Vehicle control parameter determination method, system, equipment and medium

Abstract

The invention discloses a vehicle control parameter determination method, system and device and a medium, and relates to the technical field of vehicle automatic control, the method comprises the following steps: obtaining an environment feature vector of a vehicle and a biological feature vector of a driver, the environment feature vector comprising environment feature sub-vectors of multiple dimensions, the biological feature vector comprises a plurality of dimensions of biological feature sub-vectors; acquiring an environment feature weight corresponding to the environment feature sub-vector of each dimension, and combining the environment feature weights to obtain an environment feature matrix corresponding to the environment feature vector; obtaining a biological feature weight corresponding to the biological feature sub-vector of each dimension, and combining the biological feature weights to obtain a biological feature matrix corresponding to the biological feature vector; acquiring a scene coefficient corresponding to a current driving scene of the vehicle, and performing weighted fusion according to the scene coefficient and the data to obtain a fusion decision vector; and determining a plurality of control parameters of the vehicle in the current driving scene according to the fusion decision vector.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle automatic control, and in particular to a vehicle control parameter determination method, system, device and medium. BACKGROUND

[0002] At present, with the continuous improvement of the intelligent level of automobiles, intelligent driving systems have been widely applied in vehicle control, such as realizing adaptive cruise, lane keeping and other functions. However, the current mainstream intelligent driving system adopts pre-set fixed parameters and control strategies, and its decision logic cannot be differentiated according to different driver individuals. This leads to a single driving experience provided by the system, which cannot meet the personalized needs of users for driving style. Statistics show that a considerable proportion of users think that the existing system is too conservative or does not conform to their driving habits, and this mismatch reduces the user's acceptance and usage stickiness of high-order intelligent driving functions.

[0003] However, in order to improve this situation, some adaptive schemes are proposed in the prior art. For example, some schemes divide driving styles by analyzing time slices of historical driving data, and adjust the driving mode accordingly, but such methods only rely on vehicle operation data and cannot real-time perceive the physiological and emotional state of the driver. SUMMARY

[0004] A series of simplified concepts are introduced in the summary section, which will be further described in detail in the specific embodiment section. The summary section of the present application does not mean to attempt to limit the key features and necessary technical features of the claimed technical solutions, nor to attempt to determine the protection scope of the claimed technical solutions.

[0005] In a first aspect, an embodiment of the present application provides a vehicle control parameter determination method, which comprises: obtaining an environmental feature vector of a vehicle and a biological feature vector of a driver, the environmental feature vector comprising a plurality of dimensional environmental feature sub-vectors, and the biological feature vector comprising a plurality of dimensional biological feature sub-vectors; obtaining an environmental feature weight corresponding to each dimensional environmental feature sub-vector, combining each environmental feature weight to obtain an environmental feature matrix corresponding to the environmental feature vector; obtaining a biological feature weight corresponding to each dimensional biological feature sub-vector, combining each biological feature weight to obtain a biological feature matrix corresponding to the biological feature vector; obtain a scene coefficient corresponding to a current driving scene of the vehicle, and perform environment-biological dual feature matrix weighted fusion according to the scene coefficient, the environment feature matrix, the environment feature vector, the biological feature matrix and the biological feature vector to obtain a fusion decision vector; determine a plurality of control parameters of the vehicle in the current driving scene according to the fusion decision vector.

[0006] In an embodiment of the present application, the environment feature weight corresponding to the environment feature sub-vector of each dimension is obtained by the following formula: ; wherein, is the environment feature weight corresponding to the environment feature sub-vector of the i-th dimension, is the hyperparameter of the adjustment slope of the environment feature sub-vector of the i-th dimension, is the historical mean value of the environment feature sub-vector of the i-th dimension.

[0007] In an embodiment of the present application, the biological feature weight corresponding to the biological feature sub-vector of each dimension is obtained by the following formula: ; wherein, is the biological feature weight corresponding to the biological feature sub-vector of the j-th dimension, is a preset weight coefficient, is the historical minimum value of the biological feature sub-vector of the j-th dimension, is the historical maximum value of the biological feature sub-vector of the j-th dimension.

[0008] In an embodiment of the present application, after obtaining the environment feature vector of the vehicle and the biological feature vector of the driver, the method comprises: determining a to-be-corrected environment feature sub-vector in the environment feature vector based on the current driving scene of the vehicle and the environment feature vector, and obtaining an environment correction coefficient corresponding to the to-be-corrected environment feature sub-vector in the current driving scene; correcting the to-be-corrected environment feature sub-vector based on the environment correction coefficient to obtain a corrected environment feature sub-vector; determining a to-be-corrected biological feature sub-vector in the biological feature vector based on the current driving scene of the vehicle and the biological feature vector, and obtaining a biological correction coefficient corresponding to the to-be-corrected biological feature sub-vector in the current driving scene; correcting the to-be-corrected biological feature sub-vector based on the biological correction coefficient to obtain a corrected biological feature sub-vector.

[0009] In an embodiment of the present application, the scene coefficient, the environment feature matrix, the environment feature vector, the biological feature matrix and the biological feature vector are fused by environment-biology dual feature matrix weighting to obtain a fusion decision vector, which is obtained by the following formula: ; Wherein, is the fusion decision vector, is the scene coefficient, is the environment feature matrix, is the environment feature vector, is the biological feature matrix, is the biological feature vector.

[0010] In an embodiment of the present application, the plurality of control parameters of the vehicle in the current driving scene are determined according to the fusion decision vector, which is obtained by the following formula: ; Wherein, is the i-th control parameter in the plurality of control parameters, is the basic control parameter of the i-th control parameter in the standard mode, is the maximum adjustment amount of the i-th control parameter, is the fusion decision vector in the current driving scene, is the feature sensitive vector of parameter i, is an activation function.

[0011] In an embodiment of the present application, the environment feature vector includes at least two of the road adhesion coefficient, the light intensity, the visibility level, the traffic density, the speed limit change rate, the curve curvature, the slope value and the noise decibel value; and / or The biological feature vector includes at least two of the heart rate variability, the blink frequency, the mouth corner angle, the voice emotion entropy and the steering wheel operation entropy; and / or The plurality of control parameters include at least two of the following: following distance, lane changing safety distance, curve deceleration rate, starting acceleration, steering angular velocity, overtaking decision threshold, lane centering accuracy, accelerator pedal sensitivity, brake pedal nonlinearity, voice command response delay, automatic parking into the garage accuracy and energy recovery intensity.

