Intelligent control system of electric steering gear based on distributed cooperative technology

Abstract

The present application relates to the technical field of vehicle control, in particular to an electric steering wheel intelligent control system based on distributed collaborative technology, which comprises a model parameter coefficient adjustment module, which is used for obtaining the deviation degree of each steering assist data point in the current steering data segment based on the longitudinal coordinate difference of the corresponding position data points between the current steering data segment and the historical steering data segment in the similar cluster and the stage stability index value, adjusting the parameter coefficient of the ARIMA model based on the deviation degree, obtaining the target parameter coefficient of each steering assist data point in the current steering data segment, an assist value output module, which is used for obtaining the assist prediction value at the next driving time based on the current steering data segment, the ARIMA model and the target parameter coefficient, obtaining the target steering assist value at the next driving time based on the assist prediction value and the to-be-adjusted assist value at the next driving time. And the present application can improve the vehicle assist or control stability.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and more specifically to an intelligent control system for electric steering based on distributed cooperative technology. Background Technology

[0002] Electric Power Steering (EPS) is a car steering system that uses an electric motor as the power source. Compared to traditional hydraulic power steering systems, it has significant advantages in energy efficiency, millisecond-level dynamic response, and programmable power assist characteristics, making it the mainstream configuration for modern passenger car steering systems. Moreover, by precisely controlling the motor's output torque, EPS not only effectively reduces overall vehicle energy consumption but also significantly reduces the driver's workload, improving steering precision, handling stability, and driving safety.

[0003] Current EPS control strategies primarily rely on passive response control, which uses linear or nonlinear mapping based on the current vehicle speed and steering torque to generate the assist output for the next moment. However, this mechanism is prone to assist and instability issues, leading to decreased vehicle steering stability. To address this, predictive control algorithms have been introduced, such as predicting future assist demand based on the Autoregressive Integral Moving Average (ARIMA) model to achieve pre-load assist. This means that the ARIMA model increases the motor output before the driver actually applies greater torque, smoothing the assist transition. However, in real-world road conditions, significant multi-source noise interference exists. Historical data from actual steering may contain assist noise or steering operation vibration noise. This noise can be misinterpreted by the ARIMA model as a valid trend signal, causing the predicted value to deviate significantly from the actual demand, potentially exacerbating assist fluctuations and steering wheel vibration. Therefore, reducing or avoiding the impact of noise on the prediction results to improve assist stability has become a pressing issue. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides an intelligent control system for electric steering gear based on distributed cooperative technology, the specific technical solution of which is as follows: One embodiment of the present invention provides an intelligent control system for electric power steering based on distributed cooperative technology, comprising: The data acquisition module is used to acquire the current steering data segment, the historical steering data segment, and the matching driving speed data segment of the historical steering data segment; The model parameter coefficient adjustment module is used to obtain similar clusters of the current steering data segment based on the normalized length difference between historical steering data segments and the ordinate difference of corresponding position data points between historical steering data segments; to obtain the stage stability index value of the historical steering data segment based on the ordinate difference between adjacent data points in the historical steering data segment and the corresponding matched driving speed data segment; to obtain the deviation degree of each steering assist data point in the current steering data segment based on the ordinate difference of corresponding position data points between the current steering data segment and the historical steering data segments in the similar cluster and the stage stability index value; and to adjust the parameter coefficients of the ARIMA model based on the deviation degree to obtain the target parameter coefficients of each steering assist data point in the current steering data segment. The assist value output module is used to predict the assist value at the next driving moment based on the current steering data segment, the ARIMA model and the target parameter coefficients, and to obtain the target steering assist value at the next driving moment based on the assist prediction value and the assist value to be adjusted at the next driving moment.

[0005] Beneficial effects: This invention includes a data acquisition module for acquiring the current steering data segment, historical steering data segments, and matching driving speed data segments of historical steering data segments; a model parameter coefficient adjustment module for obtaining similar clusters of the current steering data segment based on the normalized length differences between historical steering data segments and the ordinate differences of corresponding position data points between historical steering data segments; obtaining the stage stability index value of historical steering data segments based on the ordinate differences of corresponding position data points between the current steering data segment and historical steering data segments in similar clusters, and obtaining the deviation degree of each steering assist data point in the current steering data segment based on the ordinate differences of corresponding position data points between the current steering data segment and historical steering data segments in similar clusters and the stage stability index value; adjusting the parameter coefficients of the ARIMA model based on the deviation degree to obtain the target parameter coefficients of each steering assist data point in the current steering data segment; and an assist value output module for predicting the assist prediction value at the next driving moment based on the current steering data segment, the ARIMA model, and the target parameter coefficients; and obtaining the target steering assist value at the next driving moment based on the assist prediction value and the assist value to be adjusted at the next driving moment. Furthermore, this invention adjusts the parameter coefficients of the prediction model based on the degree of deviation, and uses the adjusted coefficients for prediction. This can reduce the interference of power assist noise or steering vibration noise caused by operation on the prediction results, thereby improving prediction accuracy and thus improving vehicle power assist stability, handling or control stability and driving safety. Attached Figure Description

