Method for improving torque smoothness in man-machine co-driving state for EPS response intelligent driving

Abstract

The invention relates to the technical field of automobile direction control, and discloses a method for improving torque smoothness in a man-machine co-driving state aiming at EPS response intelligent driving, which comprises the following steps: acquiring vehicle sensing data and analyzing the vehicle sensing data into working condition information, performing interval division on a vehicle speed signal under a unified outer ring sampling caliber to obtain vehicle speed interval information, and outputting the vehicle speed interval information; the method comprises the following steps: acquiring corner target information, constructing deviation for the corner target information to obtain corner deviation information, jointly describing vehicle speed and target change, executing outer ring gain mapping based on a vehicle speed interval to form a gain space, performing two-dimensional table look-up by the corner deviation information to obtain a gain set, and providing adjustable gain for a two-stage closed-loop outer ring under the condition of not changing hardware and an inner ring structure. Closed-loop adjustment is carried out on the working conditions according to the gain set, and turning angle control information is output, so that an outer ring selects gains in real time along with the vehicle speed interval and the deviation level, response lag and dynamic overshoot are reduced, responsiveness and stability under all the auxiliary driving working conditions are kept, and man-machine co-driving torque smoothness is improved.

Description

Technical Field

[0001] This invention relates to the field of automotive steering control technology, and more specifically, to a method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions. Background Technology

[0002] As the actuator and signal interface for lateral control of a vehicle, the electric power steering system is typically deployed on limited controller hardware and on-board networks. It is constrained by factors such as fixed control cycles, limited computing and storage resources, and limited communication latency and bandwidth. In existing technologies, in order to achieve stable tracking according to a given steering angle or torque target during vehicle operation, abstract control strategies such as multi-stage closed-loop regulation, state quantity feedback shaping, characteristic curve mapping, parameter tuning and amplitude limiting are commonly used. These technical solutions can generally achieve relatively stable control effects under the premise that the target operating condition range is relatively concentrated, the vehicle speed change is relatively gentle, the steering target change rate is limited, and the level of environmental disturbance is predictable.

[0003] As assisted driving functions based on trajectory planning and environmental perception are gradually integrated into the same electric power steering system, the vehicle speed continuously changes within a wide range when switching between different functions such as low-speed parking, medium-speed following, and high-speed cruising. The steering system also needs to track the angle target information with large differences in amplitude and rate of change during the steering initiation, steering holding, and return-to-center stages. This time-varying vehicle speed and angle target information works together on the control structure that adopts multi-stage closed-loop adjustment and fixed characteristic curve mapping. This makes it possible for the two-level outer loop control parameters tuned under a certain operating condition to produce response lag or dynamic overshoot under other operating conditions. Therefore, it is difficult to maintain the uniformity of steering dynamic characteristics under various assisted driving conditions without changing the existing hardware and inner loop current control structure. Accordingly, the technical problem that needs to be solved is how to achieve the responsiveness and stability of the two-level closed-loop outer loop to the angle target information under various assisted driving conditions without changing the hardware and inner loop structure of the electric power steering system, and under the condition that the vehicle speed and angle target information continuously change with the assisted driving conditions.

[0004] In view of this, the present invention proposes a method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for improving torque smoothness in EPS response intelligent driving under human-machine co-driving conditions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: Firstly, a method is provided to improve torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions, including: Acquire vehicle sensor data, perform operating condition analysis processing on the vehicle sensor data to obtain operating condition information. The vehicle sensor data includes vehicle speed signal and steering angle target information. The operating condition information is used to characterize the sampling state of the outer ring of the electric power steering system. Based on the operating condition information, the vehicle speed signal is divided into intervals to obtain vehicle speed interval information. Based on the operating condition information, the turning angle target information is deviated to obtain turning angle deviation information. The vehicle speed interval information and turning angle deviation information are used to describe the joint change characteristics of vehicle speed and turning angle target information under different assisted driving conditions. Based on the vehicle speed range information, outer loop gain mapping is performed to obtain gain space information. Based on the steering angle deviation information, two-dimensional lookup table processing is performed on the gain space information to obtain gain set information. The gain space information and gain set information are used to provide adjustable gain parameters for the two-level closed-loop outer loop without changing the hardware and inner loop structure of the electric power steering system. Based on the gain set information, closed-loop regulation is performed on the operating condition information to obtain the angle control information.

[0007] In the above scheme, the method for performing deviation construction processing on the corner target information based on the working condition information to obtain the corner deviation information includes: Based on the working condition information, the command angle information and actual angle information corresponding to the current working condition are selected from the angle target information. The command angle information and actual angle information are synchronously processed in chronological order to obtain the sampling sequence information. Extract the time position marker of each sampling moment from the sampling sequence information, and perform sequence alignment processing on the sampling sequence information according to the time position marker to obtain the aligned sequence information; Based on the alignment sequence information, the difference sample between the commanded angle information and the actual angle information is calculated at each sampling time. The difference sample is then aggregated and processed according to the time position mark to obtain the angle deviation information.