[0012] In a second aspect, the present application provides a vehicle control parameter determination system, the system comprising: a weight determination module, a weight fusion module and a parameter adjustment module. The weight determination module is configured to: obtain the environmental feature vector of the vehicle and the biometric vector of the driver, wherein the environmental feature vector includes environmental feature sub-vectors of multiple dimensions and the biometric vector includes biometric sub-vectors of multiple dimensions; obtain the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension; and combine the environmental feature weights to obtain the environmental feature matrix corresponding to the environmental feature vectors. The weight fusion module is configured to: obtain the biofeature weights corresponding to the biofeature sub-vectors of each dimension, combine the biofeature weights to obtain the biofeature matrix corresponding to the biofeature vector; obtain the scene coefficients corresponding to the current driving scene of the vehicle, and perform weighted fusion of the environmental and biological dual feature matrices based on the scene coefficients, the environmental feature matrix, the environmental feature vector, the biofeature matrix, and the biofeature vector to obtain the fusion decision vector; The parameter adjustment module is configured to determine multiple control parameters of the vehicle in the current driving scenario based on the fusion decision vector.

[0013] Thirdly, an electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of a vehicle control parameter determination method as described in any of the first aspects above.

[0014] Fourthly, this application also proposes a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of a vehicle control parameter determination method according to any one of the first aspects.

[0015] In summary, the vehicle control parameter determination method of this application constructs multi-dimensional environmental feature sub-vectors and multi-dimensional biological feature sub-vectors, calculates the weights corresponding to each dimension to form an environmental feature matrix and a biological feature matrix, and then combines the scene coefficients determined by the current driving scenario to perform weighted fusion of the two feature matrices and their original vectors to generate a fused decision vector. Finally, based on the fused decision vector, multiple control parameters of the vehicle in the current scenario are simultaneously determined. Because the driving style adjustment integrates multi-dimensional environmental feature sub-vectors and the driver's multi-dimensional biological feature sub-vectors, and dynamically calibrates and integrates the environmental feature matrix, environmental feature vectors, biological feature matrix, and biological feature vectors through weighted fusion of scene coefficients and the environmental and biological feature matrices, it overcomes the problems of perception bias and decision rigidity caused by traditional solutions that rely solely on single-dimensional data or fixed rules. This enables the vehicle's control system to more comprehensively and accurately understand complex driving scenarios and output highly matched control parameters, ultimately significantly improving the accuracy of adaptive driving style adjustment while ensuring safety.

[0016] The method for determining vehicle control parameters proposed in this application, along with other advantages, objectives, and features of this application, will be partly apparent from the following description and partly understood by those skilled in the art through study and practice of this application. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating a method for determining vehicle control parameters provided in an embodiment of this application. Figure 2 This application provides a schematic diagram of a vehicle control parameter determination system. Figure 3 This is a schematic diagram of an electronic device for determining vehicle control parameters provided in an embodiment of this application. Detailed Implementation

[0018] To better understand the technical solutions provided in the embodiments of this specification, the technical solutions of the embodiments of this specification will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0019] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The term "two or more" includes two or more cases.

[0020] Please see Figure 1 This is a flowchart illustrating a method for determining vehicle control parameters provided in an embodiment of this application, specifically including: S110. Obtain the environmental feature vector of the vehicle and the biometric vector of the driver. The environmental feature vector includes environmental feature sub-vectors of multiple dimensions, and the biometric vector includes biometric sub-vectors of multiple dimensions. For example, an environmental perception sensor array onboard the vehicle, such as cameras, radar, or high-precision maps, can collect external road conditions and traffic information in real time. The collected external road conditions and traffic information can be quantified into an environmental feature vector containing multiple independent dimensions. The environmental feature vector includes multiple environmental feature sub-vectors, each corresponding to a specific environmental quantification index, such as road conditions, lighting, or traffic density. Simultaneously, the driver's physiological and behavioral signals can be collected through in-vehicle monitoring devices facing the driver, such as cameras, microphones, or physiological sensors. The collected driver's physiological and behavioral signals can be quantified into a biometric vector containing multiple independent dimensions. The biometric vector includes multiple biometric sub-vectors, each corresponding to a specific driver state index, such as physiological characteristics, facial expressions, or vocal emotions.

[0021] S120. Obtain the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension, and combine the environmental feature weights to obtain the environmental feature matrix corresponding to the environmental feature vectors. For example, based on the real-time physical quantity represented by each environmental feature sub-vector, such as the current road surface adhesion coefficient or light intensity, and combined with a normalization function—that is, sigmoid normalization—each environmental feature sub-vector is obtained to achieve the environmental feature weight corresponding to each dimension. Subsequently, these environmental feature weights, calculated independently for all dimensions, are arranged and combined according to their corresponding feature dimension order to construct a normalized matrix, which is the environmental feature matrix. This process integrates multiple discrete weight values ​​into a unified mathematical structure, aiming to provide a data representation foundation for subsequent fusion calculations that simultaneously reflects the weights of each environmental dimension and their mutual independence.

[0022] S130. Obtain the biofeature weights corresponding to the biofeature sub-vectors of each dimension, and combine the biofeature weights to obtain the biofeature matrix corresponding to the biofeature vectors. For example, for each dimension of biometric subvector, the corresponding biometric weight is calculated by linear normalization and combined with preset weight coefficients. Then, these biometric weights are integrated into a normalized matrix form according to their dimensional order, thereby obtaining a biometric matrix that represents the importance and independence of each biometric dimension.

[0023] S140. Obtain the scene coefficients corresponding to the current driving scene of the vehicle, and perform weighted fusion of the environmental and biological dual feature matrices based on the scene coefficients, the environmental feature matrix, the environmental feature vector, the biological feature matrix, and the biological feature vector to obtain the fusion decision vector. For example, after obtaining the standardized environmental feature matrix and biometric feature matrix, the current driving scenario of the vehicle is determined based on real-time perception data, such as highway cruising, congested following, or emergency obstacle avoidance, and a preset scenario coefficient is obtained for this current driving scenario. The scenario coefficient defines the dynamic contribution ratio of each environmental feature vector and each biometric feature vector in the final decision. Subsequently, based on the scenario coefficient, the weighted environmental feature vectors of the environmental feature matrix and the weighted biometric feature vectors of the biometric feature matrix are linearly fused to output a fused decision vector that can comprehensively and balancedly represent the current comprehensive driving situation, providing a unified decision basis for the accurate calculation of subsequent control parameters.

[0024] S150. Determine multiple control parameters of the vehicle in the current driving scenario based on the fusion decision vector.