[0006] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0007] Figure 1 This is a structural block diagram of an intelligent control system for electric steering gear based on distributed collaborative technology according to the present invention. Detailed Implementation

[0008] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0009] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0010] This embodiment provides an intelligent control system for electric power steering based on distributed cooperative technology, which is described in detail below: like Figure 1 As shown in the figure, this embodiment provides an intelligent control system for electric power steering based on distributed cooperative technology, including: The data acquisition module 01 is used to acquire the current steering data segment, the historical steering data segment, and the matching driving speed data segment of the historical steering data segment.

[0011] Because current predictive control algorithms predict future power assist demand and implement pre-loaded power assist, interference noise generated in actual road conditions can affect the prediction results, potentially exacerbating power assist fluctuations and steering wheel vibration, or leading to poor control of the electric power steering system. In other words, significant multi-source noise interference exists in actual road conditions or during vehicle steering, causing power assist noise or steering operation vibration noise. Factors such as motor current fluctuations, sensor sampling jitter, and road bumps can easily cause power assist noise or steering operation vibration noise. Furthermore, interference noise can be misinterpreted as a valid trend signal by the ARIMA model, causing the predicted value to deviate significantly from the actual demand, which may further exacerbate power assist fluctuations and steering wheel vibration. To improve power assist stability, this embodiment will subsequently adjust the parameter coefficients of the ARIMA model based on the probability that the historical data used for prediction is noise, thereby improving the reliability and accuracy of the prediction results and ultimately enhancing power assist stability or control stability.

[0012] This embodiment first uses distributed collaborative technology to acquire the driving speed and target steering assist value of any vehicle at different driving times. That is, the driving speed and steering torque in this embodiment are not acquired by a single sensor, but are synchronously sensed by multiple data acquisition nodes with different physical locations and functions, and then fused through the vehicle network. Steering torque is a key parameter for subsequently acquiring the assist value to be adjusted, and the assist value to be adjusted is a key parameter for determining the target steering assist value. The assist value to be adjusted at any driving time in this embodiment is the assist value calculated by EPS based on the driving speed and steering torque of the previous driving time. Essentially, it uses the known driving speed and steering torque as indexes to interpolate and look up values ​​in a pre-calibrated LUT (Look-Up Table) for the vehicle. LUT is a pre-calibrated LUT that uses data to look up values ​​in a pre-calibrated LUT. A two-dimensional data matrix is ​​calibrated and stored for the optimal power steering output value under different combinations of driving speed and steering torque. Generally, one vehicle model corresponds to one dedicated LUT. Steering torque refers to the torque applied by the driver to the steering wheel, and steering assist value refers to the additional torque provided by EPS to assist the driver. The process of determining the target steering assist value at any driving moment in this embodiment is consistent with the process of determining the target steering assist value at the next driving moment in the example given later in this embodiment. In specific applications, the implementer needs to set the acquisition frequency according to the actual situation such as the control cycle. For example, in this embodiment, the acquisition frequency can be set to 100 Hz, that is, the driving speed and the target steering assist value are acquired synchronously in this embodiment, and the acquisition frequency is 100 Hz, or the time interval between adjacent driving moments in this embodiment is 0.01 seconds.

[0013] Next, a two-dimensional space for driving speed and a two-dimensional space for steering assist value are constructed. In the driving speed space, the horizontal axis represents time and the vertical axis represents vehicle speed. In the steering assist value space, the horizontal axis represents time and the vertical axis represents steering assist value. Then, all driving speeds acquired by the vehicle within the current preset recent time period and their corresponding acquisition times are mapped to the driving speed space, and the mapped data points are recorded as speed data points. The speed data points in the driving speed space are then connected sequentially according to time, and the resulting curve is recorded as the driving speed curve corresponding to the current driving time. Similarly, all target steering assist values ​​acquired by the vehicle within the current preset recent time period and their corresponding acquisition times are mapped to the steering assist value space, and the mapped data points are recorded as steering assist data points. The steering assist data points in the steering assist value space are then connected sequentially according to time, and the resulting curve is recorded as the driving speed curve corresponding to the current driving time. The resulting curve is denoted as the steering assist curve corresponding to the current driving time. That is, at any driving time, there is a steering assist data point and a driving speed data point. The horizontal axis of the steering assist data point at that driving time is the time corresponding to that driving time, and the vertical axis is the target steering assist value at that driving time. The horizontal axis of the driving speed data point at that driving time is the time corresponding to that driving time, and the vertical axis is the vehicle driving speed at that driving time. In specific applications, the implementer needs to set the current preset recent time period according to the actual situation. For example, in this embodiment, the current driving time and the time period with a cumulative driving time of one day before the current driving time can be used as the current preset recent time period. If the cumulative driving time of the vehicle in the three days before the current driving time is one day, then the time period formed by the three days before the current driving time and the current driving time is the current preset recent time period. In this embodiment, the driving time is the time when the vehicle is in the driving stage.