[0008] In the above scheme, the method for performing aggregation calculations on the difference samples according to time position marks to obtain the corner deviation information includes: Segmentation marking is performed on each sampling time based on the difference samples and time position markers to obtain segmentation marking information. Segmentation aggregation is then performed on the time position markers based on the segmentation marking information to obtain time segmentation information. The difference samples are segmented and aggregated based on the segmentation labeling information to obtain segmented sample information. The deviation statistics are then performed on each time segment based on the segmented sample information to obtain segmented deviation information. The segmented deviation information is weighted and aggregated based on the time segmentation information to obtain aggregated deviation information. The sampled sequence information is then subjected to deviation characterization processing based on the aggregated deviation information to obtain corner deviation information.

[0009] In the above scheme, the method for performing weighted summarization processing on the segmented deviation information based on the time segmentation information to obtain the summarized deviation information includes: Based on the time segmentation information, the segmented deviation information is processed by sample expansion according to the time segmentation and sampling order to obtain a sample record set. Each record in the sample record set corresponds to a time index and the deviation amplitude corresponding to the time index. Based on the time segmentation information, the relative position calculation is performed on the time index in the sample record set to obtain a position sequence. Each position value in the position sequence is used to characterize the normalized time position of the corresponding sample record within its respective time segment. Perform weight generation processing on the position sequence to obtain a weight sequence. Each weight value in the weight sequence is a monotonic function output of the corresponding position value in the position sequence. The deviation amplitude in the sample record set is processed by performing sample product processing according to the weight value in the weight sequence to obtain a weighted record set. The weighted record set is then processed by performing segmented accumulation processing according to time segment information to obtain the summary deviation information.

[0010] In the above scheme, the method for performing a two-dimensional lookup table process on the gain space information based on the rotation deviation information to obtain the gain set information includes: Based on the rotation deviation information, perform mesh selection processing on the two-dimensional nodes in the gain space information to obtain lookup table mesh information. Based on the lookup table mesh information, perform coordinate mapping processing on the rotation deviation information to obtain mesh index information. Based on the grid index information, interpolation calculation is performed on the node gain values ​​in the lookup table grid information to obtain the initial gain information. Based on the corner deviation information, the initial gain information is then processed by dimension adjustment to obtain the gain set information.

[0011] In the above scheme, the method for obtaining the initial gain information by performing interpolation calculations on the node gain values ​​in the lookup table grid information based on the grid index information includes: Based on the grid index information and the lookup table grid information, a neighbor filtering process is performed on the position of the two-dimensional grid node corresponding to each grid index to obtain the neighbor index information. Each record in the neighbor index information associates a grid index with the position indexes of several adjacent nodes. Based on the lookup grid information, the node gain values ​​corresponding to the neighbor index information are subjected to gain aggregation processing to obtain neighbor gain information. The neighbor gain information is aggregated according to each grid index to form an ordered gain set information by aggregating the node gain values ​​of adjacent nodes. Based on the neighbor gain information and grid index information, interpolation weight calculation is performed on each grid index to obtain weight set information. The relative position of each set of weights in the weight set information and the corresponding grid index in the two-dimensional grid maintains a monotonic relationship. The node gain values ​​in the neighbor gain information are weighted and combined according to the weights in the weight set information to obtain the initial gain information. The initial gain information provides the interpolated gain value after a two-dimensional lookup table at each grid index.

[0012] In the above scheme, the method for performing weighted combination processing on the node gain values ​​in the neighbor gain information according to the weights in the weight set information to obtain the initial gain information includes: For each node gain value in the neighbor gain information, perform sample expansion processing according to the record order to obtain the gain sample set; for each weight value in the weight set information, perform sample expansion processing according to the record order to obtain the weight sample set. The gain value of each node in the gain sample set is multiplied with the corresponding weight value in the weight sample set to obtain the product sample set. The product sample set represents the product result of the gain value of each node and the corresponding weight value at the sample level. The initial gain information is obtained by performing a summation calculation on all product results in the product sample set.

[0013] In the above scheme, the method for obtaining steering angle control information by performing closed-loop regulation processing on the operating condition information based on the gain set information includes: The rotation angle target information and angular velocity target information in the working condition information are processed synchronously to obtain the loop target information. The loop target information provides the corresponding rotation angle target quantity and angular velocity target quantity at each sampling position. The angle gain sample and speed gain sample in the gain set information are processed by loop allocation according to the sampling position to obtain loop gain information. The loop gain information associates the angle gain sample and speed gain sample with the rotation angle target quantity and angular velocity target quantity in the loop target information at each sampling position. For the target values ​​of rotation angle and angular velocity in the loop target information, closed-loop calculation processing is performed according to the angle gain sample and speed gain sample in the loop gain information to obtain the rotation angle control information.

