Road adhesion coefficient determination method, electronic device and vehicle

Abstract

The invention relates to a road adhesion coefficient determination method, an electronic device and a vehicle. The method comprises the steps of obtaining road surface data collected by a vehicle and a dynamic signal of the vehicle; according to the road surface data, determining a road surface type and a corresponding first confidence coefficient, and determining a first adhesion coefficient estimation value related to the road surface type; determining a second attachment coefficient estimation value of the vehicle and a corresponding second confidence coefficient according to the dynamic signal; and according to the first confidence coefficient and the second confidence coefficient, fusing the first adhesion coefficient estimation value and the second adhesion coefficient estimation value to obtain a target road adhesion coefficient. And traditional single numerical value estimation is expanded into'estimated value + confidence interval 'two-dimensional output, so that the accuracy of road adhesion coefficient evaluation is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle safety technology, and in particular to a method for determining the road surface adhesion coefficient, an electronic device, and a vehicle. Background Technology

[0002] In automotive active safety systems, the road adhesion coefficient (i.e., the friction coefficient between the tires and the road surface) is an input parameter for key control systems such as ABS and ESP, directly affecting vehicle braking stability and handling performance. Traditional technologies mainly rely on onboard sensors (such as wheel speed sensors and inertial measurement units) to indirectly estimate the adhesion coefficient through vehicle dynamics models. However, on low-adhesion road surfaces (such as ice, snow, and wet asphalt) or under complex conditions, sensor noise and model simplification errors can lead to inaccurate estimations. In recent years, computer vision technology has been introduced to assist in recognition. Onboard cameras capture images of the road surface ahead, which are then identified by a deep learning classifier to determine the road surface type (such as dry asphalt or snow), and finally mapped to a predefined adhesion coefficient lookup table.

[0003] The methods for determining the road adhesion coefficient in related technologies are subject to single-source dependence risks. Purely visual methods are easily affected by rain, fog, or low illumination at night, while purely dynamic methods lack sensitivity in low slip rate regions (such as constant speed driving).

[0004] Currently, no effective solution has been provided for the problem of low accuracy in the methods for determining the road surface adhesion coefficient in related technologies. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, electronic device, and vehicle for determining the road surface adhesion coefficient that can improve the accuracy of the assessment, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for determining the road surface adhesion coefficient. The method includes:

[0007] Acquire road surface data and vehicle dynamics signals collected by the vehicle;

[0008] Based on the road surface data, determine the road surface type and its corresponding first confidence level, and determine a first adhesion coefficient estimate related to the road surface type; based on the dynamic signal, determine a second adhesion coefficient estimate for the vehicle and its corresponding second confidence level.

[0009] Based on the first confidence level and the second confidence level, the first adhesion coefficient estimate and the second adhesion coefficient estimate are fused to obtain the target road surface adhesion coefficient.

[0010] In one embodiment, the target road surface adhesion coefficient is obtained by fusing the first adhesion coefficient estimate and the second adhesion coefficient estimate based on the first confidence level and the second confidence level, including:

[0011] Based on the first confidence level and the second confidence level, a first weight and a second weight are determined respectively; wherein, the first weight is proportional to the square of the first confidence level, and the second weight is proportional to the square of the second confidence level;

[0012] Based on the first weight and the second weight, the first adhesion coefficient estimate and the second adhesion coefficient estimate are nonlinearly weighted and fused to obtain the target road surface adhesion coefficient.

[0013] In one embodiment, after fusing the first adhesion coefficient estimate and the second adhesion coefficient estimate based on the first confidence level and the second confidence level to obtain the target road surface adhesion coefficient, the method further includes:

[0014] The target road surface adhesion coefficients are grouped and cached according to their corresponding road surface types.

[0015] In one embodiment, after grouping and caching the target road surface adhesion coefficients according to their corresponding road surface types, the method further includes:

[0016] Determine whether the cached data contains a first road surface type whose cumulative sample count exceeds a preset threshold;

[0017] If the determination is yes, then update the first adhesion coefficient estimate in the database related to the first road surface type.