[0025] For example, based on the fused decision vector, multiple control parameters of the vehicle in the current driving scenario are determined. By performing operations on the fused decision vector and the pre-calibrated feature sensitivity vectors of each control parameter, the abstract decision information representing the comprehensive driving situation is decoded into a set of specific control parameter values ​​that work together, thereby completing the transformation from situation perception to final control command.

[0026] In summary, the vehicle control parameter determination method proposed in this application constructs multi-dimensional environmental feature sub-vectors and multi-dimensional biological feature sub-vectors, calculates the weights corresponding to each dimension to form an environmental feature matrix and a biological feature matrix, and then combines the scenario coefficients determined by the current driving scenario to perform weighted fusion of the two feature matrices and their original vectors to generate a fused decision vector. Finally, based on the fused decision vector, multiple control parameters of the vehicle in the current scenario are simultaneously determined. Because the driving style adjustment integrates multi-dimensional environmental feature sub-vectors and the driver's multi-dimensional biological feature sub-vectors, and dynamically calibrates and integrates the environmental feature matrix, environmental feature vectors, biological feature matrix, and biological feature vectors through weighted fusion of scenario coefficients and the environmental and biological feature matrices, it overcomes the problems of perception bias and decision rigidity caused by traditional solutions that rely solely on single-dimensional data or fixed rules. This enables the vehicle's control system to more comprehensively and accurately understand complex driving scenarios and output highly matched control parameters, ultimately significantly improving the accuracy of adaptive driving style adjustment while ensuring safety.

[0027] In some examples, the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension are obtained using the following formula: (1); in, Let i be the environmental feature weights corresponding to the i-th dimension's environmental feature sub-vector. μ is a hyperparameter for adjusting the slope of the i-th dimension environmental feature subvector. i Let be the historical mean of the environmental feature sub-vectors of the i-th dimension.

[0028] For example, The real-time acquired value of the i-th dimension environmental feature sub-vector is used to control the steepness of the curve, typically set to 0.5~2.0. The historical mean of the i-th dimension environmental feature sub-vector can be obtained through offline training, such as setting the mean of the road adhesion coefficient to 0.5. For example, the environmental feature vector may include environmental feature sub-vectors such as road adhesion coefficient, light intensity, or visibility level. When calculating the corresponding environmental feature weight for the road adhesion coefficient environmental feature sub-vector, since the data range of the road adhesion coefficient is 0.1-1.0, and the current road surface is dry asphalt, requiring a high adhesion coefficient, the real-time acquired value of the road adhesion coefficient environmental feature sub-vector is 0.8. The hyperparameter of the adjustment slope corresponding to the road adhesion coefficient is... =1.0, the historical average value of the road surface adhesion coefficient. =0.5, therefore the environmental feature weight corresponding to the road adhesion coefficient (e1) is = = This indicates that the higher the road surface adhesion coefficient (closer to 1.0), the closer the weight is to 1.0, indicating that the vehicle's tire grip is good, and the safety redundancy can be appropriately reduced.

[0029] When calculating the environmental feature weights for the environmental feature subvector of light intensity (e2), since the data range of light intensity is 0-100000 lux, and the current environment is a strong light environment at noon on a sunny day, the real-time collected value of the environmental feature subvector of light intensity is 50000 lux. The hyperparameter of the adjustment slope corresponding to light intensity... =0.00001, because the data range of light intensity is large, α2 needs to be extremely small, corresponding to the historical mean of light intensity. =50000, therefore the environmental feature weight corresponding to the light intensity is = = This indicates that when the light intensity is at the mean, the weight is 0.5; if the light intensity is <1000 lux (nighttime), it will be close to 1.0, increasing the sensitivity to low light risk.

[0030] When calculating the environmental feature weights for the visibility level (e3) sub-vector, since the visibility level data ranges from 1 to 5, with level 5 being the highest, and the current situation is moderate haze with visibility of approximately 200m, the real-time acquired value for this environmental feature sub-vector is level 3. The hyperparameter α3 of the adjustment slope corresponding to the visibility level is 0.5, and the historical mean μ3 corresponding to the visibility level is 3. Therefore, the environmental feature weights corresponding to the visibility level are: = = In emergency mode, The weight needs to be multiplied by 1.8 (switching logic), that is ( =0.5×1.8=0.9), at which point the impact of visibility on decision-making is significantly enhanced.

[0031] In some examples, the biometric weights corresponding to the biometric subvectors of each dimension are obtained using the following formula: (2); in, Let be the biometric weights corresponding to the biometric subvectors of the j-th dimension. The preset weighting coefficients, Let j be the historical minimum value of the biometric subvector of the j-th dimension. The historical maximum value of the biometric subvector of the j-th dimension.

[0032] For example, The preset dynamic weighting coefficients range from 0.5 to 2.0 and are determined by a lookup table based on the current driving scenario (e.g., parent-child / emergency mode). When the biometric subvector is heart rate variability, the corresponding historical minimum value is 50ms and the historical maximum value is 150ms. The preset weighting coefficients are dynamically adjusted according to the current driving scenario, such as when the current driving scenario is parent-child mode. =1.8, to enhance the weight of voice emotion features. Biometric vectors include heart rate variability, blink frequency, and voice emotion entropy. When calculating the biometric weight corresponding to the heart rate variability (b1) biometric subvector, since the heart rate variability data range is 50-150ms, and the driver is mildly fatigued at this time, the heart rate variability is reduced. Therefore, the real-time acquisition value of the heart rate variability biometric subvector is 55ms. The preset weight coefficient β1 for heart rate variability is 2.0 (in fatigue mode, β1 is increased to 2.0 to highlight heart rate features). Therefore, the biometric weight corresponding to heart rate variability is... =2.0× =0.1. The lower the heart rate variability and the closer it is to 50ms, the closer the value after linear normalization is to 0. However, the weight increases after amplification, indicating that the system pays more attention to the driver's fatigue state.

[0033] When calculating the biometric weights corresponding to the blink frequency (b2) biometric subvector, since the blink frequency data range is 5-30 times / minute, and the driver blinks frequently and is distracted, the real-time acquisition value of the blink frequency biometric subvector is 28 times / minute. The preset weight coefficient β1 for the blink frequency is 1.2 (in conservative mode, β1=1.2). Therefore, the biometric weights corresponding to the blink frequency are: =1.2× =1.2× =1.104 indicates that the higher the blinking frequency (close to 30 times / minute), the closer the weight is to β2 (1.2), and the more the system will increase the following distance and reduce the steering speed.