[0014] Furthermore, since normal vehicle operation includes both non-steering straight-line driving and steering, and the EPS system is not involved in the non-steering portion, this embodiment, in order to more accurately predict the assistance demand of the electric power steering (EPS) system during steering, should focus on the driving data segment where the vehicle actually steering, rather than the entire data. Vehicle steering is usually a continuous process, represented as data segments in the collected data. Therefore, this embodiment needs to first extract the steering data segment from the steering assist curve. The specific extraction process is as follows: On the power steering curve, data points with non-zero power steering value or non-zero ordinate value are recorded as extracted data points. Data segments consisting of time-continuous extracted data points on the power steering curve are recorded as sub-data segments to be analyzed. For example, if points 1 to 4 on the power steering curve are all extracted data points with adjacent data points having continuous time, and points 8 to 20 are also extracted data points with adjacent data points having continuous time, then points 1 to 4 constitute one sub-data segment to be analyzed, and points 8 to 20 constitute another sub-data segment to be analyzed. Time continuity means that the times corresponding to adjacent data points are adjacent driving times. If the time span between points 1 and 2 on the power steering curve is greater than the time interval between adjacent driving times specified above, it indicates that there is an inconsistency between points 1 and 2. The driving phase consists of non-continuous data points; the first and second points are not temporally adjacent. The number of data points in the sub-data segment to be analyzed is at least one. Since the sub-data segment to be analyzed contains not only data segments generated by actual steering but also potentially very short noisy data segments, this embodiment, in order to minimize the impact of noise during subsequent prediction, will analyze the length of the obtained sub-data segment to identify noisy and steering data segments. Specifically, the total number of data points in each sub-data segment to be analyzed is recorded as the initial length of the corresponding sub-data segment. Then, the initial length of each sub-data segment to be analyzed is subjected to max-min normalization, and the result is recorded as the representative length value of the corresponding sub-data segment. The expression for the representative length value of any sub-data segment to be analyzed is: L is the initial length of the sub-data segment to be analyzed. This is the minimum of the initial lengths of all the sub-data segments to be analyzed. The maximum value among the initial lengths of all sub-data segments to be analyzed is used. Since the length of noise segments is extremely short and the length of data segments generated by real turns is relatively long, the above normalization can make the length representative value of data segments generated by real turns tend to 1 and the length representative value of noise data segments tend to 0. That is, the smaller the length representative value, the more likely it is to be a noise segment, and the larger it is, the more likely it is to be a data segment generated by real turns. Sub-data segments to be analyzed with a length representative value not less than the preset category judgment selection threshold are selected and recorded as turning data segments. That is, sub-data segments to be analyzed with a length representative value not less than the preset category judgment selection threshold are data segments generated by real turns, and sub-data segments to be analyzed with a length representative value less than the preset category judgment selection threshold are noise data segments. In specific applications, implementers can set the preset category judgment selection threshold according to the actual situation such as the value range of the length representative value. Since the above normalization can make the length representative value of data segments generated by real turns tend to 1 and the length representative value of noise data segments tend to 0, and the value range of the length representative value is from 0 to 1, this embodiment takes the midpoint of the value range of the length representative value, 0.5, as the preset category judgment selection threshold.