[0018] In one embodiment, updating the first adhesion coefficient estimate in the database related to the first road surface type includes:

[0019] Based on the cumulative number of samples of the first road surface type, the mean of the adhesion coefficients of multiple target road surfaces is calculated to obtain the mean of the road surface adhesion coefficients.

[0020] The original first adhesion coefficient estimate related to the first road surface type in the database is fused with the mean road surface adhesion coefficient to obtain an updated first adhesion coefficient estimate.

[0021] In one embodiment, the original first adhesion coefficient estimate related to the first road surface type in the database is fused with the mean road surface adhesion coefficient to obtain an updated first adhesion coefficient estimate, including:

[0022] The attenuation factor is determined based on the cumulative number of samples of the first road surface type and the preset time constant.

[0023] Based on the attenuation factor, the weights of the original first adhesion coefficient estimate and the weights of the mean road adhesion coefficient are determined respectively.

[0024] Based on the weights of the original first adhesion coefficient estimate and the weights of the mean road adhesion coefficient, the original first adhesion coefficient estimate and the mean road adhesion coefficient are weighted and summed to obtain the updated first adhesion coefficient estimate.

[0025] In one embodiment, based on the road surface data, determining the road surface type and corresponding first confidence level, and determining a first adhesion coefficient estimate related to the road surface type, includes:

[0026] The road surface data is input into a neural network model for processing.

[0027] The road surface type is output according to the main branch of the neural network model, and the first confidence level is output according to the auxiliary branch of the neural network model.

[0028] Retrieve the first adhesion coefficient estimate associated with the road surface type from the database.

[0029] In one embodiment, determining a second adhesion coefficient estimate and a corresponding second confidence level for the vehicle based on the dynamic signal includes:

[0030] The adhesion coefficient and slip ratio are calculated based on the aforementioned dynamic signals;

[0031] The second estimated value of the adhesion coefficient is calculated based on the mapping relationship between the adhesion coefficient, the slip ratio, and the road surface adhesion coefficient.

[0032] The slip rate observability index is calculated based on the slip rate, and the second confidence level is determined based on the slip rate observability index.

[0033] Secondly, this application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in the first aspect above.

[0034] Thirdly, this application also provides a vehicle, including: a body body and a control system installed in the body body, the control system including a first sensing module, a second sensing module and a control module, the first sensing module and the second sensing module being respectively connected to the control module;

[0035] The first sensing module is used to collect road surface data;

[0036] The second sensing module is used to collect the vehicle's dynamic signals;

[0037] The control module is used to execute the steps of the method described in the first aspect above.

[0038] The aforementioned method for determining the road surface adhesion coefficient, along with the electronic device and vehicle, collects road surface data and dynamic signals, respectively. Based on the road surface data, a first estimated value and a first confidence level of the adhesion coefficient are determined. Based on the dynamic signals, a second estimated value and a second confidence level of the adhesion coefficient are determined. Using the two-dimensional confidence levels, the two-dimensional adhesion coefficient estimates are fused to obtain the road surface adhesion coefficient. This extends the traditional single numerical estimation to a two-dimensional output of "estimated value + confidence interval," improving the accuracy of road surface adhesion coefficient assessment. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the structure of the control system in a vehicle in one embodiment;

[0040] Figure 2 This is a flowchart illustrating a method for determining the road surface adhesion coefficient in one embodiment;

[0041] Figure 3 This is a flowchart illustrating a database maintenance method in one embodiment;

[0042] Figure 4 This is a flowchart illustrating a database update method in one embodiment;

[0043] Figure 5 This is a schematic diagram illustrating the principle of a method for determining the road surface adhesion coefficient in a control system embodiment.

[0044] Figure 6 This is a schematic diagram of the structure of an electronic device in one embodiment. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] Unless otherwise defined, the technical or scientific terms used in this application shall have the general meaning understood by one of ordinary skill in the art to which this application pertains. Words such as “a,” “an,” “an,” “the,” “the,” and “these” used in this application do not indicate quantitative limitation and may be singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include steps or modules (units) not listed, or may include other steps or modules (units) inherent to these processes, methods, products, or devices. Words such as “connected,” “linked,” and “coupled” used in this application are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. Normally, the character " / " indicates that the objects before and after it are in an "or" relationship. The terms "first," "second," "third," etc., used in this application are merely to distinguish similar objects and do not represent a specific order of objects.