[0014] In the above scheme, the method for obtaining the steering angle control information by performing closed-loop calculation processing on the steering angle target quantity and angular velocity target quantity in the loop target information according to the angle gain sample and speed gain sample in the loop gain information includes: Based on the loop gain information, the angle target quantity and the angle gain sample are processed by angle loop calculation to obtain the angle loop information. The angle loop information at each sampling position represents the angle loop output quantity after proportional adjustment of the angle target quantity based on the angle gain sample. Based on the loop target information, the angular velocity target quantity and the speed gain sample are processed by speed loop calculation to obtain speed loop information. The speed loop information at each sampling position represents the speed loop output quantity after proportional-integral adjustment of the angular velocity target quantity based on the speed gain sample. The angle loop output in the angle loop information and the speed loop output in the speed loop information are combined in a loop to obtain the angle control information. The angle control information is used as the outer loop target input of the inner loop current control structure at each sampling position.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention acquires vehicle sensor data and performs operating condition analysis to obtain operating condition information, thus clarifying the outer ring sampling state under a unified standard. Based on this, the vehicle speed signal is divided into intervals to obtain vehicle speed interval information, and the cornering target information is deviated to obtain cornering deviation information. The changing characteristics of vehicle speed and cornering target information are jointly characterized over time. Under this characterization, outer ring gain mapping is performed based on the vehicle speed interval information to form gain space information. Then, a two-dimensional lookup table is performed on the gain space information based on the cornering deviation information to obtain gain set information. Therefore, without changing the electric power steering... Under the premise of the steering system hardware and inner loop structure, adjustable gain parameters are provided for the two-level closed-loop outer loop. Finally, the working condition information is adjusted in a closed loop based on the gain set information, and the steering angle control information is output. This allows the two-level closed-loop outer loop to select the gain in real time according to the vehicle speed range and deviation level. When switching between different working conditions such as low-speed parking, medium-speed following, and high-speed cruising, the difference in amplitude and rate of change of the steering angle target information is absorbed through the gain set information, reducing response lag and dynamic overshoot, maintaining responsiveness and stability to the steering angle target information, and improving torque smoothness in human-machine co-driving mode. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for improving torque smoothness in human-machine co-driving mode for EPS-responsive intelligent driving according to the present invention. Figure 2 This is a schematic diagram of the structure of an electronic device according to the present invention. Detailed Implementation

[0017] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the invention. In the following detailed description, many specific details are set forth for ease of explanation to provide a full understanding of the embodiments of the present invention. However, it is obvious that one or more embodiments may be implemented without these specific details. Furthermore, in the following description, descriptions of well-known structures and techniques are omitted to avoid unnecessarily obscuring the concept of the present invention.

[0018] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0019] In the technical solution of this invention, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data are all carried out in accordance with relevant laws, regulations, and standards, and necessary confidentiality measures have been taken. They do not violate public order and good morals, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0020] Example 1 Please see Figure 1 As shown, this embodiment discloses a method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions, including: S10: Acquire vehicle sensor data, perform operating condition analysis processing on the vehicle sensor data to obtain operating condition information. The vehicle sensor data includes vehicle speed signal and steering angle target information. The operating condition information is used to characterize the sampling state of the outer ring of the electric power steering system. In this embodiment, vehicle sensing data refers to the set of signals collected by the vehicle speed sensor and steering-related control unit and output through the vehicle network within the outer loop control cycle of the electric power steering system. The vehicle speed signal reflects the longitudinal speed of the vehicle at the current sampling moment. The turning angle target information is usually composed of the command turning angle signal given by the intelligent driving control unit or the vehicle control unit, which is used to characterize the target turning angle that the outer loop wants the electric power steering system to track at this sampling moment. The operating condition information is the state data obtained by organizing the above vehicle speed signal and turning angle target information into a structured record at discrete sampling moments. It is used to uniformly characterize the vehicle driving condition and turning target state corresponding to each sampling moment in the subsequent PID outer loop control calculation of the turning angle closed loop and speed closed loop.

[0021] It should be noted that performing operating condition analysis on vehicle sensor data refers to the process of establishing an operating condition record based on vehicle speed signal and steering angle target information within each outer loop sampling cycle. This process typically includes extracting the amplitude of the vehicle speed signal and the amplitude of the commanded steering angle signal at the current sampling moment, combining them with the time stamp of the current outer loop sampling cycle, and optionally adding a current driver assistance function type stamp, thereby obtaining operating condition information including vehicle speed field, commanded steering angle field, and time field. The purpose of this processing is to align the vehicle speed signal and commanded steering angle signal, which originally existed with different refresh cycles and different signal formats, to a unified outer loop sampling cycle, so that subsequent construction of angle deviation value (AngErr), execution of two-dimensional lookup table based on vehicle speed signal and angle deviation value (AngErr), and gain adjustment in the outer loop of the two-level PID control in steering angle closed loop and speed closed loop can all be directly calculated based on the same set of operating condition information.

[0022] Vehicle speed, commanded turning angle, and actual turning angle are aligned to the sampling time point of the outer ring according to the timestamp. If the refresh frequency of a certain signal is inconsistent with that of the outer ring, the change between two adjacent samples is transitioned to the target sampling time in a linear manner. If the original timestamp differs from the target sampling time by more than one and a half sampling periods, it is considered that the data at that time is missing. The most recent valid value is temporarily retained and marked as "missing" in the record. Such samples can be ignored in subsequent statistics and weighting.

