A road noise active noise reduction bench simulator device and a noise reduction method

Abstract

This invention discloses a bench simulator and noise reduction method for active road noise reduction, belonging to the field of automotive vibration and noise development technology. The device includes a driving simulator, a road noise spatial sound field reproduction system, a road noise active noise reduction system, and a data conversion system. The driving simulator dynamically generates in-vehicle binaural noise signals and under-vehicle acceleration reference signals based on vehicle speed. The data conversion system synchronously distributes signals to the sound field reproduction system and the active noise reduction system via optical fiber. The sound field reproduction system reconstructs the in-vehicle sound field using a multi-speaker array. The active noise reduction system combines the acceleration signal with feedback from the reference microphone to generate a reverse noise reduction wave. This invention solves the problems of low simulation accuracy and high real-vehicle testing costs in existing technologies, supports high-precision RNC algorithm debugging in a laboratory environment, and is suitable for the development of various vehicle models.

Description

Technical Field

[0001] This invention relates to the field of automotive vibration and noise development technology, and in particular to a bench simulator and noise reduction method for active road noise reduction, used for the development and verification of Road Noise Cancellation (RNC) algorithms, applicable to the early development stages of passenger cars, commercial vehicles and other vehicle types. Background Technology

[0002] In the automotive R&D process, road noise control is a key aspect of improving in-vehicle comfort. Current active road noise reduction (RNC) tuning mainly relies on pure computer simulation or real vehicle testing, but both of these methods have significant drawbacks, severely limiting the development efficiency and accuracy of RNC algorithms.

[0003] Current active noise cancellation tuning for road noise mainly relies on pure computer simulation or real vehicle testing. Computer simulation cannot reproduce the actual frequency response of the speaker and hardware delay, leading to algorithm errors; real vehicle testing is costly and limited by the vehicle development stage.

[0004] Pure computer simulation uses finite element models or statistical energy analysis to simulate in-vehicle noise; however, it has significant shortcomings in terms of hardware characteristic reproduction and binaural perception representation.

[0005] Lack of hardware characteristics: The frequency response curve of the speaker cannot be accurately reproduced. Real-world measurement data shows that real speakers typically exhibit fluctuations of ±3dB within their operating frequency range, while simulation models usually simplify this to an ideal linear response. This means that the nonlinear characteristics of the actual hardware cannot be considered during algorithm debugging. For example, the measured frequency response of a certain speaker model at 1kHz deviates from the simulated value by 4.2dB, directly affecting the accurate generation of the noise reduction signal.

[0006] Phase delay distortion: The phase delay of a hardware system (typically 0.5-5ms) is difficult to accurately model in simulation. Taking an audio processing link as an example, the measured overall delay is 2.8ms, while the simulation model can only roughly estimate it to be 1ms. This deviation will cause phase mismatch between the reverse sound wave and the original noise, reducing the noise reduction effect.

[0007] Nonlinear distortion is ignored: The simulation does not consider the nonlinear distortion (THD≥1%) of the loudspeaker. When the input signal amplitude is large, the nonlinear characteristics of the loudspeaker will introduce additional harmonics. The measured THD of an 8-inch loudspeaker at a sound pressure level of 90dB reaches 1.8%, but the simulation model completely ignores this, which leads to a decrease in the performance of the algorithm in practical applications.

[0008] Binaural perception bias: The simplified processing of the artificial head transfer function (HRTF) results in a spatial perception bias of ≥5dB for binaural noise. Actual measurements using the artificial head microphone revealed that the simulated binaural signals significantly malfunction in mimicking the auricular filtering effect, particularly in the 8-12kHz frequency band, where spatial localization errors can reach 7dB, severely impacting the subjective evaluation accuracy of the noise reduction algorithm.

[0009] Real-world road testing relies on actual vehicles for RNC debugging, but its high cost, long cycle, and environmental interference severely hinder the rapid iteration of algorithms.

[0010] High economic costs: A single real-vehicle test incurs costs for fuel, tire wear, and personnel hours, among others. Taking a mid-size passenger vehicle as an example, the daily testing cost is approximately ¥5,000-10,000. If full-condition debugging is carried out, the cost of the road testing phase alone may exceed ¥500,000, which is a huge burden for companies developing multiple models in parallel.

[0011] The development cycle is lengthy: Real-vehicle testing requires waiting for the completion of the vehicle assembly, which typically lags behind electronic system development by 6-12 months. During the development of a certain new energy vehicle, the RNC algorithm was delayed by 9 months before entering the debugging stage due to waiting for the real vehicle, resulting in a prolonged project cycle and a missed market opportunity.