[0047] In one embodiment, Figure 1 A schematic diagram of the internal structure of a vehicle is provided, such as... Figure 1 As shown, the vehicle includes a body body and a control system installed in the body body. The control system includes a first sensing module 101, a second sensing module 102 and a control module 103. The first sensing module 101 and the second sensing module 102 are respectively connected to the control module 103.

[0048] The first sensing module 101 is used to collect road surface data. The first sensing module 101 includes a camera and a millimeter-wave radar.

[0049] In some embodiments, the first sensing module 101 includes a camera, which can be used to acquire road surface images, perform semantic segmentation based on the road surface images, and predict the road surface type and the corresponding first confidence level.

[0050] In some embodiments, the first sensing module 101 includes a camera and a millimeter-wave radar, i.e., a millimeter-wave radar is used to replace part of the camera. For example, a 77GHz millimeter-wave radar is used to collect point cloud data to assist semantic segmentation. Specifically, point cloud reflection intensity features are extracted from the point cloud data to distinguish the difference in dielectric constant between water surface and ice layer, and a road surface semantic segmentation map is generated by combining it with a clustering algorithm. The potential low-adhesion region mask generated by the radar is then input into a lightweight UNet network to refine the classification results. At this point, the confidence level conf_radar can be evaluated based on the Doppler spectrum stability.

[0051] Feasibility verification: MIT research shows that radar is more accurate than cameras in detecting slippery surfaces in rainy weather, and has a stronger ability to penetrate rain and fog; with dynamic weighted fusion, a self-learning closed loop can still be achieved. After feedback and iterative optimization, the radar's initial classification results can provide early warning of low-adhesion areas in heavy rain conditions compared to pure vision solutions.

[0052] In some embodiments, the first sensing module 101 includes a microphone array. Road surface type is identified using the spectral characteristics of tire rolling noise. Specifically, tire noise signals can be captured by installing an external microphone array, and an LSTM network can be trained on the Mel spectrum to output class probabilities; simultaneously, the degree of environmental interference is reflected by quantifying the amplitude of sound pressure level fluctuations. Initial lookup values ​​are obtained by mapping from a voiceprint feature library.

[0053] Feasibility verification: Field tests show that dry / wet asphalt tire noise fundamental frequency offset >3dB can be distinguished; the cost of this solution is lower than that of vision systems and is not affected by lighting conditions. It has unique advantages in scenarios without GPS, such as tunnels. After calibration by the feedback mechanism, the recognition accuracy for water depth >3mm is high, meeting the same safety requirements.

[0054] The second sensing module 102 is used to acquire the vehicle's dynamic signals. The second sensing module 102 includes an inertial measurement unit and a wheel speed sensor. The dynamic signals may specifically include wheel speed, longitudinal / lateral acceleration, and yaw rate.

[0055] Control module 103 is used to execute the method for determining the road surface adhesion coefficient. Control module 103 includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the method for determining the road surface adhesion coefficient. In some embodiments, control module 103 internally installs software for executing the method for determining the road surface adhesion coefficient, and this software can be updated via over-the-air (OTA) technology.

[0056] Figure 2 A flowchart illustrating the method for determining the road surface adhesion coefficient in this embodiment is provided, including the following steps:

[0057] Step S101: Acquire road surface data and vehicle dynamic signals collected by the vehicle.

[0058] Road surface data can be road surface images (e.g., RGB images), road surface images combined with radar point cloud data, or tire noise signals. The camera can be mounted above the vehicle's windshield, covering a certain area of ​​the road surface ahead, ensuring the capture of road surface features under different conditions, such as wet / slippery conditions, snow accumulation, ice, asphalt, and gravel. The acquired road surface images may be affected by factors such as changes in lighting, shadows, and rain / fog interference.

[0059] The dynamic signals include wheel speed, longitudinal / lateral acceleration, and yaw rate.