[0023] For example, assuming the outer ring of the electric power steering system has a sampling period of 10ms, the vehicle speed signals collected in four consecutive sampling periods are 2km / h, 3km / h, 4km / h, and 5km / h, respectively. The corresponding command angle signals in the sampling periods are 5°, 10°, 15°, and 20°, respectively. During the working condition analysis process, the vehicle speed signals and command angle signals in these four sampling periods are combined into four working condition information records in chronological order: the first record is time index 1, vehicle speed 2, and command angle 5; the second record is time index 2, vehicle speed 3, and command angle 10; and so on. This working condition information sequence can be understood in form as a list structure containing multiple rows of records, with each row of records binding to the vehicle speed signal and command angle signal at the sampling time of the outer ring. This structured working condition information will be used in subsequent steps to calculate the angle deviation (AngErr) and to call the two-dimensional surface lookup table module to obtain the dynamic gain parameters of the outer ring PID.

[0024] In this embodiment, angular velocity is the rate of change of the rotation angle with respect to time. The internal calculation is uniformly in meters per second and radians, while the external calculation still uses kilometers per hour and angle values. The two are converted into equivalent values. The outer ring sampling uses a uniform outer ring sampling period.

[0025] S20: Perform interval division processing on the vehicle speed signal based on the working condition information to obtain vehicle speed interval information; perform deviation construction processing on the turning angle target information based on the working condition information to obtain turning angle deviation information. The vehicle speed interval information and turning angle deviation information are used to describe the joint change characteristics of vehicle speed and turning angle target information under different assisted driving working conditions. To avoid instability caused by relying solely on instantaneous errors, the continuous sampling sequence is divided into several segments. The division can be done using a fixed-length time window or by restarting the segment when there is a significant change in the direction of the command. Within each segment, samples closer to the end of the segment are given slightly higher weights, while samples closer to the beginning of the segment are given slightly lower weights. This ensures that the most recent error has a greater impact on the statistical results of the segment. By default, the weight of the sample at the end of the segment is approximately twice that at the beginning of the segment. The entire sample is then normalized to ensure that the sum of all weights within a segment equals "a complete sample". Each segment is assigned a "representative deviation value". When there are multiple segments, the average value is taken according to the importance of each segment. Unless otherwise specified, the importance is divided equally among the number of segments.

[0026] Methods for obtaining corner deviation information by performing deviation construction processing on corner target information based on working condition information include: Based on the working condition information, the command angle information and actual angle information corresponding to the current working condition are selected from the angle target information. The command angle information and actual angle information are synchronously processed in chronological order to obtain the sampling sequence information. Extract the time position marker of each sampling moment from the sampling sequence information, and perform sequence alignment processing on the sampling sequence information according to the time position marker to obtain the aligned sequence information; Based on the alignment sequence information, the difference sample between the commanded angle information and the actual angle information is calculated at each sampling time. The difference sample is then aggregated and processed according to the time position mark to obtain the angle deviation information.

[0027] In this embodiment, the vehicle speed interval information is interval marking data formed by grouping continuously changing vehicle speed signals in the operating condition information according to a pre-set vehicle speed range. For example, 0-10 km / h can be defined as the first vehicle speed interval, and 10-30 km / h can be defined as the second vehicle speed interval. When the vehicle speed signals in the outer ring sampling period are 2, 3, 4, and 5 km / h respectively, the corresponding vehicle speed interval information can be uniformly marked as the first vehicle speed interval. The turning angle deviation information is the sequence or statistical result of the angle deviation value (AngErr) calculated based on the commanded turning angle signal and the actual turning angle signal in the outer ring sampling period. The vehicle speed interval information and the turning angle deviation information together describe the joint change characteristics between vehicle speed change and turning angle target information deviation under different assisted driving conditions, providing a structured input basis for the subsequent two-dimensional surface lookup table module based on vehicle speed signal and angle deviation value (AngErr).

[0028] It should be noted that when performing deviation construction processing on the cornering target information based on the operating condition information, the commanded cornering information and actual cornering information corresponding to the current operating condition are first selected from the cornering target information according to the operating condition information. Then, the commanded cornering information and actual cornering information are synchronously processed in chronological order to obtain the sampling sequence information. In the specific implementation, the commanded cornering information comes from the commanded cornering signal given by the intelligent driving control unit, and the actual cornering information comes from the cornering feedback signal output by the cornering sensor. For example, in the aforementioned example, four consecutive sampling cycles... During the period, the commanded rotation angle information is 5°, 10°, 15° and 20° respectively, and the actual rotation angle information is 0°, 4°, 11° and 18° respectively. Then the sampling sequence information can be written as 4 records arranged according to the time position mark: Record 1 is time 1, commanded rotation angle 5°, actual rotation angle 0°; Record 2 is time 2, commanded rotation angle 10°, actual rotation angle 4°; and so on. This sampling sequence information ensures that the commanded rotation angle information and the actual rotation angle information correspond one-to-one at each sampling time when constructing the angle deviation value (AngErr) in the future.