[0012] Significant environmental interference: Road background noise (wind noise, ambient sound) can reduce the test signal-to-noise ratio (SNR) by ≥10dB. In urban road tests, the instantaneous peak value of background noise can reach 75dB(A), while the amplitude of the target road noise signal in the mid-to-high frequency band is only about 60dB(A), resulting in a signal-to-noise ratio as low as -15dB, which seriously affects the debugging accuracy and makes it difficult to adjust the algorithm parameters.

[0013] Limited operating condition coverage: Real-world vehicle testing cannot fully cover all kinds of extreme operating conditions. For example, in a low temperature environment of -30℃, changes in the stiffness of the vehicle suspension system will lead to a significant change in road noise characteristics. However, due to limitations in testing conditions, the debugging of such operating conditions is often simplified, affecting the environmental adaptability of the algorithm.

[0014] The core problem with existing technologies lies in the inability to construct a high-fidelity road noise simulation system in a laboratory environment to achieve a full-element reproduction of "hardware characteristics, binaural perception, and dynamic operating conditions." A technological gap exists between pure computer simulation and real-vehicle testing; the former lacks physical realism, while the latter lacks flexibility and controllability. This leads to a vicious cycle in RNC algorithm development: inaccurate simulation – real-vehicle trial and error – high costs. Therefore, there is an urgent need for a bench system capable of accurately simulating real-vehicle road noise environments in a laboratory setting to overcome existing technological bottlenecks and improve the efficiency and quality of algorithm development. Summary of the Invention

[0015] In view of the above-mentioned deficiencies of the prior art, the present invention aims to provide a bench simulator device and noise reduction method for active road noise reduction. By constructing a high-fidelity three-dimensional sound field reproduction system, it realizes accurate laboratory simulation of the actual vehicle road noise environment, solves the problems of low simulation accuracy and high actual vehicle testing cost in the prior art, and provides an efficient and accurate development and verification platform for RNC algorithm.

[0016] To achieve the above objectives, this invention provides a benchtop simulator device for active road noise reduction. Its core lies in the collaborative operation of multiple modules to achieve dynamic generation of road noise signals, sound field reconstruction, active noise reduction, and synchronous signal transmission. Specifically, it includes the following mutually cooperating systems:

[0017] The driving simulator incorporates a speed-based road noise generation model to output real-time binaural noise signals and undercarriage acceleration reference signals. The simulator includes a dynamic road noise generation module and a driving interaction unit. The dynamic road noise generation module calculates noise based on the vehicle transfer function (VTF) model, and the low-frequency road noise amplitude meets the following requirements: The driving interaction unit includes a steering wheel torque feedback module and a seat vibration module;

[0018] The road noise spatial sound field reproduction system includes a speaker array, a subwoofer, and a sound field calibration module, which is used to reconstruct the three-dimensional sound field inside the vehicle. The speaker array includes 4-8 full-range speakers with a frequency response range of 45-20kHz and a maximum sound pressure level of ≥105dB, which are arranged in the ceiling, seat headrests, and foot areas, and 1-4 subwoofers with a frequency response range of 20-150Hz and a power of ≥150W, which are arranged under the vehicle floor.

[0019] The active noise cancellation system includes an RNC algorithm controller, a reference microphone array, and a noise-canceling speaker, used to generate inverse acoustic waves to cancel noise. The RNC algorithm controller includes an adaptive filtering module and a delay compensation module. The adaptive filtering module is based on the FxLMS algorithm, with an adjustable order of 64-512 and a convergence step size μ=0.001-0.1. The delay compensation module pre-generates inverse acoustic waves that are 0.5-5ms ahead to compensate for hardware delay.

[0020] The data conversion system achieves high-precision transmission and distribution of multi-channel signals through an optical fiber synchronization module. The data conversion system includes an ADAT optical fiber synchronization box, which supports 8-channel 24bit / 96kHz audio transmission with a synchronization error of ≤0.5ms, and a multi-channel distributor, which distributes the binaural noise signal to the speaker channel of the sound field reproduction system and the acceleration signal to the reference channel of the noise reduction system.

[0021] Preferably, the spatial layout of the full-range loudspeakers satisfies the following: the Z-coordinate of the ceiling loudspeakers is ≥1.5m, and the horizontal spacing is 0.8-1.2m; the headrest loudspeakers are embedded inside the seat, with a spacing of 0.3±0.05m.

[0022] Preferably, the RNC algorithm controller automatically adjusts parameters under dynamic operating conditions:

[0023] When the vehicle speed is 0-60km / h, the filter order N=128 and the step size μ=0.01;

[0024] When the vehicle speed is 60-100km / h, the filter order N=256 and the step size μ=0.005;

[0025] When the vehicle speed is greater than 100 km / h, the filter order N = 512 and the step size μ = 0.002.