[0034] When calculating the biometric weights corresponding to the biometric subvector of voice emotion entropy (b3), since the data range of voice emotion entropy is 0.1-0.9, the higher the entropy value, the more unstable the emotion. Furthermore, the driver was emotionally agitated, such as screaming or speaking angrily. Therefore, the real-time collected value of the biometric subvector of voice emotion entropy was 0.85, and the preset weight coefficient β1 corresponding to the voice emotion entropy was 1.8 (in parent-child mode). =1.8 (prioritizing responses to voice commands), therefore the biometric weight corresponding to the voice emotion entropy is... =1.8× =1.8× =1.8 × 0.9375 = 1.6875, indicating that when the emotional entropy is close to 0.9, the weight is close to... (1.8) The system will prioritize responding to voice commands (such as “Brake!”) and trigger emergency braking.

[0035] In some examples, after obtaining the vehicle's environmental feature vector and the driver's biometric vector, the process includes: Based on the current driving scenario of the vehicle and the environmental feature vector, determine the environmental feature sub-vector to be corrected in the environmental feature vector, and obtain the environmental correction coefficient corresponding to the environmental feature sub-vector to be corrected in the current driving scenario; The environmental feature subvector to be corrected is corrected based on the environmental correction coefficient to obtain the corrected environmental feature subvector. Based on the current driving scenario of the vehicle and the biometric vector, determine the biometric sub-vector to be corrected in the biometric vector, and obtain the biometric correction coefficient corresponding to the biometric sub-vector to be corrected in the current driving scenario. The biological feature subvector to be corrected is corrected based on the biological correction coefficient to obtain the corrected biological feature subvector.

[0036] For example, after acquiring the environmental feature vector, the system first selects one or more sub-vectors that have the most critical impact on the current driving scenario (e.g., emergency mode, rain mode, or mountain road mode) from multiple environmental feature sub-vectors based on the real-time determined current driving scenario. These are then identified as the environmental feature sub-vectors to be corrected. For instance, in emergency mode, visibility level is identified as the environmental feature sub-vector to be corrected; in rain mode, road adhesion coefficient is identified as the environmental feature sub-vector to be corrected. Subsequently, the system obtains the environmental correction coefficient specifically set for the environmental feature sub-vector to be corrected according to Table 1, which is used to amplify the contribution of this feature dimension in specific scenarios. The correction process involves multiplying the environmental correction coefficient with the environmental feature sub-vector to be corrected. For example, in emergency mode, the preset environmental correction coefficient for the environmental feature sub-vector to be corrected in the visibility level dimension is 1.8. Multiplying the environmental feature sub-vector to be corrected by 1.8 yields the corrected environmental feature sub-vector, thus completing the targeted enhancement of the environmental feature sub-vector.

[0037] The system synchronously selects the sub-vector most relevant to the driver's state in the current driving scenario from multiple biometric sub-vectors, based on the same current driving scenario, and identifies it as the biometric sub-vector to be corrected. For example, in parent-child mode, voice emotion entropy is identified as the biometric sub-vector to be corrected; in fatigue mode, heart rate variability is identified as the biometric sub-vector to be corrected. Subsequently, the system obtains the biometric correction coefficient set for the biometric sub-vector to be corrected according to Table 1. This biometric correction coefficient is also a scalar multiplier used to adjust the feature intensity. The correction process is the same as that for environmental feature sub-vectors, and will not be elaborated further here. This breaks through the limitations of traditional solutions that use fixed or uniform processing modes for all feature dimensions. By identifying and dynamically enhancing the core influencing factors under different driving scenarios, it achieves scenario-based fine-tuning of feature weights. This allows the subsequent fusion decision vector to more sensitively and accurately reflect the key risk factors or driver intentions in specific scenarios (such as emergency, parent-child, fatigue), thereby providing a more targeted and discriminative decision basis for the generation of control parameters, ultimately improving the adaptability and adjustment accuracy of the vehicle system in complex and changing scenarios.

[0038] The current driving scenario follows the following scenario arbitration priority rules: a three-level priority mechanism is adopted: 1. Safety scenarios (emergency / snow mode) have the highest priority, forcibly interrupting other scenarios upon triggering, with biometric weight increased by 60%; 2. Driving state scenarios (fatigue / aggressive mode) are secondary, dynamically arbitrated based on vital signs parameters such as heart rate variability (HRV<60ms); 3. Environmental scenarios (highway / congestion mode) have the lowest priority, yielding to safety scenarios. In case of conflict, a predefined transfer strategy is executed: if "fatigue + emergency" is triggered simultaneously, a smooth switch to emergency mode is performed within 0.5 seconds, using an exponential smoothing algorithm to avoid parameter mutations. For conflicts within the same scenario: if "fatigue mode" and "parent-child mode" are triggered simultaneously, the scenario with higher biometric weight is prioritized (fatigue mode b1×2.0 > parent-child mode b4×1.8).

[0039]

[0040] Table 1 In some examples, the scene coefficients, the environmental feature matrix, the environmental feature vector, the biological feature matrix, and the biological feature vector are weighted and fused using both environmental and biological feature matrices to obtain a fused decision vector, which is obtained through the following formula: (3); in, Let be the fusion decision vector. The coefficients for the aforementioned scenarios. The environmental feature matrix, The environmental feature vector, The biometric matrix, The biological feature vector is described above.

[0041] For example, the scene coefficient ( This is used to dynamically allocate the weights of environmental features and biological features. The specific allocation rules are shown in Table 1. For example, in the current driving scenario, which is in emergency mode... =0.4, the total weight of biometric features is increased by 60% in the current driving scenario of the vehicle (S k Under the current driving scenario, the fusion decision vector ( The formula for calculating ) is: ,in, For the current driving scenario (S) k The corresponding scene coefficients (0≤) ≤1), For the current driving scenario (S) k The environmental feature matrix under ) For the current driving scenario (S) kBiometric matrix under ) The environmental feature vector, The biological feature vector is described above.

[0042] In some examples, the determination of multiple control parameters of the vehicle in the current driving scenario based on the fused decision vector is obtained by the following formula: (4); in, Let i be the i-th control parameter among multiple control parameters. The basic control parameter for the i-th control parameter in standard mode. Let i be the maximum adjustment amount of the i-th control parameter. This is the fusion decision vector for the current driving scenario. Let i be the feature sensitivity vector of parameter i. This is the activation function.