[0029] It is understandable that after obtaining the sampling sequence information, the time position marker of each sampling moment is extracted based on the sampling sequence information, and the sampling sequence information is aligned according to the time position marker to obtain the aligned sequence information. The aligned sequence information can be understood structurally as two sequences arranged according to the same time index: one is the command angle information sequence, and the other is the actual angle information sequence. In the aforementioned example, the command angle information sequence in the aligned sequence information is 5, 10, 15, 20, and the actual angle information sequence is 0, 4, 11, 18. Based on this, the difference sample between the command angle information and the actual angle information is calculated at each sampling moment according to the aligned sequence information. For example, the difference samples are 5, 6, 4, 2 in sequence. The difference samples are aggregated according to the time position marker, and the angle deviation values ​​(AngErr) at each sampling moment are combined into a time series or statistically summarized according to subsequent time segments to obtain the angle deviation information. This angle deviation information not only retains the error change trend at each sampling moment, but also provides a unified deviation dimension for subsequent optimization of gain adjustment by time segment or operating condition segment.

[0030] Methods for performing aggregation calculations on difference samples according to time position markers to obtain corner deviation information include: Segmentation marking is performed on each sampling time based on the difference samples and time position markers to obtain segmentation marking information. Segmentation aggregation is then performed on the time position markers based on the segmentation marking information to obtain time segmentation information. The difference samples are segmented and aggregated based on the segmentation labeling information to obtain segmented sample information. The deviation statistics are then performed on each time segment based on the segmented sample information to obtain segmented deviation information. The segmented deviation information is weighted and aggregated based on the time segmentation information to obtain aggregated deviation information. The sampled sequence information is then subjected to deviation characterization processing based on the aggregated deviation information to obtain corner deviation information.

[0031] Methods for performing weighted summarization processing on segmented deviation information based on time segmentation information to obtain summarized deviation information include: Based on the time segmentation information, the segmented deviation information is processed by sample expansion according to the time segmentation and sampling order to obtain a sample record set. Each record in the sample record set corresponds to a time index and the deviation amplitude corresponding to the time index. Based on the time segmentation information, the relative position calculation is performed on the time index in the sample record set to obtain a position sequence. Each position value in the position sequence is used to characterize the normalized time position of the corresponding sample record within its respective time segment. Perform weight generation processing on the position sequence to obtain a weight sequence. Each weight value in the weight sequence is a monotonic function output of the corresponding position value in the position sequence. The deviation amplitude in the sample record set is processed by performing sample product processing according to the weight value in the weight sequence to obtain a weighted record set. The weighted record set is then processed by performing segmented accumulation processing according to time segment information to obtain the summary deviation information.

[0032] In this embodiment, the aggregation calculation processing of the difference samples according to the time position mark is a process of introducing time segmentation and statistical operation on the basis of the obtained difference samples. The difference samples are essentially derived from the angle deviation value (AngErr) between the command angle information and the actual angle information at each sampling time. The segmentation mark information is the mark result of assigning the time segment to each sampling time according to the time position mark. The time segmentation information gives the start and end time index and the number of sampling points corresponding to each time segment based on the segmentation mark information. By performing segmentation mark processing on each sampling time according to the difference samples and the time position mark to obtain the segmentation mark information, and then performing segmentation aggregation processing on the time position mark according to the segmentation mark information to obtain the time segmentation information, the angle deviation value (AngErr) arranged in time order can be organized into time segments with physical meaning without changing the difference sample value, so that the subsequent deviation statistics can distinguish the error performance of different turning stages.

[0033] Furthermore, after obtaining the segmentation label information and time segmentation information, the difference samples are processed by segmentation aggregation based on the segmentation label information to obtain segmented sample information. The segmented sample information can be understood as a set of difference samples divided by time segments. Each time segment corresponds to a set of continuous angle deviation values ​​(AngErr) samples. Based on this, deviation statistical processing is performed on each time segment according to the segmented sample information to obtain segmented deviation information. The deviation statistical processing can adopt conventional methods in this field, such as absolute value averaging or variance, to characterize the overall deviation degree of angle deviation values ​​(AngErr) within the time segment. For example, in the aforementioned Example 1, if the difference samples at the four sampling times are 5, 6, 4, and 2, and are divided into two time segments, then the segmented deviation information can represent the average deviation level of the first two sampling times and the last two sampling times, respectively, thereby obtaining a more smooth and representative deviation dimension at the time segmentation level than the point-by-point difference samples.