[0060] Step S102: Based on the road surface data, determine the road surface type and the corresponding first confidence level, and determine the first adhesion coefficient estimate related to the road surface type; based on the dynamic signal, determine the second adhesion coefficient estimate of the vehicle and the corresponding second confidence level.

[0061] The first adhesion coefficient estimate represents a preliminary estimate of the maximum usable friction coefficient (i.e., adhesion coefficient) of the road surface, and the first confidence level represents the reliability of the road surface type identified based on road surface data. The second adhesion coefficient estimate represents the actual usable adhesion coefficient of the current road surface derived from vehicle dynamics, and the second confidence level represents the reliability of the adhesion coefficient estimate derived from vehicle dynamic response.

[0062] As one possible implementation, road surface data can be input into a neural network model for processing. The main branch of the neural network model can be used to output the road surface type, and the auxiliary branch of the neural network model can be used to output the first confidence level. The first adhesion coefficient estimate related to the road surface type can be queried from the database.

[0063] Specifically, the neural network model can be a lightweight MobileNet structure, with the input being an RGB image after dehazing and enhancement. The output layer is expanded into a two-branch structure, with the main branch outputting discrete road surface categories (e.g., dry asphalt, wet cement, ice and snow, gravel), and the auxiliary branch outputting a scalar value in the interval [0, 1] as the first confidence level. The first confidence level can be calculated by combining image sharpness scoring (based on gradient magnitude variance) and temporal consistency verification (fluctuation threshold of continuous frame classification results), i.e., based on the semantic category probability distribution of the current road surface output by the model. Then, the estimated value of the first adhesion coefficient related to the road surface type is determined by looking up a table in a pre-configured database.

[0064] (2) As an implementation method, the adhesion coefficient and slip ratio can be calculated based on the dynamic signal; the second adhesion coefficient estimate can be calculated based on the mapping relationship between the adhesion coefficient, slip ratio, slip ratio and road adhesion coefficient; the slip ratio observability index can be calculated based on the slip ratio; and the second confidence level can be determined based on the slip ratio observability index.

[0065] Specifically, during acceleration / braking, tire forces are calculated based on the vehicle's mechanical model, and the coefficient of friction is obtained from these tire forces. After filtering the IMU and wheel speed sensor signals, a reference vehicle speed is calculated, and the slip ratio is calculated based on this reference speed. The mapping relationship between the slip ratio and the road surface adhesion coefficient is obtained, and the currently calculated coefficient of friction and slip ratio are substituted into this mapping relationship to calculate a second estimated value of the coefficient of friction. The mapping relationship between the slip ratio and the road surface adhesion coefficient can be obtained through calibration under constant speed driving conditions.

[0066] The slip ratio observability index is the differential of the slip ratio with respect to time: |dλ / dt|, where λ represents the slip ratio and t represents time. For example, if the slip ratio observability index is greater than or equal to a preset threshold, the second confidence level is set to 0.8~1.0. Conversely, if the slip ratio observability index is less than the preset threshold, the second confidence level is set to 0.2~0.3.

[0067] Step S103: Based on the first confidence level and the second confidence level, the first adhesion coefficient estimate and the second adhesion coefficient estimate are fused to obtain the target road surface adhesion coefficient.

[0068] As one possible implementation, the first confidence level and the second confidence level can be normalized, the first weight and the second weight can be determined respectively, and then the first weight, the second weight and the corresponding attachment coefficient estimate can be weighted and summed.

[0069] As another possible implementation, a first weight and a second weight can be determined based on a first confidence level and a second confidence level, respectively; wherein the first weight is proportional to the square of the first confidence level, and the second weight is proportional to the square of the second confidence level; based on the first weight and the second weight, the estimated values ​​of the first adhesion coefficient and the estimated values ​​of the second adhesion coefficient are nonlinearly weighted and fused to obtain the target road surface adhesion coefficient.