[0043] For example, the maximum adjustment of the i-th control parameter is either a positive or negative adjustment range. The feature sensitivity vector of parameter i is obtained through offline training. A training set is constructed using samples of 500 drivers (including different ages and driving styles), covering the 18 sub-scenario triggering conditions in Table 1 (e.g., parent-child mode = child voice + child seat signal). During training, the control parameter adjustment error (e.g., following distance deviation < 0.1s) is used as the optimization objective, and iterative updates are performed using gradient descent. Values ​​to ensure the accuracy of parameter adjustments: Overtaking threshold in aggressive mode Reduce following distance by 20% in conservative mode Improvement by 30%. The model performs incremental learning every 100km and dynamically corrects its performance. To adapt to changes in driver habits, Using a sigmoid activation function, the decision value is mapped to an adjustment ratio of [0,1]. For example, following distance adjustment in emergency mode. Current driving scenario (S k This is the emergency mode, triggered by a scream detection combined with a sudden turn of the steering wheel. The environmental feature vector E = [0.8, 50000, 3, 40, 5, 0.03, 2º, 65] (values ​​e1~e8 sequentially), and the biometric feature vector B = [55, 28, -12°, 0.85, 1.1] (values ​​b1~b5 sequentially). The baseline value for following distance is... =1.5s, maximum adjustment amount =+0.5s, the corresponding scenario coefficient in emergency mode =0.4, preset weight coefficient β=1.6, environmental feature matrix is = diag([0.72, 0.65, 0.88, 0.42, 0.35, 0.22, 0.18, 0.45]) (The weights of each dimension are obtained by sigmoid normalization). = 0.72×0.8 + 0.65×50000 / 1e5 + 0.88×3 / 5 + 0.42×40 / 100 + 0.35×5 / 20 + 0.22×0.03 / 0.1 + 0.18×2 / 15 + 0.45×65 / 120 = 0.576 + 0.325 + 0.528 + 0.168 + 0.0875 + 0.066 + 0.024 + 0.24375 = 2.018. The biometric matrix is... = diag([1.6×0.1, 1.6×0.9, 1.6×0.15, 1.6×0.9, 1.6×0.85]) = diag([0.16, 1.44, 0.24, 1.44, 1.36]) (β=1.6×normalized values ​​of each feature). = 0.16×55 / 150 + 1.44×28 / 30 + 0.24×(-12) / 15 + 1.44×0.85 / 0.9 + 1.36×1.1 / 1.2= 0.0587 + 1.344 - 0.192 + 1.36 + 1.253 = 3.824. The fusion decision vector is... = 0.4×2.018 + (1-0.4)×3.824 = 0.807 + 2.294 = 3.101. The following distance parameter is p1 = 1.5 + 0.5×σ(3.101×0.72) (V1=0.72 is the following distance sensitivity vector), σ(2.233) = 0.90 (calculated by the sigmoid function). p1 = 1.5 + 0.5×0.904 = 1.952s (actual output 1.95s, rounded to 0.01s). In emergency mode, the following distance increases from 1.5s to 1.95s (+30%), which is consistent with the biometric-driven safety strategy and the theoretical adjustment trend.

[0044] In some examples, multiple control parameters of the vehicle in the current driving scenario are determined based on the fused decision vector, including: Determine whether the target control parameter data is within a preset safety threshold. If the target control parameter data is within a preset safety threshold, acquire multiple sensor data and determine the data difference between the multiple sensor data. If the data difference is less than or equal to a preset difference threshold, then it is determined whether the driver has the ability to take over. If the driver has the ability to take over, the current control parameter data of the vehicle will be adjusted to the target control parameter data.

[0045] For example, after determining multiple control parameters of the vehicle in the current driving scenario, these parameters need to undergo three levels of verification. The first level of verification involves hard constraints on parameter boundaries, specifically: all control parameters are strictly limited to safety thresholds (e.g., following distance ≥ 0.8s). Hardware-in-the-loop (HIL) simulation verification must achieve a 100% pass rate before deployment. After passing the first level of verification, a second level of verification is performed: multi-source consistency detection. This multi-source consistency verification ensures perception reliability through cross-validation of multi-sensor data, avoiding decision-making biases caused by single sensor failures (e.g., camera obstruction or radar false detection). Its core logic is: for the same physical target (e.g., lateral distance or obstacle distance), data is simultaneously collected by different sensors such as cameras, millimeter-wave radar, and lidar. The data difference is calculated; if it exceeds a threshold, a weighted fusion of environmental and biological dual-feature matrices is initiated; if it is within the threshold, the average is taken. Finally, the fusion result is compared with the safety boundaries of the control parameters to determine whether to execute the control command. The threshold setting for multi-source consistency verification needs to be determined comprehensively by combining sensor accuracy, scenario risk level, and control parameter safety redundancy. The core principle is: the threshold should be ≤ twice the maximum error of the sensor (covering the 95% confidence interval) and less than the safety redundancy of the control parameters.

[0046] For example, following distance (P1) can be adjusted from 0.8 to 2.0 seconds, with a threshold of 0.1 seconds. The threshold is set based on sensor accuracy: millimeter-wave radar time measurement error ±0.05 seconds, camera time synchronization error ±0.03 seconds, and the threshold is twice the maximum error (0.05 × 2 = 0.1 seconds). Safety redundancy: The safety boundary for following distance is a minimum of 0.8 seconds (mandatory by regulations). The threshold of 0.1 seconds is much smaller than 0.8 seconds, ensuring that even with deviations in multi-sensor data, the fusion result remains within a safe range. For example, if the radar measures a following distance of 1.5 seconds and the camera measures 1.35 seconds, the difference of 0.15 seconds > 0.1 seconds, triggering a weighted fusion of environmental and biological dual-feature matrices (radar weight 0.7, camera weight 0.3), avoiding excessively close following due to time synchronization errors.

[0047] Lane change safety distance (P2) is adjustable from 1.5 to 3.5 meters, with a threshold of 0.3 meters (high-speed scenario) / 0.5 meters (urban scenario). The threshold is set based on sensor accuracy: LiDAR distance error ±0.1 meters, millimeter-wave radar ±0.2 meters, resulting in a threshold of (0.2 × 2 = 0.4 meters), but must be less than the safety redundancy of 0.3 meters (P2 - 0.3 meters is the safety boundary in high-speed scenarios). Therefore, it is strictly set at 0.3 meters in high-speed scenarios. Scenario risk: Lane change collisions at high speeds have severe consequences, requiring a stricter threshold; in urban areas with lower vehicle speeds, the threshold can be relaxed to 0.5 meters. Example: In a high-speed scenario, if the LiDAR measures a lateral distance of 2.8 meters and the camera measures 2.4 meters, the difference of 0.4 meters > 0.3 meters, and the system immediately cancels the lane change (because the fused distance may be lower than the safety boundary of 2.5 meters).