[0034] The process of weighted aggregation of segmented deviation information based on time segmentation information is a secondary aggregation of segmented deviation information at the time segmentation level to obtain aggregated deviation information. The sample expansion process involves expanding the segmented deviation information into a sample record set according to time segmentation and sampling order. Each record in the sample record set corresponds to a time index and the deviation magnitude of the time segment to which that index belongs. Then, relative position calculation is performed on the time indices in the sample record set based on the time segmentation information to obtain a position sequence. Each position value in the position sequence represents the normalized time position of the corresponding sample record within its respective time segment. After obtaining the position sequence, weights are applied to the position sequence. The generation process generates a weight sequence, where each weight value is a monotonic function output of the corresponding position value in the position sequence. For example, the weight of sample records near the end of the time segment can be increased. Then, the deviation amplitude in the sample record set is processed by performing sample product processing according to the weight values ​​in the weight sequence to obtain a weighted record set. Based on the time segment information, the weighted record set is processed by performing segmented accumulation processing according to the time segment to obtain the summary deviation information. Then, based on the summary deviation information, the sampling sequence information is processed by deviation characterization to obtain the corner deviation information. This corner deviation information reflects both the comprehensive level of the angle deviation value (AngErr) over the entire sampling time and retains the relative importance of different time segments in the overall error.

[0035] Furthermore, compared to the existing technology that adjusts the outer loop gain of the steering angle closed-loop and speed closed-loop PID based on the instantaneous angle deviation value (AngErr) only at a single moment or under a few operating conditions, this embodiment introduces segmented marking information, time segmented information, segmented sample information, and summarized deviation information. It performs segmented statistical and weighted summarization processing on the time axis for the difference samples, so that the final steering angle deviation information can comprehensively reflect the deviation characteristics between the commanded steering angle signal and the actual steering angle signal under different time periods and different assisted driving conditions in the same data structure. This provides a more stable and representative input for the subsequent two-dimensional surface lookup table module based on vehicle speed signal and angle deviation value (AngErr). Without changing the EPS hardware and inner loop current control structure, the dynamic gain combination selected by the two-level outer loop PID control under conditions such as automatic parking and lane keeping is more in line with the real deviation distribution, which is beneficial to improving the smoothness of steering torque output and suppressing response overshoot.

[0036] S30: Based on the vehicle speed range information, perform outer loop gain mapping processing to obtain gain space information. Based on the steering angle deviation information, perform two-dimensional lookup table processing on the gain space information to obtain gain set information. The gain space information and gain set information are used to provide adjustable gain parameters for the two-level closed-loop outer loop without changing the hardware and inner loop structure of the electric power steering system. Methods for obtaining gain set information by performing two-dimensional lookup table processing on gain spatial information based on rotation deviation information include: Based on the rotation deviation information, perform mesh selection processing on the two-dimensional nodes in the gain space information to obtain lookup table mesh information. Based on the lookup table mesh information, perform coordinate mapping processing on the rotation deviation information to obtain mesh index information. Based on the grid index information, interpolation calculation is performed on the node gain values ​​in the lookup table grid information to obtain the initial gain information. Based on the corner deviation information, the initial gain information is then processed by dimension adjustment to obtain the gain set information.

[0037] The gain table is built with two dimensions: "vehicle speed" and "angle deviation". It is recommended to set the vehicle speed to cover several typical points from low speed to high speed (e.g., 5-6 gears from standstill to high speed). The angle deviation is set to 4-5 gears from zero to the common maximum deviation. During operation, the current condition is first located within the small square enclosed by two adjacent vehicle speed points and two adjacent deviation points. A smooth transition is only performed between these four nodes: the closer to the node, the closer to the gain value of that node, and the farther away, the smaller the impact. If the current value falls outside the grid boundary, it is first clamped to the nearest boundary and then processed in the above way. This smooth transition is applied to three types of gain parameters at the same time, namely the angle loop proportional gain, the speed loop proportional gain, and the speed loop integral gain.

[0038] To ensure reproducibility, the calibration examples should cover at least the vehicle speed gears of 0, 5, 10, 20, 40, and 80 km / h, and the angle deviation gears of 0, 1, 2, 3, and 5°. If the deviation falls outside the boundaries during operation, it should be treated as the nearest boundary.

[0039] In this embodiment, it can be understood that the gain space information refers to a two-dimensional surface lookup table structure established during the calibration phase based on vehicle speed range information and angle deviation value (AngErr). This structure is discretized according to vehicle speed range information in the vehicle speed dimension and according to steering angle deviation information in the deviation dimension. At each two-dimensional node, the gain parameter values ​​required for the corresponding outer loop control are stored, including the proportional gain Kp1' for the steering angle closed-loop PID, the proportional gain Kp2' for the speed closed-loop PID, and the integral gain Ki2'. The gain set information is a set of dynamic gain parameters obtained by performing a two-dimensional lookup table in the above-mentioned gain space information based on the current vehicle speed range information and the current steering angle deviation information during operation. This gain set information provides adjustable gain parameters that match the current operating conditions for the two-level PID control of steering angle closed loop and speed closed loop in the outer loop of EPS without changing the hardware of the electric power steering system and the inner loop current control structure. This allows the same vehicle to automatically switch to different outer loop gain combinations under different vehicle speeds and different (AngErr) combinations.