[0026] Preferably, the reference microphone array includes a front axle accelerometer with a sensitivity of 50 or 100 mV / g and a frequency response of 0.5-1000 Hz, and 1-4 reference microphones arranged at the bottom of the seat and the middle of the ceiling with a dynamic range ≥120 dB.

[0027] Preferably, the sound field calibration module uses an artificial head microphone to collect binaural noise and compensates for amplitude ±6dB and phase ±180° with a 1024th order FIR filter.

[0028] Preferably, it supports switching between multiple vehicle modes. In sedan mode, the sound field reproduction system adopts a 7-speaker + 1 subwoofer layout with a low-frequency cutoff frequency of 30Hz; in SUV mode, it adopts an 8-speaker + 2 subwoofer layout with a low-frequency cutoff frequency of 20Hz and a sound pressure level increase of 3dB.

[0029] Preferably, the steering wheel torque feedback module of the driving interaction unit has a torque... The unit is N·m. The seat vibration module adjusts the amplitude according to the road spectrum frequency f, using the following formula: It can simulate the vibration characteristics of seats under different road surface conditions.

[0030] Preferably, the mid-to-high frequency noise generated by the dynamic road noise generation module is generated by the tire tread pattern and suspension stiffness parameters to form an octave spectrum, and the 1 / 3 octave band fluctuation is ≤3dB.

[0031] Preferably, the noise-canceling speakers are arranged at the four corners of the vehicle cabin to form a multi-channel canceling sound field with a frequency response range of 60-16kHz.

[0032] Another aspect of the present invention provides a noise reduction method for a bench simulator device based on the above-mentioned active road noise reduction, characterized by comprising the following steps:

[0033] A mixed sound field of road noise and motor noise is generated by using vehicle speed signals. The road noise signal is calculated based on the vehicle transfer function (VTF) model, and the motor noise is generated by matching with a database.

[0034] The sound field transfer function is calibrated using an artificial head microphone to generate an FIR equalization filter to compensate for frequency response deviations during the sound field reproduction process.

[0035] The algorithm calculates the noise reduction signal in real time based on the FxLMS algorithm and outputs the reverse sound wave through the delay compensation module. The algorithm dynamically adjusts the filter order and convergence step size according to the vehicle speed.

[0036] Preferably, the FIR equalization filter uses the Kaiser window function, β=8, order 1024, compensation amplitude ±6dB, and phase ±180°.

[0037] Preferably, the mixing ratio of road noise and motor noise in the mixed sound field is adjusted according to the noise characteristics of different vehicle models to ensure that the similarity to the noise environment of the actual vehicle is ≥90%.

[0038] Beneficial technical effects of the present invention:

[0039] (1) Full-element dynamic simulation: For the first time, the dynamic coupling simulation of "vehicle speed-road noise-driving interaction" has been realized. The driving simulator not only generates road noise signals related to vehicle speed, but also provides a realistic driving experience through steering wheel torque feedback and seat vibration module, so that the debugging personnel can feel a driving scene close to the real vehicle in the laboratory environment, which is something that existing pure computer simulation cannot achieve.

[0040] (2) High-precision sound field reproduction: The road noise spatial sound field reproduction system achieves high-fidelity reproduction of the three-dimensional sound field through multi-speaker array layout and precise sound field calibration. Compared with the traditional dual-speaker stereo simulation, the sound field reproduction system of the present invention improves the binaural positioning accuracy by 40% and extends the low-frequency response range by 15Hz, which can more realistically simulate the spatial distribution characteristics of in-vehicle noise.

[0041] (3) Adaptive algorithm control: The RNC algorithm controller can dynamically adjust the filter order and convergence step size according to the vehicle speed, and can maintain good noise reduction effect under different vehicle speed conditions. The test data shows that this adaptive strategy shortens the convergence time of the algorithm by 50% when the vehicle speed changes abruptly, and has stronger adaptability to operating conditions compared with the algorithm with fixed parameters.

[0042] (4) Multi-vehicle compatibility design: The device supports switching between multiple vehicle modes. By adjusting the speaker layout, subwoofer configuration, and algorithm parameters, it can be applied to road noise simulation for different vehicle models such as sedans and SUVs. For example, the sedan mode uses a 7-speaker + 1 subwoofer layout with a low-frequency cutoff frequency of 30Hz; the SUV mode uses an 8-speaker + 2 subwoofer layout with a low-frequency cutoff frequency of 20Hz, increasing the sound pressure level by 3dB. This flexible configuration greatly improves the versatility and practicality of the device.

[0043] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0044] Figure 1 This is a system structure diagram of a preferred embodiment of the present invention.