[0070] Specifically, the target pavement adhesion coefficient is calculated using a nonlinear weighting function:

[0071] μ_opt = [f(conf_vis) × μ_vis + f(conf_dyn) × μ_dyn] / [f(conf_vis) + f(conf_dyn)], where f(x) = x^k (k>1);

[0072] f(conf_vis)=conf_vis² / Σ(conf_vis²+conf_dyn²);

[0073] f(conf_dyn)=conf_vis² / Σ(conf_vis²+conf_dyn²);

[0074] Where μ_opt represents the target road surface adhesion coefficient, conf_vis represents the first confidence level, μ_vis represents the first adhesion coefficient estimate, f(conf_vis) represents the first weight, conf_dyn represents the second confidence level, μ_dyn represents the second adhesion coefficient estimate, and f(conf_dyn) represents the second weight.

[0075] In this embodiment, nonlinear weighted fusion ensures that the contribution of high-confidence sources is amplified quadratically, thus enhancing the dominance of the high-confidence interval. For example, when conf_vis=0.2 and conf_dyn=0.8 in rainy or foggy weather, k=2 makes the weight of the dynamic subsystem reach 94%, avoiding visual noise contamination of the results. By breaking through the fixed weight limitation and dynamically allocating the contribution ratio according to real-time confidence, the fusion error is reduced under extreme conditions.

[0076] In steps S101 to S103 above, road surface data and dynamic signals are collected respectively. Based on the road surface data, a first estimated value and a first confidence level of the adhesion coefficient are determined. Based on the dynamic signals, a second estimated value and a second confidence level of the adhesion coefficient are determined. Based on the two-dimensional confidence levels, the two-dimensional adhesion coefficient estimates are fused to obtain the road surface adhesion coefficient. This expands the traditional single numerical estimation to a two-dimensional output of "estimated value + confidence interval", improving the accuracy of road surface adhesion coefficient assessment.

[0077] In some embodiments, Figure 3 A flowchart illustrating a database maintenance method is provided, such as... Figure 3 As shown, after obtaining the target road surface adhesion coefficient by fusing the first adhesion coefficient estimate and the second adhesion coefficient estimate based on the first confidence level and the second confidence level, the method further includes:

[0078] Step S201: Group and cache the target road surface adhesion coefficients according to the corresponding road surface types.

[0079] For example, independent queues can be established according to the identified road surface type (such as "wet asphalt", "compacted snow", "gravel road", etc.). Each queue stores the most recent valid sample and records the cumulative number of samples. This setup allows the system to independently calculate the distribution characteristics (such as mean, variance, and confidence interval) of the adhesion coefficient for each type of road surface, avoiding estimation bias caused by the mixing of data from different road surface types. It also facilitates the construction or updating of adhesion coefficients associated with each road surface type.

[0080] To construct a dynamic credibility-driven closed-loop self-learning framework, in some embodiments, after grouping and caching the target road surface adhesion coefficients according to their corresponding road surface types, database maintenance also includes:

[0081] Step S202: Determine whether there is a first road surface type in the cached data whose cumulative number of samples exceeds a preset threshold. If the determination is yes, proceed to step S203; if the determination is no, return to step S202.

[0082] Step S203: Update the estimated first adhesion coefficient in the database related to the first road surface type.

[0083] For example, when the cumulative number of samples for a certain road surface type exceeds a preset threshold, a database update process is triggered. Based on the cumulative number of samples for the first road surface type, the mean of the adhesion coefficients of multiple target road surfaces is calculated to obtain the average road surface adhesion coefficient. The original first adhesion coefficient estimates related to the first road surface type in the database are then merged with the average road surface adhesion coefficient to obtain the updated first adhesion coefficient estimate.

[0084] Specifically, Figure 4 A flowchart illustrating the database update method is provided, such as... Figure 4 As shown, the process includes the following steps:

[0085] Step S301: Determine the attenuation factor based on the cumulative number of samples of the first road surface type and the preset time constant.

[0086] The calculation formula is: β=exp(-N / τ). β represents the attenuation factor (0<β<1), N represents the cumulative number of samples of the first road surface type, τ represents the preset time constant, and exp represents the exponential function.

[0087] Step S302: Based on the attenuation factor, determine the weight of the original first adhesion coefficient estimate and the weight of the mean road adhesion coefficient.

[0088] The weight of the original first adhesion coefficient estimate is β, and the weight of the mean road adhesion coefficient is (1−β).