[0040] The process of performing outer-loop gain mapping based on vehicle speed interval information to obtain gain space information can be understood as follows: Two-dimensional surface lookup table modules are pre-arranged on several vehicle speed intervals given by the vehicle speed interval information. Two-dimensional grids of vehicle speed and deviation dimensions are constructed using the interval markers and the value range of the corner deviation information from the vehicle speed interval information. The corresponding Kp1', Kp2', and Ki2' node gain values ​​are stored at each two-dimensional node, thus forming gain space information covering all vehicle speed intervals and typical corner deviation values ​​(AngErr). During runtime, grid selection processing is performed on the two-dimensional nodes in the gain space information based on the corner deviation information to obtain lookup table grid information. Then, coordinate mapping processing is performed on the corner deviation information based on the lookup table grid information to obtain grid index information. The grid index information clearly defines the cell in the two-dimensional grid where the current vehicle speed interval and the current corner deviation information are located, enabling subsequent interpolation calculations to be performed within the local grid, thereby avoiding a global search across the entire gain space information.

[0041] In this embodiment, it can be understood that performing interpolation calculation on the node gain values ​​in the lookup table grid information based on the grid index information to obtain the initial gain information means that after determining the two-dimensional grid cell corresponding to the current operating condition, extracting the node gain values ​​of Kp1', Kp2', and Ki2' of several nodes around the grid cell from the lookup table grid information, calculating the corresponding continuous gain value in a linear or other monotonic interpolation method based on the relative position of the current vehicle speed range and the current steering angle deviation information within the grid cell in the grid index information, and summarizing the interpolation results into the initial gain information. The initial gain information simultaneously includes the proportional gain Kp1' of the steering angle closed-loop PID and the proportional gain Kp2' and integral gain Ki2' of the speed closed-loop PID. Subsequently, performing dimensionality sorting processing on the initial gain information based on the steering angle deviation information, splitting or rearranging the initial gain information corresponding to different vehicle speed ranges and different deviation ranges into gain set information organized by the sampling time index according to the needs of the outer loop control, so that the gain set information can be directly read by the outer loop of the two-level PID of steering angle closed loop and speed closed loop and participate in the control calculation.

[0042] Methods for obtaining initial gain information by performing interpolation calculations on the node gain values ​​in the lookup table grid information based on the grid index information include: Based on the grid index information and the lookup table grid information, a neighbor filtering process is performed on the position of the two-dimensional grid node corresponding to each grid index to obtain the neighbor index information. Each record in the neighbor index information associates a grid index with the position indexes of several adjacent nodes. Based on the lookup grid information, the node gain values ​​corresponding to the neighbor index information are subjected to gain aggregation processing to obtain neighbor gain information. The neighbor gain information is aggregated according to each grid index to form an ordered gain set information by aggregating the node gain values ​​of adjacent nodes. Based on the neighbor gain information and grid index information, interpolation weight calculation is performed on each grid index to obtain weight set information. The relative position of each set of weights in the weight set information and the corresponding grid index in the two-dimensional grid maintains a monotonic relationship. The node gain values ​​in the neighbor gain information are weighted and combined according to the weights in the weight set information to obtain the initial gain information. The initial gain information provides the interpolated gain value after a two-dimensional lookup table at each grid index.

[0043] Methods for obtaining initial gain information by performing weighted combination processing on the node gain values ​​in the neighbor gain information according to the weights in the weight set information include: For each node gain value in the neighbor gain information, perform sample expansion processing according to the record order to obtain the gain sample set; for each weight value in the weight set information, perform sample expansion processing according to the record order to obtain the weight sample set. The gain value of each node in the gain sample set is multiplied with the corresponding weight value in the weight sample set to obtain the product sample set. The product sample set represents the product result of the gain value of each node and the corresponding weight value at the sample level. The initial gain information is obtained by performing a summation calculation on all product results in the product sample set.

[0044] Understandably, the grid index information refers to the index data obtained after determining the position of the current operating condition in the two-dimensional surface lookup table module based on the vehicle speed range information and the turning angle deviation information. Each grid index information identifies a two-dimensional grid cell in the gain space information for the combination of a vehicle speed signal and an angle deviation value (AngErr). The lookup table grid information refers to the set of two-dimensional grid nodes pre-stored for the entire gain space information. Each two-dimensional grid node position corresponds to a set of outer loop gain node gain values, including the proportional gain Kp1' of the turning angle closed-loop PID and the proportional gain Kp2' and integral gain Ki2' of the speed closed-loop PID. Based on the grid index information and the lookup table grid information, the neighbor point filtering process is performed on the two-dimensional grid node position corresponding to each grid index to obtain the neighbor point index information. Each record in the neighbor point index information associates a grid index with the position indexes of several adjacent nodes. Through this process, the continuous vehicle speed signal and angle deviation value (AngErr) are mapped to the discrete two-dimensional node neighborhood, providing a local search range for subsequent interpolation calculations based on the neighborhood node gain.