[0045] This figure illustrates the overall structure of the road noise active noise reduction bench simulator of the present invention, comprising four main modules: a driving simulator (101), a road noise spatial sound field reproduction system (102), a road noise active noise reduction system (103), and a data conversion system (104). The driving simulator generates binaural noise signals and acceleration reference signals through a dynamic road noise generation module, which are then transmitted to the sound field reproduction system and the active noise reduction system via the ADAT fiber optic synchronization box of the data conversion system. The speaker array of the sound field reproduction system reconstructs the in-vehicle sound field based on the binaural noise signals, while the active noise reduction system uses the acceleration signal and reference microphone feedback to generate inverse sound waves to achieve noise cancellation. The figure clearly indicates the signal flow and connection relationships between the modules, demonstrating the system's integration and collaborative working principle. Detailed Implementation

[0046] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0047] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0048] like Figure 1 As shown, the road noise active noise reduction bench simulator device provided by the present invention is based on the core function of achieving dynamic generation of road noise signals, sound field reconstruction, active noise reduction, and synchronous signal transmission through the collaborative work of multiple modules. Specifically, it includes the following mutually cooperating systems:

[0049] Driving simulator (101): As the system's dynamic signal source, it incorporates a road noise generation model based on vehicle speed, enabling real-time output of in-vehicle binaural noise signals and under-vehicle acceleration reference signals that closely match those of the actual vehicle. This simulator includes:

[0050] Dynamic road noise generation module: Based on the vehicle transfer function (VTF) model, it calculates the noise signal in real time using the vehicle speed v (km / h). For low-frequency road noise in the 20-200Hz range, its amplitude follows the formula... (dB), the model was obtained through regression analysis of road noise data from more than 10 typical vehicle models, with a fitting error ≤2.5dB. The mid-to-high frequency noise (200-5000Hz) is generated by octave spectrum from parameters such as tire tread and suspension stiffness, with 1 / 3 octave band fluctuation controlled within ≤3dB to ensure that the spectral characteristics of mid-to-high frequency noise are consistent with those of the actual vehicle.

[0051] Driving Interaction Unit: Includes a steering wheel torque feedback module and a seat vibration module to enhance the realism of simulated driving. The steering wheel torque feedback increases linearly with vehicle speed, from 0 N·m at 0 km / h to 5 N·m at 120 km / h. This parameter is obtained by fitting a steering wheel torque-vehicle speed curve from a real vehicle, with a correlation coefficient... The seat vibration module adjusts the amplitude according to the road spectrum frequency f (Hz), using the following formula: It can simulate the vibration characteristics of seats under different road surface conditions.

[0052] Road noise spatial sound field reproduction system (102): used to reconstruct the three-dimensional sound field inside the vehicle, including a full-range speaker array, subwoofer and sound field calibration module:

[0053] Speaker Array: A combination of 7-8 full-range speakers and 1-2 subwoofers is used. The full-range speakers are JBL 305P models, with a frequency response of 45-20kHz and a maximum sound pressure level ≥105dB, positioned in the ceiling, headrests, and footwell areas. The ceiling speakers have a Z-coordinate ≥1.5m and a horizontal spacing of 0.8-1.2m; the headrest speakers are embedded within the seats, with a spacing of 0.3±0.05m. This layout, optimized through acoustic simulation, achieves over 90% uniformity of the in-vehicle sound field. The subwoofers are Yamaha HS8S, with a frequency response of 20-150Hz and a power ≥150W, positioned under the vehicle floor to effectively enhance the reproduction of low-frequency road noise.

[0054] Sound field calibration module: Head Acoustics HMS III artificial head microphone is used to collect binaural noise, which is compensated for ±6dB amplitude and ±180° phase using a 1024th-order FIR filter. During calibration, a reference signal of the target sound field is first acquired, and then the difference between the actual reproduced sound field and the reference signal is compared to generate a Kaiser window function (β=8) FIR filter, ensuring that the similarity between the calibrated sound field and the target sound field is ≥90%.

[0055] Active noise cancellation system (103): realizes the generation and cancellation of reverse sound waves, including RNC algorithm controller, reference microphone array and noise-canceling speaker:

[0056] The RNC algorithm controller's core components are an adaptive filtering module and a delay compensation module. The adaptive filtering is based on the FxLMS algorithm, with an adjustable filter order of 64-512 and a convergence step size μ = 0.001-0.1. Under dynamic conditions, the system automatically adjusts parameters: at vehicle speeds of 0-60 km / h, the filter order N = 128, and the step size μ = 0.01; at vehicle speeds of 60-100 km / h, N = 256, and μ = 0.005; at vehicle speeds > 100 km / h, N = 512, and μ = 0.002. This dynamic adjustment strategy, trained with extensive real-vehicle data, ensures the algorithm maintains optimal convergence at different vehicle speeds. The delay compensation module pre-generates a reverse sound wave leading the speaker's frequency response by 0.5-5 ms to compensate for hardware delay. Actual measurements show that this module can control the total delay to within 1 ms.