[0089] Step S303: Based on the weights of the original first adhesion coefficient estimate and the average road surface adhesion coefficient, perform a weighted summation on the original first adhesion coefficient estimate and the average road surface adhesion coefficient to obtain the updated first adhesion coefficient estimate.

[0090] The calculation formula is as follows:

[0091] table_new(type)=β×μ_table_old(type)+(1−β)×mean(μ_opt)_type;

[0092] Where table_new(type) represents the database, which stores the correlation between road surface type and the first adhesion coefficient estimate, μ_table_old(type) represents the original first adhesion coefficient estimate, and mean(μ_opt)_type represents the mean of the road surface adhesion coefficient.

[0093] For example, after the system repeatedly identifies the "snow-covered" road surface type and outputs a stable μ_opt≈0.2, it automatically corrects the original calibration value of 0.35 to a more accurate level.

[0094] This embodiment constructs a closed-loop self-learning framework driven by dynamic credibility. By establishing a continuous learning loop, the database is continuously optimized based on historical driving data. After accumulating a certain mileage, it can improve the accuracy of identifying unknown road surface types, eliminate long-term biases caused by static databases, and ensure rapid convergence in high-frequency scenarios.

[0095] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0096] In one embodiment, Figure 5 A schematic diagram of a method for determining the road surface adhesion coefficient in a control system is provided. The control system sets up a visual path and a dynamic path respectively to comprehensively determine the road surface adhesion coefficient.

[0097] For the visual path, a forward-looking camera is used to acquire road surface images. These images are preprocessed, for example, by performing dehazing and enhancement. The preprocessed images are then input into a lightweight neural network, MobileNet, for prediction, yielding the semantic category probability distribution of the current road surface. This distribution includes, but is not limited to, types such as dry asphalt, slippery asphalt, ice and snow, gravel, and mud. Based on this semantic category probability distribution, the road surface type and a first confidence score (conf-vis) are obtained. Finally, the first adhesion coefficient estimate (u-vis) corresponding to the current road surface type is retrieved from the database.

[0098] For the dynamic path, wheel speed / IMU sensors are used to collect the vehicle's dynamic signals. The dynamic signals are processed by a dynamic solver to output the second adhesion coefficient estimate u-dyn and the second confidence level conf-dyn.

[0099] The adaptive fusion unit combines the first adhesion coefficient estimate u-vis and the second adhesion coefficient estimate u-dyn based on the first confidence level conf-vis and the second confidence level conf-dyn to obtain the target road surface adhesion coefficient u-opt.

[0100] The target road surface adhesion coefficient u-opt is sent back to the database cache to determine whether there is a first road surface type in the cache data whose cumulative number of samples exceeds a preset threshold. If it is determined to be yes, the estimated value of the first adhesion coefficient related to the first road surface type in the database is updated to form a closed-loop self-learning feedback loop.

[0101] This embodiment constructs a closed-loop self-learning framework driven by dynamic credibility:

[0102] By setting up a dual-dimensional reliability quantification mechanism, not only are the adhesion coefficient estimates (μ-vis, μ-dyn) of each subsystem output, but their dynamic reliability indices (conf-vis, conf-dyn) are also generated simultaneously. This expands the traditional single numerical estimation into a dual-dimensional output of "estimated value + confidence interval", thereby improving the accuracy of road adhesion coefficient assessment.

[0103] By employing a credibility-weighted adaptive fusion algorithm, a non-linear weighting function is used to overcome the limitation of fixed weights and dynamically allocate contribution ratios based on real-time credibility, thereby reducing fusion errors under extreme conditions.

[0104] By using closed-loop feedback-driven online table updates, a continuous learning cycle is established. Historical driving data continuously optimizes the table lookup benchmark database, improving the accuracy of identifying unknown road surface types after accumulating a certain distance, thus eliminating long-term biases caused by static databases.

[0105] In one embodiment, Figure 6 A schematic diagram of the structure of an electronic device is also provided, such as Figure 6As shown, the electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores sensor data and results calculated based on the sensor data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a method for determining the road surface adhesion coefficient.

