+Q △d· △d; where Mo represents the reference angle calibration matrix, which is dimensionless. During factory calibration, signals are collected according to the ideal installation angle, and the reference matrix is ​​obtained through experimental fitting or system calibration. It is usually a fixed array directivity matrix containing the reference phase relationship of each antenna element. In the formula, when both the installation angle deviation △Px and the translation deviation △d are zero, the system defaults to using the reference angle calibration matrix Mo. When an angle or translation deviation is detected, a correction term is introduced on this reference. Q △Px and Q △d Let Q represent the convergence control coefficients for the installation angle deviation ΔPx and the translation deviation Δd, respectively. These coefficients are dimensionless, and Q... △Px +Q △d =1; Source of angle deviation correction: In radar installation, if there is an installation angle deviation △Px, it will cause the receiving beam direction to be at an angle with the target direction; the basic idea of ​​angle compensation comes from geometric trigonometry and the principle of direction cosine correction: that is, by introducing a correction amount proportional to the angle deviation, the directivity matrix of the receiving array is adjusted. Source of translational deviation correction: In radar installation, if there is a positional translational deviation Δd, the path length of the target echo will exhibit a systematic deviation; the formula for calculating the translational deviation Δd is derived from the electromagnetic wave propagation and phase difference ranging formulas. In mathematics, linear combination methods are often used to map multiple correction factors onto the same reference matrix. This formula is based on this idea, and introduces the installation angle deviation ΔPx and the translation deviation Δd into the reference angle calibration matrix Mo in a linear form. The reference angle calibration matrix Mo represents the system reference matrix under ideal installation conditions. The convergence control coefficient Q is determined by iterative learning. The iterative learning method is based on the statistical characteristics of historical calibration data and vehicle operation scenarios. It dynamically updates the correction weights of installation angle deviation △Px and translation deviation △d, thereby improving the convergence speed and stability of the comprehensive calibration matrix. Dimensional consistency analysis: The installation angle deviation △Px and the translation deviation △d are dimensionless parameters. The reference angle calibration matrix Mo and the convergence control coefficient Q are also dimensionless. Therefore, the output of the comprehensive calibration matrix Mfinal is also dimensionless.

[0034] After generating the comprehensive calibration matrix Mfinal, the calibration evaluation unit calibrates the installation of the UWB vehicle-mounted millimeter-wave radar based on the comprehensive calibration matrix Mfinal. After calibration, the comprehensive calibration matrix Mfinal is applied to the multi-source echo data of the current UWB vehicle-mounted millimeter-wave radar, and the installation angle deviation ΔPx and translation deviation Δd are recalculated to obtain the corrected detection result set Rfinal. Based on the detection result set Rfinal and the ideal detection result Rideal, residuals are calculated to obtain residual values. A preset residual threshold Eth is then compared with the residual values ​​to evaluate the effectiveness of the correction. The specific evaluation content is as follows: When the residual value is less than or equal to the residual threshold Eth, the correction is deemed valid, the calibration is completed, and the comprehensive calibration matrix Mfinal is output. When the residual value is greater than the residual threshold Eth, the correction is deemed invalid, and the process proceeds to the second optimization. The residual threshold Eth stores historical calibration records through the vehicle's central control system and dynamically corrects the threshold using machine learning models (such as Kalman filtering or Bayesian estimation). For example, when it is detected that 95% of the residual values ​​in the past 50 corrections are below a certain level, the system will automatically update the threshold to that level.

[0035] The secondary optimization unit enters a secondary optimization phase after determining that the correction is invalid. The secondary optimization phase involves re-acquiring multi-source echo parameters to form a new UWB echo feature vector set Vu and a millimeter-wave echo feature vector set Vm. Based on the new UWB echo feature vector set Vu and the millimeter-wave echo feature vector set Vm, the installation angle deviation and translation deviation results are recalculated. The recalculated installation angle deviation ΔPx and translation deviation Δd are then fused to form an updated comprehensive calibration matrix Mfinal. During the fusion process, an iterative convergence mechanism is used to iteratively correct the updated calibration matrix multiple times until the corrected residual value is less than or equal to the residual threshold Eth condition.