[0057] Reference microphone array: Includes a front axle accelerometer and 1-4 reference microphones. The front axle accelerometer is selected with a sensitivity of 100mV / g and a frequency response of 0.5-1000Hz, used to collect vibration signals from under the vehicle. The reference microphones are located under the seats and in the center of the headliner, with a frequency response of 20-20000Hz and a dynamic range ≥120dB, capable of accurately picking up in-vehicle noise signals.

[0058] Data conversion system (104): Ensures high-precision synchronous transmission and distribution of signals, including:

[0059] ADAT Fiber Optic Synchronization Box: Supports 8-channel 24bit / 96kHz audio transmission with a synchronization error of ≤0.5ms. The use of fiber optic transmission effectively avoids electromagnetic interference and ensures signal purity.

[0060] Multi-channel distributor: Distributes the binaural noise signal generated by the driving simulator to the speaker channel of the sound field reproduction system, and at the same time distributes the acceleration signal to the reference channel of the noise reduction system. The distribution logic is implemented through hardware timing control to ensure that the synchronization error of the two signals is ≤0.1ms.

[0061] Based on the above-described device, this embodiment provides a noise reduction method that achieves active control of road noise through the following steps:

[0062] Hybrid sound field generation: A hybrid sound field of road noise and motor noise is generated based on the vehicle speed signal. The road noise signal is calculated by the dynamic road noise generation module of the driving simulator based on the VTF model, while the motor noise is generated by matching a database established by collecting actual vehicle motor operating data. The mixing ratio of the two is adjusted according to the noise characteristics of different vehicle models to ensure that the similarity between the hybrid sound field and the actual vehicle noise environment is ≥90%.

[0063] Sound field calibration and filter generation: The sound field transfer function is calibrated using an artificial head microphone to generate an FIR equalizer filter. The specific process is as follows: An artificial head is placed inside a simulated vehicle cabin, a standard test signal is played, and the response signals at both ears are collected. The sound field transfer function is obtained through a system identification method, and then an FIR equalizer filter is designed. This filter can compensate for frequency response deviations during sound field reproduction, ensuring that the error between the reproduced sound field's frequency response and the target frequency response is ≤1.5dB.

[0064] Reverse acoustic wave generation: The noise-reduced signal is calculated in real time based on the FxLMS algorithm, and the reverse acoustic wave is output through a delay compensation module. During the algorithm operation, signals from the reference microphone and accelerometer are collected in real time, and the optimal noise-reduced signal is obtained through adaptive filtering calculation. At the same time, an advance compensation signal is generated according to the group delay characteristics of the loudspeaker to ensure that the reverse acoustic wave is in phase and at the same frequency as the original noise, thus achieving the best noise reduction effect.

[0065] The specific content of the present invention will be illustrated below through specific embodiments and comparative embodiments.

[0066] Example 1: Low-speed road noise tuning in car mode

[0067] (1) Operating conditions

[0068] Vehicle parameters: Simulates a compact sedan with a cabin volume of 2.8m³, seat spacing of 1.1m, target speed of 40km / h, and road conditions of flat asphalt road.

[0069] (2) System Configuration

[0070] Sound field reproduction system: Employs 7 full-range speakers (3 in the ceiling, 2 in the headrests, and 2 in the footwells), using the JBL 305P model, with a frequency response of 45-20kHz and a maximum sound pressure level of 108dB. It is paired with one Yamaha HS8S subwoofer, with a cutoff frequency of 25Hz and a sound pressure level calibrated to 75dB@50Hz. The speaker layout strictly adheres to requirements: ceiling speakers are positioned 1.6m at the Z-coordinate and 1.0m horizontally, while the headrest speakers are spaced 0.3m apart.

[0071] RNC algorithm configuration: Two reference microphones are placed at the bottom of the seat and one in the middle of the ceiling, using electret microphones with a sensitivity of 100dB. The FxLMS algorithm order N=128, convergence step size μ=0.01, noise reduction frequency band is set to 20-300Hz, and the delay compensation module pre-generates a 2.5ms lead signal.

[0072] Driving simulator: Dynamic road noise generation module based on formula The system generates low-frequency road noise signals in the 20-200Hz range, and generates octave band spectra for mid- and high-frequency noise using tire tread parameters. The steering wheel torque feedback is set to 1.6 N·m, and the seat vibration module calculates the amplitude based on a frequency of f=50Hz. .

[0073] (3) Testing process

[0074] Start the driving simulator, set the vehicle speed to 40km / h, and generate road noise signal and acceleration reference signal.

[0075] The data conversion system distributes the signal to the sound field reproduction system and the active noise reduction system through the ADAT fiber optic synchronization box, with a synchronization error of 0.3ms.