[0015] After filtering out the steering data segments, the steering data segments containing steering assist data points at the current driving time are obtained and recorded as the current steering data segment. That is, if the x-coordinate of the last data point in a steering data segment is the time corresponding to the current driving time, then that steering data segment is the current steering data segment. All steering data segments obtained above, except for the current steering data segment, are recorded as historical steering data segments. Subsequent adjustments to the model parameter coefficients of the data points in the current steering data segment are mainly based on the historical steering data segments. These parameter coefficients include the autoregressive coefficient (AR) and the moving average coefficient (MA). Furthermore, if none of the steering data segments contain the current driving time... If the steering assist data point is the current driving time, then the steering assist value to be adjusted at the next driving time will be directly used as the target steering assist value at the next driving time. The steering assist value to be adjusted at the next driving time is the assist value calculated by EPS based on the driving speed and steering torque of the steering wheel at the current driving time. In addition, there may be a situation where the current driving time is the start time of the current driving cycle. If the current driving time is the start time of the current driving cycle, then the initial assist value preset by the manufacturer for the vehicle will be directly used as the target assist value at the corresponding time. That is, the target assist value at the start time of each driving cycle in the current preset recent time period is the initial assist value preset by the manufacturer for the vehicle.

[0016] After obtaining the historical steering data segments, the matching driving speed data segments are determined based on the driving speed curve. Obtaining the matching driving speed data segments facilitates subsequent analysis of the contribution of the historical steering data segments to calculating the deviation of data points in the current steering data segment. The deviation is the basis for adjusting the parameter coefficients. The process of obtaining the matching driving speed data segments of the historical steering data segments is as follows: On the driving speed curve, obtain the curve segment that corresponds to the time period of each historical steering data segment, and record it as the matching driving speed data segment of the corresponding historical steering data segment; that is, if the time period corresponding to a certain historical steering data segment is from time a0 to time an, then the curve segment formed by all speed data points on the driving speed curve whose horizontal coordinate is located in the period from time a0 to time an is the matching driving speed data segment of that historical steering data segment.

[0017] Therefore, this embodiment obtains the current steering data segment, the historical steering data segment, and the matching driving speed data segment of the historical steering data segment through the above process.

[0018] The model parameter coefficient adjustment module 02 is used to obtain similar clusters of the current steering data segment based on the normalized length difference between historical steering data segments and the ordinate difference of corresponding position data points between historical steering data segments; to obtain the stage stability index value of the historical steering data segment based on the ordinate difference between adjacent data points in the historical steering data segment and the corresponding matched driving speed data segment; to obtain the deviation degree of each steering assist data point in the current steering data segment based on the ordinate difference of corresponding position data points between the current steering data segment and the historical steering data segments in the similar cluster and the stage stability index value; and to adjust the parameter coefficients of the ARIMA model based on the deviation degree to obtain the target parameter coefficients of each steering assist data point in the current steering data segment.

[0019] During vehicle operation, various steering situations can occur. Drivers have their own driving habits, and their actions under the same steering conditions tend to be similar. Therefore, similar steering habits are also similar. Thus, deviation analysis can be performed on real-time steering data using similar steering habit characteristics. When the difference between a data point in a real-time steering data segment and a data point in a historical steering data segment with the most similar steering characteristics is significant, it indicates that the data point has abnormal noise characteristics unrelated to the driver's driving habits. Subsequent predictions need to reduce the impact of these data points on the prediction results. In other words, the current steering data segment may still contain data points affected by noise; to ensure prediction accuracy, subsequent predictions need to reduce the impact of these data points on the prediction results. Based on the above, this embodiment... The next step is to cluster the historical turning data segments, grouping segments with similar turning characteristics into one class. Since the similar lengths of the data segments and the difference in the ordinates of corresponding data points within the segments can reflect the similarity of turning characteristics, this embodiment will first obtain the metric distance between historical turning data segments based on the differences in the representative length values ​​and the differences in the ordinates of corresponding data points. The metric distance can measure the similarity or difference in turning characteristics between corresponding historical turning data segments. Then, K-means clustering is performed on all historical turning data segments based on the metric distance between them, and the resulting clusters are all recorded as data segment clusters. The metric distance calculated here is equivalent to the Euclidean distance between samples during clustering. The number of cluster centers is obtained using the elbow method.

[0020] For ease of understanding, this embodiment will subsequently describe the process of obtaining the metric distance between historical steering data segment A1 and historical steering data segment A2 as an example. Historical steering data segment A1 and historical steering data segment A2 belong to the historical steering data segments obtained above, but are not the same data segment; therefore, the specific process of obtaining the metric distance between historical steering data segment A1 and historical steering data segment A2 is as follows: The absolute value of the difference between the length representative value of historical steering data segment A1 and the length representative value of historical steering data segment A2 is denoted as the first distance value; the normalized result of the mean of the absolute values ​​of the differences in the ordinates between data points with the same position in historical steering data segment A1 and historical steering data segment A2 is denoted as the second distance value; the product of the first distance value and the second distance value is denoted as the metric distance between historical steering data segment A1 and historical steering data segment A2; the specific expression for the metric distance between historical steering data segment A1 and historical steering data segment A2 is as follows:

[0021] in, This is the distance metric between historical turning data segment A1 and historical turning data segment A2. This represents the length of the historical redirection data segment A1. The length of the historical data segment A2 is represented by Norm(), which is the normalization function. Let min(NA1, NA2), where min() is the minimum function, NA1 is the total number of data points in historical turning data segment A1, and NA2 is the total number of data points in historical turning data segment A2. This represents the ordinate value of the g-th data point in the historical turning data segment A1. This represents the ordinate value of the g-th data point in the historical turning data segment A2; The smaller the value, the closer the lengths of historical steering data segment A1 and historical steering data segment A2 are, and the more similar the steering characteristics between them. The smaller the value, the closer the ordinates are between data points in the same position in historical steering data segment A1 and historical steering data segment A2, and the more similar the steering characteristics are between historical steering data segment A1 and historical steering data segment A2. smaller and The smaller, The smaller, therefore The smaller the number of segments, the more similar the turning characteristics are between historical turning data segment A1 and historical turning data segment A2, and the easier it is for them to be clustered into one class during subsequent clustering.

[0022] After clustering, the metric distance between the current turning data segment and each historical turning data segment in each data segment cluster is calculated. In this embodiment, the calculation process of the metric distance between any two different data segments is the same as the calculation process of the metric distance between historical turning data segment A1 and historical turning data segment A2. Based on the metric distance between the current turning data segment and each historical turning data segment in each data segment cluster, the clusters most similar to the turning features of the current turning data segment are selected, that is, the similar clusters of the current turning data segment are determined. The specific process of determining the similar clusters of the current turning data segment is as follows: Calculate the metric distance between the current steering data segment and the historical steering data segments in each data segment cluster. The average metric distance between the current steering data segment and each historical steering data segment in each data segment cluster is denoted as the difference index value of the corresponding data segment cluster. That is, the difference index value of any data segment cluster is the average metric distance between the current steering data segment and each historical steering data segment in that cluster. Select the data segment cluster corresponding to the smallest difference index value as the similarity cluster of the current steering data segment. In other words, the smaller the difference index value, the more similar the steering characteristics of the historical steering data segments in that cluster are to the steering characteristics of the current steering data segment. Alternatively, the data segments in the similar cluster are more similar to the steering characteristics of the current steering data segment compared to the data segments in other clusters. Subsequently, the deviation characteristics of each data point in the current steering data segment will be measured based on the determined similarity clusters.

[0023] However, among all historical steering data segments in a similar cluster, not all are conventional steering data segments; some may be unconventional. Unconventional data segments cannot reflect the driver's steering behavior or habits. These unconventional segments are historical steering data segments that do not reflect the driver's steering behavior or habits, and are typically caused by sensor noise, sudden operations, or system malfunctions. They lack representativeness of real driving habits. Therefore, if unconventional historical steering data segments are treated the same as conventional steering data segments when calculating deviation, it will affect the subsequent judgment of the deviation degree of data points in the current steering data segment. Furthermore, the power assist stability and driving stability of the time period corresponding to the historical steering data segment are also affected. This can reflect the possibility that the corresponding data segment is a regular data segment. Furthermore, the difference in the ordinate between adjacent data points in the historical steering data segment and its corresponding matched driving speed data segment can reflect the power assist stability and driving stability of the time period corresponding to the historical steering data segment. Therefore, before calculating the degree of deviation, this embodiment will first obtain the stage stability index value of each historical steering data segment in the similar cluster based on the difference in the ordinate between adjacent data points in the matched driving speed data segment of each historical steering data segment in the similar cluster. That is, the stage stability index value of the historical steering data segment can reflect the power assist stability and driving stability of the time period corresponding to the historical steering data segment. The specific calculation process for the stage stability index value of any historical steering data segment in the similar cluster is as follows: The neighborhood difference of a data point is obtained by calculating the absolute value of the difference in the x-coordinate between a data point and its neighboring data points in a data segment. For the b-th data point in any data segment, if b is not equal to 1 and B, where B is the total number of data points in the data segment, the neighborhood difference of the b-th data point is the average of the first difference and the second difference of the b-th data point in the data segment. The first difference of the b-th data point is the absolute value of the difference in the y-coordinate between the b-th data point and the (b+1)-th data point in the data segment, and the second difference of the b-th data point is the absolute value of the difference in the y-coordinate between the b-th data point and the (b+1)-th data point in the data segment. The absolute value of the difference in the ordinate between data points is used. If b equals 1, the neighborhood difference of the b-th data point is the absolute value of the difference in the ordinate between the b-th and (b+1)-th data points in the data segment. If b equals B, the neighborhood difference of the b-th data point is the absolute value of the difference in the ordinate between the b-th and (b-1)-th data points in the data segment. The result of negatively correlated mapping of the mean neighborhood difference of all data points in the historical steering data segment is denoted as the first measure of the historical steering data segment. The neighborhood difference of all data points in the matching driving speed data segment of the historical steering data segment is... The negative correlation mapping result of the mean difference is denoted as the second measure value of this historical turning data segment. At this point, a negative exponential function with a base of constant e is used for the negative correlation mapping, i.e., the first and second measure values ​​are exp(-H1) and exp(-H2) respectively, where exp() is an exponential function with a base of constant e, H1 is the first measure value, and H2 is the second measure value. The sum of the first and second measure values ​​of this historical turning data segment is denoted as the stage stability index value of this historical turning data segment. The larger the first measure value, the more stable the assist value of this historical turning data segment. The larger the value, the more stable the driving speed is within the same time period as the historical steering data segment. The larger the first and second values, the larger the stage stability index value of the historical steering data segment. The larger the stage stability index value of the historical steering data segment, the more stable the assist value and the driving speed within the same time period are. In this case, the historical steering data segment is better able to represent steering habits. Subsequent calculations of deviation should increase the participation of the historical steering data segment to ensure the reliability of the calculated deviation.