[0048] Cornering deceleration rate (P3) is adjustable from 5% to 10%, with a threshold of 1%. The threshold is set based on sensor accuracy: IMU (Inertial Measurement Unit) deceleration rate error ±0.5%, and the threshold is set to twice the error (0.5% × 2 = 1%). For safety control: a deceleration rate deviation >1% may cause vehicle sideslip (e.g., setting the deceleration rate to 8% but actually executing 9.5%). Strict verification of the consistency of data from multiple sensors (IMU + ESP (Electronic Stability Program)) is required. For example: if the IMU measures a deceleration rate of 7.2% and the ESP sensor measures 8.5%, the difference is 1.3% > 1%. The system uses a weighted fusion of environmental and biological dual-feature matrices (IMU weight 0.6, ESP weight 0.4) to avoid the risk of fishtailing due to excessive deceleration.

[0049] Starting acceleration (P4) is adjustable from 0.2g to 0.6g, with a threshold of 0.05g. The threshold is set based on sensor accuracy: the accelerometer error is ±0.025g, and the threshold is twice the error (0.025 × 2 = 0.05g). Comfort constraints: Passengers are sensitive to changes in acceleration (0.05g is approximately 0.5m / s²). Excessive deviation can cause a "jerkiness," so it's crucial to ensure consistency between the motor controller and the accelerometer data. For example: if the motor controller outputs 0.4g and the sensor reports 0.46g, the difference (0.06g > 0.05g) will cause the system to reduce its output to 0.43g (the fused value), balancing power and comfort.

[0050] Steering angular velocity (P5) has an adjustment range of 0.5° / s to 2.5° / s, with a threshold of 0.2° / s. The threshold is set based on sensor accuracy: the steering angle sensor error is ±0.1° / s, and the threshold is twice the error (0.1 × 2 = 0.2° / s). Handling stability: A steering angular velocity deviation > 0.2° / s will cause lane departure (e.g., set to 1.5° / s, actual deviation is 1.8° / s). It is necessary to verify the consistency between the steering wheel angle sensor and the camera's lane line detection data. For example: Sensor A measures a steering angular velocity of 1.2° / s, and Sensor B measures 1.5° / s, a difference of 0.3° / s > 0.2° / s. The system uses mean fusion (1.35° / s) and reduces the steering gain.

[0051] The overtaking decision threshold (P6) has an adjustment range of 10%-25%, with a threshold of 2%. The threshold is set based on sensor accuracy: ±1% of the preceding vehicle's speed measurement error (e.g., ±1 km / h at 100 km / h), and the threshold is set to twice this error (1% × 2 = 2%). Decision logic: The overtaking threshold determines whether to trigger a lane change (e.g., "overtake if the preceding vehicle's speed is 15% lower than the set speed"). If the speed difference between multiple sensors is greater than 2%, it may lead to misjudgment (e.g., if the actual preceding vehicle speed is only 13% lower, it may be misjudged as 15%). For example: Navigation data shows the preceding vehicle's speed is 90 km / h, while radar measures 88 km / h, a difference of 2.2% > 2%. The system delays the overtaking decision until the data difference is less than 2%.

[0052] Lane centering accuracy (P7) is adjustable from ±20 to ±50cm, with a threshold of 5cm (highway) / 10cm (urban). The threshold is set based on sensor accuracy: LiDAR lateral position error ±2cm, camera ±3cm, and a threshold of (3×2=6cm), with a strict 5cm threshold for high-speed scenarios. Safety constraints: High-speed lane departure risks are high, requiring that the difference in lane line detection data from multiple sensors (LiDAR + camera) be <5cm to avoid driving over the lines. For example: LiDAR measures vehicle centering +3cm, camera measures +8cm, the difference is 5cm = threshold, the system takes the average +5.5cm, still within the ±30cm safe range.

[0053] Accelerator pedal sensitivity (P8) is adjustable from 50% to 150%, with a threshold of 5%. The threshold is set based on sensor accuracy: pedal travel sensor error ±2.5%, with the threshold set to twice the error (2.5% × 2 = 5%). Human-machine interaction: A sensitivity deviation >5% will cause a "mismatch between pedal pressure and acceleration feedback" (e.g., setting 100% sensitivity but actually outputting 108%). It is necessary to verify the consistency of data between the pedal sensor and the motor controller. For example: if the pedal sensor feedback sensitivity is 110%, but the controller actually executes 116%, the difference of 6% > 5%. The system will reduce the sensitivity to 113% to prevent the driver from misinterpreting "excessive power."

[0054] Brake pedal nonlinearity (P9) has an adjustment range of 0.3-0.8 and a threshold of 0.05. The threshold is set based on sensor accuracy: brake pressure sensor error ±0.025, with the threshold set at twice the error (0.025 × 2 = 0.05). Braking smoothness: Nonlinearity describes the relationship between pedal travel and braking force. A deviation >0.05 can lead to "light braking with heavy braking" or "heavy braking with light braking," requiring verification of the consistency between the pedal force sensor and ESP data. For example: if the sensor measures a nonlinearity of 0.6, and the ESP feedback is 0.66, the difference is 0.06 > 0.05. The system uses a weighted fusion of environmental and biological dual feature matrices (sensor weight 0.5, ESP weight 0.5) to ensure linear brake pedal feel.

[0055] Voice command response delay (P) 10 The adjustable range for voice command response latency is 0.3s-1.0s, with a threshold of 0.1s. The threshold is set based on the following: sensor accuracy: microphone pickup latency ±0.03s, voice module processing latency ±0.04s, threshold is (0.04×2=0.08s), rounded up to 0.1s. Interactive experience: A response latency difference >0.1s will make users feel "command stuttering" (e.g., saying "accelerate," 0.3s vs 0.45s response). It is necessary to verify the consistency of the microphone and vehicle system timestamps. For example: if the microphone records a command trigger time of 0.5s, and the vehicle system receives it at 0.65s, the difference is 0.15s > 0.1s. The system will synchronize the timestamps and recalculate the response latency.

[0056] Automatic parking accuracy (P) 11The automatic parking accuracy adjustment range is ±5-±15cm, with a threshold of 2cm. The threshold setting is based on: Sensor accuracy: Ultrasonic radar distance error ±1cm, the threshold is twice the error (1×2=2cm). Parking safety: High parking accuracy is required (≤±10cm). If the difference in data from multiple sensors is >2cm (e.g., 30cm on the left vs. 27cm on the right), it may lead to scratches. For example: The ultrasonic radar on the left measures 25cm, and the one on the right measures 22cm, a difference of 3cm > 2cm. The system will activate the lidar for auxiliary calibration to ensure that the final parking deviation is <5cm.