[0076] The sound field reproduction system drives the speaker array based on the binaural noise signal to reconstruct the in-vehicle sound field. The artificial head microphone collects the noise signal at both ears, and the frequency response error is ≤1.2dB after calibration.

[0077] The active noise cancellation system receives acceleration signals and reference microphone signals, runs the FxLMS algorithm to generate inverse sound waves, and cancels road noise in real time.

[0078] (4) Test results

[0079] Noise reduction test: The sound level was measured at the driver's right ear using a B&K2669 sound level meter. The A-weighted noise reduction reached 12dB. The noise reduction effect was the best in the mid-high frequency range (200-500Hz), reaching 15dB, and the noise reduction in the low frequency range (20-100Hz) was 8dB.

[0080] Sound field uniformity: The sound pressure level was measured at 9 test points inside the vehicle (driver's head, front passenger's head, rear left and right seat heads, etc.), with a maximum deviation of 3.2dB and a uniformity of 92%.

[0081] System latency: The total latency from signal generation by the driving simulator to reverse acoustic wave output is 0.8ms, which meets the requirements for real-time noise reduction.

[0082] Example 2: High-speed full-frequency noise reduction in SUV mode

[0083] (1) Operating conditions

[0084] Vehicle parameters: Simulates a mid-size SUV, with a cabin volume of 4.5m³, seat spacing of 1.3m, target speed of 100km / h, and road conditions as a highway.

[0085] (2) System Configuration

[0086] Sound field reproduction system: Employs 8 full-range speakers (4 in the ceiling, 2 in the headrests, and 2 in the footwells), JBL 305P model, paired with 2 Yamaha HS8S subwoofers in parallel, with a cutoff frequency of 20Hz and a sound pressure level boosted to 78dB@30Hz. The ceiling speakers are positioned at a Z-coordinate of 1.7m with a horizontal spacing of 1.1m, while the headrest speakers are spaced 0.32m apart.

[0087] RNC algorithm configuration: two reference microphones are placed at the bottom of the seat and two in the middle of the ceiling. The FxLMS algorithm order is N=512, the convergence step size is μ=0.002, the noise reduction frequency band is extended to 20-1000Hz, and the delay compensation is 3.5ms.

[0088] Driving simulator: Low-frequency road noise amplitude The mid-to-high frequency noise is generated based on the SUV suspension stiffness parameters, with a steering wheel torque feedback of 5 N·m, and the seat vibration amplitude is calculated as follows with frequency f=80Hz. .

[0089] (3) Testing process

[0090] The driving simulator operates at a vehicle speed of 100 km / h, generating road noise and acceleration signals.

[0091] The data conversion system transmits signals synchronously, and the sound field reproduction system reconstructs the sound field, focusing on compensating for the low-frequency resonance peak at 80Hz.

[0092] The active noise cancellation system automatically adjusts algorithm parameters based on high-speed operating conditions to generate reverse sound waves in real time.

[0093] (4) Test results

[0094] Noise reduction effect: Noise reduction ≥8dB across the entire frequency band, with a noise reduction of 18dB at the 80Hz resonant peak, achieving a 95% elimination rate. Noise reduction is 10dB in the mid-to-high frequency band (500-1000Hz) and 6dB in the high frequency band (1-5kHz).

[0095] Sound field characteristics: The spatial correlation of the in-vehicle sound field is ≥0.9, and the standing wave ratio at 80Hz is ≤1.3, which effectively suppresses the low-frequency resonance problem common in SUV models.

[0096] Algorithm performance: The algorithm converges in 150ms and runs stably under high-speed conditions without any howling or instability.

[0097] Example 3: Algorithm Adaptability Test under Dynamic Operating Conditions

[0098] (1) Operating conditions

[0099] Operating conditions: The vehicle speed accelerates from 0 to 120 km / h in 30 seconds. The road surface is a Belgian cobblestone road, simulating road noise changes under complex road conditions.

[0100] (2) System configuration:

[0101] Sound field reproduction system: It adopts a car-style layout of 7 speakers + 1 subwoofer, dynamically tracks changes in vehicle speed, and adjusts the speaker gain in real time.

[0102] RNC algorithm configuration: The filter order is automatically adjusted from 128 to 512 according to the vehicle speed, the convergence step size is reduced from 0.01 to 0.001, and the delay compensation module dynamically adjusts the lead time according to the delay of the speaker group.

[0103] Driving simulator: Generates road noise signals in real time that match the acceleration process, including linear increase in amplitude of low-frequency road noise and frequency shift of mid-to-high-frequency noise.

[0104] (3) Testing process

[0105] The system is started and the 0-120km / h acceleration condition is executed, with the data conversion system synchronizing the signal in real time.