[0024] After calculating the stage stability index values ​​of each historical steering data segment in the similar cluster, the deviation degree of each steering assist data point in the current steering data segment is obtained based on the difference in the ordinates of the corresponding position data points between the current steering data segment and the historical steering data segments in the similar cluster, as well as the stage stability index values ​​of the historical steering data segments in the similar cluster. The deviation degree can reflect the degree of noise interference to the corresponding steering assist data point or its reliability when used for prediction. The specific process for obtaining the deviation degree of the i-th steering assist data point in the current steering data segment is as follows: The data segments in the similar clusters are all similar historical steering data segments. The set of all steering assist data points in all similar historical steering data segments whose positions are the same as those of the i-th steering assist data point in the current steering data segment is denoted as the reference set of the i-th steering assist data point. That is, the data points in the reference set of the i-th steering assist data point belong to similar historical steering data segments, and their positions in the similar historical steering data segments are the same as those of the i-th steering assist data point in the current steering data segment. The feature values ​​of each steering assist data point in the reference set of the i-th steering assist data point are obtained. The characteristic value of the j-th steering assist data point in the set is the result of multiplying the absolute value of the difference in the ordinate between the i-th and j-th steering assist data points by the stage stability index value of the historical steering data segment to which the j-th steering assist data point belongs. The normalized result of the mean of the characteristic values ​​of all steering assist data points in the reference set of the i-th steering assist data point is denoted as the deviation degree of the i-th steering assist data point. Here, the maximum and minimum normalization function Norm() is used for normalization, that is, normalization is performed based on the maximum and minimum deviation degrees of all steering assist data points in the current steering data segment. The expression for the deviation degree of the i-th steering assist data point is:

[0025] in, The deviation of the i-th steering assist data point. Let i be the number of data points in the reference set for the i-th steering assist data point. Let be the ordinate value of the i-th steering assist data point. Let j be the ordinate value of the steering assist data point in the reference set of the i-th steering assist data point. Let be the stage stability index value of the historical steering data segment to which the j-th steering assist data point belongs; and The larger and A larger value indicates a higher degree of deviation for the i-th steering assist data point, which also suggests a greater likelihood of noise interference. Subsequent predictions require reducing the impact of this data point on the prediction results, i.e., significantly reducing the parameter coefficients. Furthermore, referencing the stage stability index value when measuring deviation allows data segments with high stage stability to contribute more significantly to the deviation, making the calculation results more reflective of the noise level or better characterize the noise level of the i-th steering assist data point.

[0026] Next, the parameter coefficients of the ARIMA model are adjusted based on the deviation of each steering assist data point in the current steering data segment to obtain the target parameter coefficients for each steering assist data point in the current steering data segment. The specific adjustment process is as follows: the product of the negative correlation mapping result of the deviation of each steering assist data point in the current steering data segment and the original parameter coefficient of the corresponding steering assist data point is denoted as the target parameter coefficient of the corresponding steering assist data point. Here, the negative correlation mapping result of the deviation of the steering assist data point is a constant 1 minus the deviation of that steering assist data point. That is, the expression for the target parameter coefficient of the i-th steering assist data point is: , The original parameter coefficients for the i-th steering assist data point are used. The above adjustments reduce the parameter coefficients of points with more noise characteristics, thereby decreasing the impact of noise on the prediction results or minimizing interference from assist noise or steering jitter noise caused by operation, resulting in more accurate predictions. Furthermore, the original parameter coefficients in this embodiment are model parameter coefficients calculated using traditional methods, such as maximum likelihood estimation or conditional least squares.