[0036] In this embodiment, the system's integrated calibration unit incorporates the installation angle deviation ΔPx and translation deviation Δd into the reference angle calibration matrix Mo, and uses the convergence control coefficient Q for linear adjustment to obtain the integrated calibration matrix Mfinal. This design aims to avoid the excessive influence of a single correction factor on the results. If only the angle deviation ΔPx is considered, errors are easily amplified by multipath interference in a tunnel environment; while if only the translation deviation Δd is compensated, the directivity of the receiving beam is ignored. Through linear fusion, the two deviations are incorporated into the same matrix framework, ensuring that the final correction result is both stable and convergent, physically equivalent to maintaining consistency of the calibration matrix in both "angle space" and "position space." After generating the integrated calibration matrix Mfinal, the calibration evaluation unit does not directly output the result, but performs self-checking and secondary judgment by comparing the residuals of the detection result set Rfinal with the ideal result Rideal. This setting aims to address the fact that in real-world environments, multipath effects and interference signals may cause deviations to remain after a single correction. Without residual verification, a situation may arise where the result appears stable but is distorted. Dynamic evaluation using the residual threshold Eth allows the system to filter out seemingly correct but invalid corrections in real time. Furthermore, machine learning methods are introduced to dynamically update the residual threshold Eth, enabling the system to adaptively adjust the evaluation criteria based on historical operating experience, avoiding insufficient adaptability caused by a fixed threshold. When the correction is still ineffective, a secondary optimization unit is triggered to re-acquire data and perform iterative convergence correction. The purpose of this mechanism is to prevent the problem of "single-correction solidifying error". For example, when a vehicle is traveling in a slippery tunnel, the UWB signal path may be momentarily interfered with, causing the one-time correction to fail. Without the secondary optimization mechanism, the system will output an incorrect calibration matrix. However, through iterative convergence, re-acquiring and fusing multiple rounds of data, the influence of occasional noise can be gradually weakened, allowing the residual to eventually stabilize within the threshold range, ensuring the robustness of the calibration. Therefore, the beneficial effects of this implementation method are: simultaneous correction of multi-dimensional errors through matrix fusion, ensuring the reliability of the results through residual evaluation, and ensuring correct calibration can still be restored in extreme scenarios through secondary optimization. The final effect is to significantly improve the installation consistency of millimeter-wave radar in complex environments, reduce ranging and angle detection errors caused by installation deviations, and maintain centimeter-level positioning and sensing accuracy, especially in high multipath environments such as tunnels and underground parking garages.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A UWB-based vehicle-mounted millimeter-wave radar installation and calibration system, characterized in that: It includes an echo joint acquisition module, an angle deviation calibration module, a translation deviation correction module, and a calibration control module; The echo joint acquisition module synchronously acquires multi-source echo parameters inside the tunnel wall through the vehicle-mounted UWB transmitter, UWB receiver array, and millimeter-wave radar equipment, and transmits the multi-source echo parameters to the vehicle-mounted central control system to form UWB echo feature vector set Vu and millimeter-wave echo feature vector set Vm. The angle deviation calibration module calculates the installation angle deviation △Px by comparing the UWB echo feature vector set with the millimeter wave echo feature vector set, and performs a preliminary comparison and evaluation with the preset angle threshold Pth. Based on the preliminary comparison and evaluation results, it triggers translation correction. The translation deviation correction module calculates the translation deviation Δd of the millimeter-wave radar based on the phase difference between the UWB receiving arrays after triggering translation correction. The calibration control module combines the installation angle deviation △Px and the translation deviation △d to form a comprehensive calibration matrix Mfinal. If the correction result corresponding to the comprehensive calibration matrix Mfinal meets the preset threshold condition, the comprehensive calibration matrix is ​​output; otherwise, re-acquisition and secondary optimization are triggered.

2. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 1, characterized in that: The echo joint acquisition module includes a UWB echo acquisition unit, a millimeter-wave echo acquisition unit, and a vector generation unit; The UWB echo acquisition unit transmits UWB signals through a UWB transmitter installed on the vehicle, and the reflected signals are received by a UWB receiving array. The reflected signals are then feature-extracted in the UWB receiving array to obtain UWB echo parameters. The UWB echo parameters include the echo intensity parameter A of the i-th UWB reflection path. u，i The echo delay parameter T of the i-th UWB reflection path u，i The phase parameter Xw of the i-th UWB reflection path u，i .

3. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 2, characterized in that: The millimeter-wave echo acquisition unit acquires millimeter-wave echo signals with the same propagation path as the UWB signal through a millimeter-wave radar device installed on the vehicle, and performs signal processing on the echo signals to obtain millimeter-wave echo parameters. The millimeter-wave echo parameters include the echo intensity parameter A of the k-th path of the millimeter wave. m,k The azimuth parameter O of the k-th path of the millimeter wave m,k And the path identifier parameter L of the k-th path of the millimeter wave. m,k ; The millimeter-wave echo parameters and UWB echo parameters are combined to generate multi-source echo parameters.

4. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 3, characterized in that: The vector generation unit encapsulates multi-source echo parameters in the form of data frames and transmits them to the vehicle central control system via the vehicle bus. The data frames contain timestamp information. After receiving the multi-source echo parameters, the vehicle central control system performs dimensionless processing, eliminating the dimensions of all parameters in the multi-source echo parameters and normalizing them to the same range. Then, it synchronizes the UWB echo parameters and millimeter-wave echo parameters based on the timestamp information of the dimensionless multi-source echo parameters, and generates the UWB echo feature vector set Vu and the millimeter-wave echo feature vector set Vm.

5. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 4, characterized in that: The angle deviation calibration module includes an installation angle deviation analysis unit and a deviation angle evaluation unit; The installation angle deviation analysis unit calculates the installation angle deviation △Px based on the UWB echo feature vector set Vu and the millimeter wave echo feature vector set Vm. The installation angle deviation △Px is the difference between the mean values ​​of the millimeter wave echo feature vector set Vm and the UWB echo feature vector set Vu, and is obtained by taking the arctangent value of the ratio and comparing it with the mean value of the UWB echo feature vector set Vu.

6. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 5, characterized in that: The deviation angle evaluation unit, after obtaining the installation angle deviation △Px, performs a preliminary comparison evaluation between the installation angle deviation △Px and the preset angle threshold Pth to determine the degree of deviation of the installation angle. Based on the preliminary comparison evaluation results, it triggers translation correction. The specific evaluation content is as follows: When the installation angle deviation ΔPx is less than or equal to the angle threshold Pth, the angle deviation is considered to be within an acceptable range, and no calibration is required. When the installation angle deviation △Px is greater than the angle threshold Pth, it is determined that the installation angle is obviously wrong, and translation correction is triggered.

7. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 6, characterized in that: The translation deviation correction module includes a translation deviation analysis unit; The translation deviation analysis unit, after triggering translation correction, sets the phase parameter Xw of the i-th UWB reflection path. u，i The phase parameter Xw of the j-th UWB reflection path with different reflection paths u，j The phase difference ΔXw is obtained by performing difference calculation, and the translation deviation Δd is calculated and output based on the phase difference ΔXw.

8. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 1, characterized in that: The calibration control module includes a comprehensive calibration unit, a calibration evaluation unit, and a secondary optimization unit; The integrated calibration unit, based on the initially set reference angle calibration matrix Mo, introduces the installation angle deviation ΔPx and the translation deviation Δd respectively, and performs linear adjustment through the convergence control coefficient Q to obtain the integrated calibration matrix Mfinal.

9. The vehicle-mounted millimeter-wave radar installation and calibration system based on UWB according to claim 8, characterized in that: After generating the comprehensive calibration matrix Mfinal, the calibration evaluation unit calibrates the installation of the UWB vehicle-mounted millimeter-wave radar based on the comprehensive calibration matrix Mfinal. After calibration, the comprehensive calibration matrix Mfinal is applied to the multi-source echo data of the current UWB vehicle-mounted millimeter-wave radar, and the installation angle deviation ΔPx and translation deviation Δd are recalculated to obtain the corrected detection result set Rfinal. Based on the detection result set Rfinal and the ideal detection result Rideal, residual calculation is performed to obtain the residual value. A preset residual threshold Eth is then used to compare and evaluate the residual value to determine the effectiveness of the correction. The specific evaluation content is as follows: When the residual value is less than or equal to the residual threshold Eth, the correction is deemed valid, the calibration is completed, and the comprehensive calibration matrix Mfinal is output. When the residual value is greater than the residual threshold Eth, the correction is deemed invalid, and the process proceeds to the second optimization.

10. The UWB-based vehicle-mounted millimeter-wave radar installation and calibration system according to claim 9, characterized in that: The secondary optimization unit enters secondary optimization after determining that the correction is invalid. The secondary optimization re-acquires multi-source echo parameters to form a new UWB echo feature vector set Vu and a millimeter-wave echo feature vector set Vm. Based on the new UWB echo feature vector set Vu and millimeter-wave echo feature vector set Vm, the installation angle deviation result and translation deviation result are recalculated. The recalculated installation angle deviation ΔPx and translation deviation Δd are then fused to form an updated comprehensive calibration matrix Mfinal. During the fusion process, an iterative convergence mechanism is used to iteratively correct the updated calibration matrix multiple times until the corrected residual value is less than or equal to the residual threshold Eth condition.

Images