[0106] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0107] Acquire road surface data and vehicle dynamics signals collected by the vehicle;

[0108] Based on road surface data, determine the road surface type and its corresponding first confidence level, as well as the first adhesion coefficient estimate related to the road surface type; based on dynamic signals, determine the second adhesion coefficient estimate of the vehicle and its corresponding second confidence level.

[0109] The target road surface adhesion coefficient is obtained by fusing the first and second adhesion coefficient estimates based on the first and second confidence levels.

[0110] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0111] Based on the first confidence level and the second confidence level, the first adhesion coefficient estimate and the second adhesion coefficient estimate are fused to obtain the target road surface adhesion coefficient, including:

[0112] Based on the first confidence level and the second confidence level, the first weight and the second weight are determined respectively; wherein, the first weight is proportional to the square of the first confidence level, and the second weight is proportional to the square of the second confidence level;

[0113] Based on the first weight and the second weight, the first adhesion coefficient estimate and the second adhesion coefficient estimate are nonlinearly weighted and fused to obtain the target road surface adhesion coefficient.

[0114] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0115] After obtaining the target road surface adhesion coefficient by fusing the first adhesion coefficient estimate and the second adhesion coefficient estimate based on the first confidence level and the second confidence level, the method further includes:

[0116] The target road surface adhesion coefficient is grouped and cached according to the corresponding road surface type.

[0117] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0118] After grouping and caching the target pavement adhesion coefficient according to the corresponding pavement type, the method also includes:

[0119] Determine whether the cached data contains a first road surface type whose cumulative sample count exceeds a preset threshold;

[0120] If the determination is yes, then update the estimated value of the first adhesion coefficient related to the first road surface type in the database.

[0121] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0122] Update the database with the first adhesion coefficient estimate related to the first road surface type, including:

[0123] Based on the cumulative number of samples of the first road surface type, the mean of the adhesion coefficients of multiple target road surfaces is calculated to obtain the mean of the road surface adhesion coefficients.

[0124] The original first adhesion coefficient estimate related to the first road surface type in the database is fused with the mean road surface adhesion coefficient to obtain the updated first adhesion coefficient estimate.

[0125] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0126] The original first adhesion coefficient estimates related to the first road surface type in the database are merged with the mean road surface adhesion coefficient to obtain updated first adhesion coefficient estimates, including:

[0127] The attenuation factor is determined based on the cumulative number of samples of the first road surface type and the preset time constant.

[0128] Based on the attenuation factor, the weights of the original first adhesion coefficient estimate and the weights of the mean road adhesion coefficient are determined respectively.

[0129] Based on the weights of the original first adhesion coefficient estimate and the mean road adhesion coefficient, the original first adhesion coefficient estimate and the mean road adhesion coefficient are weighted and summed to obtain the updated first adhesion coefficient estimate.

[0130] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0131] Based on road surface data, determine the road surface type and its corresponding first confidence level, as well as the estimated first adhesion coefficient related to the road surface type, including:

[0132] The road surface data is input into a neural network model for processing.

[0133] The main branch of the neural network model outputs the road surface type, and the auxiliary branch of the neural network model outputs the first confidence level.

[0134] Retrieve the first adhesion coefficient estimate related to the road surface type from the database.

[0135] In one embodiment, when a processor executes a computer program, it performs the following steps:

[0136] Based on the dynamic signals, determine the estimated second adhesion coefficient of the vehicle and the corresponding second confidence level, including:

[0137] The adhesion coefficient and slip ratio are calculated based on the dynamic signals.

[0138] The second estimated value of the adhesion coefficient is calculated based on the mapping relationship between the adhesion coefficient, the slip ratio, and the road surface adhesion coefficient.

[0139] The slip rate observability index is calculated based on the slip rate, and the second confidence level is determined based on the slip rate observability index.

[0140] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated in this embodiment.

[0141] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0142] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining the road surface adhesion coefficient, characterized in that, include: Acquire road surface data and vehicle dynamics signals collected by the vehicle; Based on the road surface data, determine the road surface type and its corresponding first confidence level, and determine a first adhesion coefficient estimate related to the road surface type; based on the dynamic signal, determine a second adhesion coefficient estimate for the vehicle and its corresponding second confidence level. Based on the first confidence level and the second confidence level, the first adhesion coefficient estimate and the second adhesion coefficient estimate are fused to obtain the target road surface adhesion coefficient.