[0027] The assist value output module 03 is used to predict the assist prediction value at the next driving moment based on the current steering data segment, the ARIMA model and the target parameter coefficients, and to obtain the target steering assist value at the next driving moment based on the assist prediction value and the assist value to be adjusted at the next driving moment.

[0028] After adjusting the parameter coefficients, the predicted steering assist value for the next driving moment is predicted based on the current steering data segment, the ARIMA model, and the target parameter coefficients of each steering assist data point in the current steering data segment. That is, the input to the model is the current steering data segment. Given the input and the parameter coefficients of the data points in the input, the process of predicting future data based on the ARIMA model is a well-known technique. Then, based on the predicted steering assist value for the next driving moment and the steering assist value to be adjusted for the next driving moment, the target steering assist value for the next driving moment is obtained, and the target steering assist value for the next driving moment is used as the final output of EPS. The target steering assist value for the next driving moment is the average of the predicted steering assist value for the next driving moment and the steering assist value to be adjusted for the next driving moment. The steering assist value to be adjusted for the next driving moment is the assist value calculated by EPS based on the vehicle's driving speed and the steering torque of the steering wheel at the current driving moment. After obtaining the target steering assist value for the next driving moment, the system relies on the actuator in the vehicle's electric power steering system to assist the driver in completing the steering operation by precisely controlling the output torque of the motor. In this embodiment, the target steering assist value for the next driving moment mainly covers the control cycle between the next driving moment and the next driving moment after that. The target steering assist value for the next driving moment covers the time range [t+1, t+2), where t is the current driving moment.

[0029] Thus, this embodiment completes the intelligent control of the electric power steering system. Furthermore, by reducing power steering noise or steering vibration noise caused by operation and minimizing its interference with prediction, this embodiment improves power steering stability, control stability, and driving safety. Additionally, in this embodiment, only the specific numerical values ​​of the parameters involved in the calculations are considered; dimensions are not considered.

[0030] In summary, this embodiment includes a data acquisition module for acquiring the current steering data segment, historical steering data segments, and matching driving speed data segments of historical steering data segments; a model parameter coefficient adjustment module for obtaining similar clusters of the current steering data segment based on the normalized length differences between historical steering data segments and the ordinate differences of corresponding position data points between historical steering data segments; obtaining the stage stability index value of historical steering data segments based on the ordinate differences of corresponding position data points between the current steering data segment and historical steering data segments in similar clusters, and obtaining the deviation degree of each steering assist data point in the current steering data segment based on the ordinate differences of corresponding position data points between the current steering data segment and historical steering data segments in similar clusters and the stage stability index value; adjusting the parameter coefficients of the ARIMA model based on the deviation degree to obtain the target parameter coefficients of each steering assist data point in the current steering data segment; and an assist value output module for predicting the assist prediction value at the next driving moment based on the current steering data segment, the ARIMA model, and the target parameter coefficients; and obtaining the target steering assist value at the next driving moment based on the assist prediction value and the assist value to be adjusted at the next driving moment. Furthermore, this embodiment adjusts the parameter coefficients of the prediction model based on the degree of deviation, and uses the adjusted coefficients for prediction. This can reduce the interference of power assist noise or steering vibration noise caused by operation on the prediction results, thereby improving prediction accuracy and thus improving vehicle power assist stability, handling or control stability and driving safety.

[0031] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An intelligent control system for electric power steering based on distributed cooperative technology, characterized in that, The intelligent control system for the electric power steering includes: The data acquisition module is used to acquire the current steering data segment, the historical steering data segment, and the matching driving speed data segment of the historical steering data segment; The model parameter coefficient adjustment module is used to obtain similar clusters of the current steering data segment based on the normalized length difference between historical steering data segments and the ordinate difference of corresponding position data points between historical steering data segments; to obtain the stage stability index value of the historical steering data segment based on the ordinate difference between adjacent data points in the historical steering data segment and the corresponding matched driving speed data segment; to obtain the deviation degree of each steering assist data point in the current steering data segment based on the ordinate difference of corresponding position data points between the current steering data segment and the historical steering data segments in the similar cluster and the stage stability index value; and to adjust the parameter coefficients of the ARIMA model based on the deviation degree to obtain the target parameter coefficients of each steering assist data point in the current steering data segment. The assist value output module is used to predict the assist value at the next driving moment based on the current steering data segment, the ARIMA model and the target parameter coefficients, and to obtain the target steering assist value at the next driving moment based on the assist prediction value and the assist value to be adjusted at the next driving moment.

2. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, The x-axis of the steering assist data points in the steering data segment represents time, and the y-axis represents the target steering assist value. The x-axis of the speed data points in the driving speed data segment represents time, and the y-axis represents driving speed.

3. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, The methods for obtaining the current steering data segment, the historical steering data segment, and the matching driving speed data segment of the historical steering data segment include: The curves formed by the continuous data points of the vehicle's driving speed and target steering assist value acquired within the current preset recent time period are respectively denoted as the driving speed curve and the steering assist curve. The continuous data point segments with non-zero assist values ​​in the steering assist curve are extracted and denoted as sub-data segments to be analyzed. The steering data segments are selected based on the normalized length of each sub-data segment to be analyzed. The steering data segment containing the current driving time is denoted as the current steering data segment, and all steering data segments other than the current steering data segment are denoted as historical steering data segments. On the driving speed curve, obtain the curve segment that corresponds to the time period of each historical steering data segment, and record it as the matching driving speed data segment of the corresponding historical steering data segment.

4. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 3, characterized in that, Methods for obtaining the redirection data segment include: The result of performing maximum and minimum normalization on the length of each sub-data segment to be analyzed is recorded as the representative length value of the corresponding sub-data segment to be analyzed. Sub-data segments to be analyzed whose representative length value is not less than the preset category selection threshold are all recorded as turning data segments.

5. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, The methods for obtaining similar clusters in the current redirected data segment include: Based on the normalized length difference between historical steering data segments and the ordinate difference of corresponding position data points between historical steering data segments, the metric distance between historical steering data segments is obtained. Based on the metric distance, all historical steering data segments are clustered to obtain data segment clusters. The mean of the metric distance between the current steering data segment and the historical steering data segments in each data segment cluster is recorded as the difference index value of the corresponding data segment cluster. The data segment cluster corresponding to the smallest difference index value is selected as the similarity cluster of the current steering data segment.

6. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 5, characterized in that, Methods for obtaining the metric distance between historical transition data segments include: For historical steering data segments A1 and A2: the absolute value of the difference between the normalized result of the data segment length of historical steering data segment A1 and the normalized result of the data segment length of historical steering data segment A2 is recorded as the first distance value; the normalized result of the mean of the absolute values ​​of the differences in the ordinates of data points with the same position between historical steering data segments A1 and A2 is recorded as the second distance value; and the product of the first distance value and the second distance value is recorded as the metric distance between historical steering data segments A1 and A2.

7. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, Methods for obtaining the stage stability index values ​​of historical transition data segments include: The result of negatively correlated mapping of the mean neighborhood difference of all data points in the historical steering data segment is denoted as the first measure of the corresponding historical steering data segment. The result of negatively correlated mapping of the mean neighborhood difference of all data points in the matched driving speed data segment of the historical steering data segment is denoted as the second measure of the corresponding historical steering data segment. The absolute value of the difference in the abscissa between a data point in a data segment and its neighboring data points is the neighborhood difference of the corresponding data point. The sum of the first measure and the second measure of each historical steering data segment is denoted as the stage stability index value of the corresponding historical steering data segment.

8. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, Methods for obtaining the degree of deviation include: For the i-th steering assist data point in the current steering data segment, the set of all steering assist data points in similar historical steering data segments that are in the same position as the i-th steering assist data point in the current steering data segment is denoted as the reference set of the i-th steering assist data point. The data segments in the similar cluster are all similar historical steering data segments. The normalized result of the mean of the feature values ​​of all steering assist data points in the reference set of the i-th steering assist data point is denoted as the deviation degree of the i-th steering assist data point. The feature value of the j-th steering assist data point in the reference set is the result of multiplying the absolute value of the difference between the ordinates of the i-th and j-th steering assist data points by the stage stability index value of the historical steering data segment to which the j-th steering assist data point belongs.

9. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, The methods for obtaining the target parameter coefficients for each steering assist data point in the current steering data segment include: The product of the negative correlation mapping result of the deviation degree of each steering assist data point in the current steering data segment and the original parameter coefficient of the corresponding steering assist data point is denoted as the target parameter coefficient of the corresponding steering assist data point.

10. The intelligent control system for electric steering gear based on distributed cooperative technology as described in claim 1, characterized in that, The target steering assist value at the next driving moment is the average of the predicted assist value and the assist value to be adjusted at the next driving moment. The assist value to be adjusted at the next driving moment is the assist value calculated by EPS based on the driving speed and steering torque of the steering wheel at the current driving moment.

Images