[0106] The sound field reproduction system dynamically adjusts the sound field parameters to ensure the accuracy of sound field reproduction at different vehicle speeds.

[0107] The RNC algorithm controller monitors vehicle speed and noise characteristics in real time and automatically optimizes algorithm parameters.

[0108] (4) Test results

[0109] System response: The data conversion system latency is always ≤0.5ms, and the dynamic tracking error of the sound field reproduction system is ≤1dB, which can accurately follow changes in vehicle speed.

[0110] Algorithm adaptability: Dynamic adjustment of filter order and step size enables the algorithm to maintain stable convergence during acceleration, with noise reduction fluctuation ≤2dB.

[0111] Noise reduction effect: The noise reduction is stable at 10±2dB across the entire vehicle speed range, and still maintains 8dB noise reduction at 120km / h, which proves the system's good adaptability under dynamic conditions.

[0112] Comparative Example 1: Traditional Pure Computer Simulation Method

[0113] (1) Simulation settings

[0114] A commercial finite element analysis software was used to build a car body model, and the tire-road excitation spectrum was input to simulate the in-vehicle road noise at a speed of 40 km / h.

[0115] The RNC algorithm uses fixed parameters, a filter order of 128, a step size of 0.01, and does not consider hardware delay compensation.

[0116] (2) Simulation results

[0117] Noise reduction prediction: Software simulation shows a noise reduction of 10dB at the driver's right ear, but does not take into account speaker frequency response deviation and phase delay.

[0118] Differences from the real vehicle: When the noise reduction algorithm obtained from the simulation was applied to the real vehicle test, the actual noise reduction was only 6.5dB, with a deviation of 35%. The main reason is that the simulation did not take into account the 4.2dB frequency response deviation of the speaker at 1kHz and the 2.8ms hardware delay.

[0119] Comparative Example 2: Traditional Real-Vehicle Testing Method

[0120] (1) Test setup

[0121] Using the same passenger car as in Example 1, RNC debugging was performed at a speed of 40 km / h on a flat asphalt road. The cost of a single test was ¥8,000.

[0122] Due to the limitations of real vehicle testing, only constant speed conditions can be tested, and dynamic acceleration processes cannot be simulated.

[0123] (2) Test results

[0124] Noise reduction: After actual vehicle testing, the noise reduction at the driver's right ear was 9dB, which is lower than the 12dB in Embodiment 1 of this invention.

[0125] Development efficiency: It takes 3 days and costs ¥24,000 to complete the debugging under this working condition, while the device of this invention can complete the debugging under the same working condition in only 4 hours in a laboratory environment, at a cost of less than ¥1,000.

[0126] Operating conditions coverage: Real vehicle testing is difficult to cover extreme temperatures (such as -30℃) and complex road conditions, while the device of this invention can easily simulate various environmental conditions through software parameter adjustment.

[0127] in conclusion

[0128] The device of this invention achieves excellent noise reduction performance at different vehicle speeds and vehicle types. In sedan mode at a speed of 40 km / h, the noise reduction reaches 12 dB; in SUV mode at 100 km / h, the noise reduction across the entire frequency band is ≥8 dB, with a 95% elimination rate of the 80Hz resonance peak. Compared with traditional methods, the noise reduction performance is improved by 30%-50%. Under dynamic operating conditions, the noise reduction remains stable at 10±2 dB across the entire vehicle speed range, demonstrating the robustness of the system.

[0129] The road noise spatial sound field reproduction system achieves high-fidelity reproduction of a three-dimensional sound field through multi-speaker array layout and precise calibration. After calibration, the frequency response error between the sound field and the target sound field is ≤1.5dB, the in-vehicle sound field uniformity is ≥90%, and the binaural positioning accuracy is improved by 40%. It can realistically simulate the spatial distribution characteristics of real vehicle noise and provide a reliable acoustic environment for RNC algorithm debugging.

[0130] The device supports rapid algorithm debugging in a laboratory environment, avoiding the high cost and long cycle of real-vehicle testing. Taking Example 1 as an example, the device of this invention can complete debugging in just 4 hours at a cost of less than ¥1,000, while traditional real-vehicle testing requires 3 days and costs ¥24,000, improving development efficiency by 18 times and reducing costs by 96%. This rapid iteration capability shortens the algorithm development cycle by 6-12 months, significantly accelerating the vehicle development progress.