[0057] Energy recovery intensity (P) 12 The energy recovery intensity can be adjusted from 10% to 40%, with a threshold of 3%. The threshold is set based on the following: Sensor accuracy: Battery Management System (BMS) current measurement error ±1.5%, with the threshold set to twice the error (1.5% × 2 = 3%). Range and comfort: A recovery intensity deviation >3% will affect range (e.g., set to 30%, actual is 26%) or cause a dragging sensation (actual is 34%). It is necessary to verify the consistency of data between the BMS and the motor controller. For example: If the BMS measures a recovery intensity of 28%, and the motor controller reports 32%, the difference is 4% > 3%. The system will use the average of 30% to balance energy recovery and driving smoothness.

[0058] Specific examples of lane changing scenarios: Current driving scenario: The system is preparing to perform an automatic lane change. The current safe lane change distance parameter is P2 = 1.5 m (minimum safe distance for high-speed scenarios). Data acquisition: The camera's detection distance is: (d) camera = 1.2 m) (due to visual errors caused by strong light), the detection range of millimeter-wave radar is: (d radar = 1.5m) (radar echo stable), the detection range of the lidar is: (d lidar = 1.3m) (point cloud clustering result), consistency judgment: Δd=max(1.2,1.5,1.3) min(1.2, 1.5, 1.3) = 1.5 1.2 = 0.3m ≤ 0.5m, at this point the data consistency is good, so the mean is taken and the data is merged. The merged distance is: d fused =1.2+1.5+1.33=1.33m. Safety check: Safety boundary: (P2-0.3 = 1.5 - 0.3 = 1.2 m). Comparison result: (1.33 m > 1.2 m), the safety conditions are met, and lane changing is allowed. Abnormal scenario comparison: If the lidar malfunctions and outputs (d... lidar= 0.8 m), then: (Δd = 1.5 - 0.8 = 0.7m>0.5m), at this time, the environmental biological dual feature matrix weighted fusion is triggered: weight reference sensor accuracy pre-allocation: d lidar (0.5) + d radar (0.3) + d camera (0.2). d fused =0.5×0.8+0.3×1.5+0.2×1.2=0.4+0.45+0.24=1.09m. Since (1.09 m<1.2 m), the system immediately cancels the lane change and displays an alarm "Insufficient lateral distance, lane change canceled" on the central control screen.

[0059] After passing the Level 2 verification, the Level 3 verification is conducted, which assesses the driver's willingness to take over. Specifically, this involves judging the driver's willingness to take over based on facial orientation (gaze deviation > 2 seconds) or steering wheel torque (< 0.5 N·m for 5 seconds). In high-risk scenarios, the system proactively alerts the driver to take over, providing a 3-second buffer period.

[0060] Detailed control example (parent-child mode): The triggering conditions are: the voice module recognizes the voiceprint of a child under 8 years old (200-600Hz frequency characteristics), the rear child seat pressure sensor detects a weight >15kg, and the ISOFIX (child seat anchoring system) interface is locked. Parameter adjustment sequence: Longitudinal control parameters: Starting acceleration P4 decreases from 0.4g to 0.25g (a reduction of 37.5%). Braking deceleration rate limit ≤0.3g (to prevent motion sickness in children). Following distance P1 increases from 1.5s to 2.0s (increasing collision buffer). Lateral control parameters: Steering angle P5 decreases from 2.0° / s to 1.2° / s (a reduction of 40%). Lane change safety distance P2 decreases from 1.0m to 1.5m (increasing lateral redundancy). Interactive control parameters: Voice command response delay P... 10 Set to 0.3s (highest priority), entertainment system volume limited to ≤60dB (to protect children's hearing), air conditioning automatically switches to recirculation + PM2.5 filtration mode. Enhanced safety monitoring: rear camera captures child status every 30 seconds, and if a child is detected unbuckled, a three-level alarm (visual to auditory to tactile) is issued within 3 seconds, and vehicle speed is automatically limited to ≤80km / h (highway sections).

[0061] In some examples, the environmental feature vector includes at least two of the following: road surface adhesion coefficient, light intensity, visibility level, traffic density, speed limit change rate, curve curvature, slope value, and noise decibel value; and / or The biometric vector includes at least two of the following: heart rate variability, blink frequency, mouth corner upturn angle, vocal emotion entropy, and steering wheel operation entropy; and / or The multiple control parameters include at least two of the following distance, lane change safety distance, cornering deceleration rate, starting acceleration, steering angle velocity, overtaking decision threshold, lane centering accuracy, accelerator pedal sensitivity, brake pedal nonlinearity, voice command response delay, automatic parking accuracy, and energy recovery intensity.

[0062] For example, the environmental feature vector includes at least two of the following: road surface adhesion coefficient, light intensity, visibility level, traffic density, speed limit change rate, curve curvature, slope value, and noise decibel value, as shown in Table 2.

[0063]

[0064] Table 2 Biometric vectors include at least two of the following: heart rate variability, blink frequency, mouth corner upturn angle, vocal emotion entropy, and steering wheel operation entropy. Biometric vectors are shown in Table 3.

[0065]

[0066] Table 3 Multiple control parameters include at least two of the following distance, lane change safety distance, cornering deceleration rate, starting acceleration, steering angle rate, overtaking decision threshold, lane centering accuracy, accelerator pedal sensitivity, brake pedal nonlinearity, voice command response delay, automatic parking accuracy, and energy recovery intensity. These multiple control parameters are shown in Table 4.

[0067]

[0068] Table 4 like Figure 2 As shown, this application proposes a vehicle control parameter determination system, which includes: a weight determination module 21, a weight fusion module 22, and a parameter adjustment module 23; The weight determination module 21 is configured to: obtain the environmental feature vector of the vehicle and the biometric vector of the driver, wherein the environmental feature vector includes environmental feature sub-vectors of multiple dimensions and the biometric vector includes biometric sub-vectors of multiple dimensions; obtain the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension; and combine the environmental feature weights to obtain the environmental feature matrix corresponding to the environmental feature vectors. The weight fusion module 22 is configured to: obtain the biofeature weights corresponding to the biofeature sub-vectors of each dimension, combine the biofeature weights to obtain the biofeature matrix corresponding to the biofeature vector; obtain the scene coefficients corresponding to the current driving scene of the vehicle, and perform weighted fusion of the environmental and biological dual feature matrices based on the scene coefficients, the environmental feature matrix, the environmental feature vector, the biofeature matrix, and the biofeature vector to obtain the fusion decision vector; The parameter adjustment module 23 is configured to determine multiple control parameters of the vehicle in the current driving scenario based on the fusion decision vector.

[0069] The effects of applying the aforementioned method in the above system can be found in the description of the aforementioned method embodiments, and will not be repeated here.