2. The method for determining the road surface adhesion coefficient according to claim 1, characterized in that, Based on the first confidence level and the second confidence level, the first adhesion coefficient estimate and the second adhesion coefficient estimate are fused to obtain the target road surface adhesion coefficient, including: Based on the first confidence level and the second confidence level, a first weight and a second weight are determined respectively; wherein, the first weight is proportional to the square of the first confidence level, and the second weight is proportional to the square of the second confidence level; Based on the first weight and the second weight, the first adhesion coefficient estimate and the second adhesion coefficient estimate are nonlinearly weighted and fused to obtain the target road surface adhesion coefficient.

3. The method for determining the road surface adhesion coefficient according to claim 1, characterized in that, After obtaining the target road surface adhesion coefficient by fusing the first adhesion coefficient estimate and the second adhesion coefficient estimate based on the first confidence level and the second confidence level, the method further includes: The target road surface adhesion coefficients are grouped and cached according to their corresponding road surface types.

4. The method for determining the road surface adhesion coefficient according to claim 3, characterized in that, After grouping and caching the target road surface adhesion coefficients according to their corresponding road surface types, the method further includes: Determine whether the cached data contains a first road surface type whose cumulative sample count exceeds a preset threshold; If the determination is yes, then update the first adhesion coefficient estimate in the database related to the first road surface type.

5. The method for determining the road surface adhesion coefficient according to claim 4, characterized in that, Update the database with the first adhesion coefficient estimate associated with the first road surface type, including: Based on the cumulative number of samples of the first road surface type, the mean of the adhesion coefficients of multiple target road surfaces is calculated to obtain the mean of the road surface adhesion coefficients. The original first adhesion coefficient estimate related to the first road surface type in the database is fused with the mean road surface adhesion coefficient to obtain an updated first adhesion coefficient estimate.

6. The method for determining the road surface adhesion coefficient according to claim 5, characterized in that, The original first adhesion coefficient estimate related to the first road surface type in the database is fused with the mean road surface adhesion coefficient to obtain an updated first adhesion coefficient estimate, including: The attenuation factor is determined based on the cumulative number of samples of the first road surface type and the preset time constant. Based on the attenuation factor, the weights of the original first adhesion coefficient estimate and the weights of the mean road adhesion coefficient are determined respectively. Based on the weights of the original first adhesion coefficient estimate and the weights of the mean road adhesion coefficient, the original first adhesion coefficient estimate and the mean road adhesion coefficient are weighted and summed to obtain the updated first adhesion coefficient estimate.

7. The method for determining the road surface adhesion coefficient according to claim 1, characterized in that, Based on the road surface data, determine the road surface type and its corresponding first confidence level, and determine a first adhesion coefficient estimate related to the road surface type, including: The road surface data is input into a neural network model for processing. The road surface type is output according to the main branch of the neural network model, and the first confidence level is output according to the auxiliary branch of the neural network model. Retrieve the first adhesion coefficient estimate associated with the road surface type from the database.

8. The method for determining the road surface adhesion coefficient according to claim 1, characterized in that, Based on the dynamic signal, determining the estimated second adhesion coefficient and the corresponding second confidence level of the vehicle includes: The adhesion coefficient and slip ratio are calculated based on the aforementioned dynamic signals; The second estimated value of the adhesion coefficient is calculated based on the mapping relationship between the adhesion coefficient, the slip ratio, and the road surface adhesion coefficient. The slip rate observability index is calculated based on the slip rate, and the second confidence level is determined based on the slip rate observability index.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

10. A vehicle, characterized in that, include: The vehicle body and the control system installed in the vehicle body, the control system including a first sensing module, a second sensing module and a control module, the first sensing module and the second sensing module being respectively connected to the control module; The first sensing module is used to collect road surface data; The second sensing module is used to collect the vehicle's dynamic signals; The control module is used to perform the steps of the method according to any one of claims 1 to 8.

Images