[0131] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A benchtop simulator device for active road noise reduction, characterized in that, include: The driving simulator has a built-in road noise generation model based on vehicle speed, which is used to output in-vehicle binaural noise signals and under-vehicle acceleration reference signals in real time. The driving simulator includes a dynamic road noise generation module and a driving interaction unit. The dynamic road noise generation module calculates noise based on the vehicle transfer function (VTF) model, and the low-frequency road noise amplitude meets the following requirements. The driving interaction unit includes a steering wheel torque feedback module and a seat vibration module; The road noise spatial sound field reproduction system includes a speaker array, a subwoofer, and a sound field calibration module, which is used to reconstruct the three-dimensional sound field inside the vehicle. The speaker array includes 4-8 full-range speakers with a frequency response range of 45-20kHz and a maximum sound pressure level of ≥105dB, which are arranged in the ceiling, seat headrests, and foot areas, and 1-4 subwoofers with a frequency response range of 20-150Hz and a power of ≥150W, which are arranged under the vehicle floor. The active noise cancellation system includes an RNC algorithm controller, a reference microphone array, and a noise-canceling speaker, used to generate inverse acoustic waves to cancel noise. The RNC algorithm controller includes an adaptive filtering module and a delay compensation module. The adaptive filtering module is based on the FxLMS algorithm, with an adjustable order of 64-512 and a convergence step size μ=0.001-0.

1. The delay compensation module pre-generates inverse acoustic waves that are 0.5-5ms ahead to compensate for hardware delay. The data conversion system achieves high-precision transmission and distribution of multi-channel signals through an optical fiber synchronization module. The data conversion system includes an ADAT optical fiber synchronization box, which supports 8-channel 24bit / 96kHz audio transmission with a synchronization error of ≤0.5ms, and a multi-channel distributor, which distributes the binaural noise signal to the speaker channel of the sound field reproduction system and the acceleration signal to the reference channel of the noise reduction system.

2. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The spatial layout of the full-range loudspeakers meets the following requirements: the Z-coordinate of the ceiling loudspeakers is ≥1.5m, and the horizontal spacing is 0.8-1.2m; the headrest loudspeakers are embedded inside the seats, with a spacing of 0.3±0.05m.

3. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The RNC algorithm controller automatically adjusts parameters under dynamic operating conditions: When the vehicle speed is 0-60km / h, the filter order N=128 and the step size μ=0.01; When the vehicle speed is 60-100km / h, the filter order N=256 and the step size μ=0.005; When the vehicle speed is greater than 100 km / h, the filter order N = 512 and the step size μ = 0.

002.

4. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The reference microphone array includes a front axle accelerometer with a sensitivity of 50 or 100 mV / g and a frequency response of 0.5-1000 Hz, and 1-4 reference microphones arranged at the bottom of the seat and the center of the ceiling with a dynamic range ≥120 dB.

5. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The sound field calibration module uses an artificial head microphone to collect binaural noise and compensates for amplitude ±6dB and phase ±180° with a 1024th order FIR filter.

6. The bench simulator device for active road noise reduction according to claim 1, characterized in that, It supports switching between multiple vehicle modes. In sedan mode, the sound field reproduction system adopts a 7-speaker + 1 subwoofer layout with a low frequency cutoff frequency of 30Hz; in SUV mode, it adopts an 8-speaker + 2 subwoofer layout with a low frequency cutoff frequency of 20Hz and a sound pressure level increase of 3dB.

7. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The steering wheel torque feedback module of the driving interaction unit The unit is N·m. The seat vibration module adjusts the amplitude according to the road spectrum frequency f, using the following formula: It can simulate the vibration characteristics of seats under different road surface conditions.

8. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The mid-to-high frequency noise generated by the dynamic road noise generation module is generated by the tire tread pattern and suspension stiffness parameters to form an octave spectrum, with a 1 / 3 octave band fluctuation of ≤3dB.

9. The bench simulator device for active road noise reduction according to claim 1, characterized in that, The noise-canceling speakers are arranged at the four corners of the vehicle cabin to form a multi-channel canceling sound field with a frequency response range of 60-16kHz.

10. A noise reduction method based on the device described in claims 1-9, characterized in that, Including the following steps: A mixed sound field of road noise and motor noise is generated by using vehicle speed signals. The road noise signal is calculated based on the vehicle transfer function (VTF) model, and the motor noise is generated by matching with a database. The sound field transfer function is calibrated using an artificial head microphone to generate an FIR equalization filter to compensate for frequency response deviations during the sound field reproduction process. The algorithm calculates the noise reduction signal in real time based on the FxLMS algorithm and outputs the reverse sound wave through the delay compensation module. The algorithm dynamically adjusts the filter order and convergence step size according to the vehicle speed.

11. The method according to claim 10, characterized in that, The FIR equalizer uses the Kaiser window function, β=8, order 1024, compensation amplitude ±6dB, and phase ±180°.

12. The method according to claim 10, characterized in that, The mixing ratio of road noise and motor noise in the hybrid sound field is adjusted according to the noise characteristics of different vehicle models to ensure that the similarity with the noise environment of the actual vehicle is ≥90%.

Images