[0070] like Figure 3 As shown, this application embodiment also provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, it implements the steps of any of the above-described methods for determining vehicle control parameters.

[0071] Since the electronic device described in this embodiment is the device used to implement a vehicle control parameter determination device in the embodiments of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any device used by those skilled in the art to implement the method in the embodiments of this application falls within the scope of protection of this application.

[0072] In practical implementation, when the computer program 311 is executed by the processor, it can achieve the following: Figure 1 Any of the corresponding implementation methods in the embodiments.

[0073] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0074] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program code.

[0075] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0076] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0077] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0078] This application also provides a computer program product, which includes computer software instructions that, when executed on a processing device, cause the processing device to execute the LDPC decoding method of a solid-state drive controller.

[0079] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0080] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0081] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0082] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0083] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0084] Obviously, those skilled in the art can make various modifications and variations to this specification without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims and their equivalents, this specification is also intended to include such modifications and variations.

Claims

1. A method for determining vehicle control parameters, characterized in that, The method includes: The environmental feature vector of the vehicle and the biometric vector of the driver are obtained. The environmental feature vector includes environmental feature sub-vectors of multiple dimensions, and the biometric vector includes biometric sub-vectors of multiple dimensions. Obtain the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension, and combine the environmental feature weights to obtain the environmental feature matrix corresponding to the environmental feature vectors. Obtain the biofeature weights corresponding to the biofeature subvectors of each dimension, and combine the biofeature weights to obtain the biofeature matrix corresponding to the biofeature vectors. Obtain the scene coefficients corresponding to the current driving scenario of the vehicle, and perform weighted fusion of the environmental and biological dual feature matrices based on the scene coefficients, the environmental feature matrix, the environmental feature vector, the biological feature matrix, and the biological feature vector to obtain the fusion decision vector; The vehicle's multiple control parameters in the current driving scenario are determined based on the fusion decision vector.

2. The method for determining vehicle control parameters according to claim 1, characterized in that, The environmental feature weights corresponding to the environmental feature sub-vectors of each dimension are obtained through the following formula: ； in, Let i be the environmental feature weights corresponding to the i-th dimension's environmental feature sub-vector. Let be the hyperparameter for adjusting the slope of the i-th dimension of the environmental feature sub-vector. Let be the historical mean of the environmental feature sub-vectors of the i-th dimension.

3. The method for determining vehicle control parameters according to claim 1, characterized in that, The biometric weights corresponding to the biometric subvectors of each dimension are obtained using the following formula: ； in, Let be the biometric weights corresponding to the biometric subvectors of the j-th dimension. The preset weighting coefficients, Let j be the historical minimum value of the biometric subvector of the j-th dimension. The historical maximum value of the biometric subvector of the j-th dimension.

4. The method for determining vehicle control parameters according to any one of claims 1-3, characterized in that, After obtaining the vehicle's environmental feature vector and the driver's biometric vector, the process includes: Based on the current driving scenario of the vehicle and the environmental feature vector, determine the environmental feature sub-vector to be corrected in the environmental feature vector, and obtain the environmental correction coefficient corresponding to the environmental feature sub-vector to be corrected in the current driving scenario; The environmental feature subvector to be corrected is corrected based on the environmental correction coefficient to obtain the corrected environmental feature subvector. Based on the current driving scenario of the vehicle and the biometric vector, determine the biometric sub-vector to be corrected in the biometric vector, and obtain the biometric correction coefficient corresponding to the biometric sub-vector to be corrected in the current driving scenario. The biological feature subvector to be corrected is corrected based on the biological correction coefficient to obtain the corrected biological feature subvector.

5. The method for determining vehicle control parameters according to claim 1, characterized in that, The scene coefficients, the environmental feature matrix, the environmental feature vector, the biological feature matrix, and the biological feature vector are weighted and fused using both environmental and biological feature matrices to obtain a fused decision vector, which is obtained through the following formula: ； in, Let be the fusion decision vector. The coefficients for the aforementioned scenarios. The environmental feature matrix, The environmental feature vector, The biometric matrix, The biological feature vector is described above.

6. The method for determining vehicle control parameters according to claim 1, characterized in that, The determination of multiple control parameters of the vehicle in the current driving scenario based on the fused decision vector is obtained through the following formula: ； in, Let i be the i-th control parameter among multiple control parameters. The basic control parameter for the i-th control parameter in standard mode. Let i be the maximum adjustment amount of the i-th control parameter. This is the fusion decision vector for the current driving scenario. Let i be the feature sensitivity vector of parameter i. This is the activation function.

7. The method for determining vehicle control parameters according to claim 1, characterized in that, The environmental feature vector includes at least two of the following: road surface adhesion coefficient, light intensity, visibility level, traffic density, speed limit change rate, curve curvature, slope value, and noise decibel value; and / or The biometric vector includes at least two of the following: heart rate variability, blink frequency, mouth corner upturn angle, vocal emotion entropy, and steering wheel operation entropy; and / or The multiple control parameters include at least two of the following distance, lane change safety distance, cornering deceleration rate, starting acceleration, steering angle velocity, overtaking decision threshold, lane centering accuracy, accelerator pedal sensitivity, brake pedal nonlinearity, voice command response delay, automatic parking accuracy, and energy recovery intensity.

8. A vehicle control parameter determination system, characterized in that, The system includes: a weight determination module, a weight fusion module, and a parameter adjustment module; The weight determination module is configured to: obtain the environmental feature vector of the vehicle and the biometric vector of the driver, wherein the environmental feature vector includes environmental feature sub-vectors of multiple dimensions and the biometric vector includes biometric sub-vectors of multiple dimensions; obtain the environmental feature weights corresponding to the environmental feature sub-vectors of each dimension; and combine the environmental feature weights to obtain the environmental feature matrix corresponding to the environmental feature vectors. The weight fusion module is configured to: obtain the biofeature weights corresponding to the biofeature sub-vectors of each dimension, combine the biofeature weights to obtain the biofeature matrix corresponding to the biofeature vector; obtain the scene coefficients corresponding to the current driving scene of the vehicle, and perform weighted fusion of the environmental and biological dual feature matrices based on the scene coefficients, the environmental feature matrix, the environmental feature vector, the biofeature matrix, and the biofeature vector to obtain the fusion decision vector; The parameter adjustment module is configured to determine multiple control parameters of the vehicle in the current driving scenario based on the fusion decision vector.

9. An electronic device, comprising: The memory and processor are characterized in that the processor is used to execute a computer program stored in the memory to implement the steps of a vehicle control parameter determination method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of a vehicle control parameter determination method as described in any one of claims 1-7.

Images