[0045] It should be noted that after obtaining the neighbor index information, gain aggregation processing is performed on the node gain values ​​corresponding to the neighbor index information according to the lookup table grid information to obtain neighbor gain information. The neighbor gain information aggregates the node gain values ​​of adjacent nodes according to each grid index to form ordered gain set information, so that all Kp1', Kp2', Ki2' node gain values ​​corresponding to a certain working condition are arranged in a fixed order in structure, which facilitates the item-by-item correspondence with the subsequent weight set information. At the same time, according to the neighbor gain information and grid index information, interpolation weight calculation processing is performed on each grid index to obtain weight set information. The relative position of each set of weights in the weight set information and the corresponding grid index in the two-dimensional grid maintains a monotonic relationship. In other words, when the current vehicle speed signal and angle deviation value (AngErr) are closer to a certain node, the weight value of the corresponding node is larger. When the current working condition is located in the middle of multiple nodes, the weights of each node are smoothly distributed according to the grid geometric position, so that the initial gain information keeps changing continuously in the two-dimensional vehicle speed and (AngErr) plane during interpolation.

[0046] In this embodiment, the weighted combination processing of node gain values ​​in the neighbor gain information according to the weights in the weight set information is a process of refining the neighbor gain information and weight set information into sample-level operations. For each node gain value in the neighbor gain information, sample expansion processing is performed according to the record order to obtain a gain sample set. For each weight value in the weight set information, sample expansion processing is performed according to the record order to obtain a weight sample set. This ensures that each node gain value in the gain sample set corresponds one-to-one with the corresponding weight value in the weight sample set in terms of record number. Based on this, a product calculation processing is performed on each node gain value in the gain sample set and the corresponding weight value in the weight sample set to obtain a product sample set. The product sample set represents the gain of each node at the sample level. The product of the value and the corresponding weight value is then summed across all product results in the product sample set to obtain initial gain information. This initial gain information provides interpolated gain values ​​after a two-dimensional lookup table at each grid index. These values ​​can be directly used as dynamic gain parameters for the corner closed-loop PID and speed closed-loop PID in the outer loop under the corresponding operating conditions to participate in the EPS outer loop control calculation. Compared to existing technologies that only store fixed Kp1, Kp2, and Ki2 at discrete nodes, this embodiment uses interpolation calculation to make Kp1', Kp2', and Ki2' continuously adjustable in the joint space of vehicle speed signal and angle deviation value (AngErr). This is beneficial for obtaining a more balanced control effect between torque smoothness and response speed under different operating conditions.

[0047] S40: Perform closed-loop regulation processing on the operating condition information based on the gain set information to obtain the angle control information.

[0048] Methods for obtaining steering angle control information by performing closed-loop regulation processing on operating condition information based on gain set information include: The rotation angle target information and angular velocity target information in the working condition information are processed synchronously to obtain the loop target information. The loop target information provides the corresponding rotation angle target quantity and angular velocity target quantity at each sampling position. The angle gain sample and speed gain sample in the gain set information are processed by loop allocation according to the sampling position to obtain loop gain information. The loop gain information associates the angle gain sample and speed gain sample with the rotation angle target quantity and angular velocity target quantity in the loop target information at each sampling position. For the target values ​​of rotation angle and angular velocity in the loop target information, closed-loop calculation processing is performed according to the angle gain sample and speed gain sample in the loop gain information to obtain the rotation angle control information.

[0049] In this embodiment, the gain set information refers to the outer loop dynamic gain parameter set obtained by looking up a table on a two-dimensional surface based on the vehicle speed signal and the angle deviation value (AngErr). It includes at least the angle gain sample corresponding to the cornering closed-loop PID and the speed gain sample corresponding to the speed closed-loop PID. The operating condition information provides the cornering target information and the angular velocity target information at each sampling time in the outer loop sampling period. By performing synchronous processing on the cornering target information and the angular velocity target information in the operating condition information, the two are arranged according to a unified time index at each sampling position to obtain the loop target information. The loop target information simultaneously provides the corresponding cornering target quantity and angular velocity target quantity at each sampling position, providing a consistent target input basis for the subsequent simultaneous calculation of cornering closed-loop and speed closed-loop at the same sampling position.

[0050] After obtaining the loop target information, the angle gain samples and speed gain samples in the gain set information are processed according to the sampling position to obtain the loop gain information. At each sampling position, the loop gain information associates the angle gain samples and speed gain samples with the target values ​​of rotation angle and angular velocity in the loop target information. From the perspective of those skilled in the art, this can be understood as providing a set of angle gain parameters Kp1' that match the current vehicle speed signal and angle deviation value (AngErr) for the rotation angle closed-loop PID in each sampling cycle of the outer loop, and providing a set of speed gain parameters Kp2' and Ki2' that match the same operating condition for the speed closed-loop PID. It is ensured that these gain parameters correspond one-to-one with the corresponding target values ​​of rotation angle and angular velocity on the time axis. In this way, it is not necessary to search for or reorganize the gain parameters during operation in subsequent closed-loop calculations. Instead, the outer loop control calculation can be performed directly according to the sampling position of the loop target information and the loop gain information.

[0051] The methods for obtaining steering control information by performing closed-loop calculations on the steering angle and angular velocity target values ​​in the loop target information, based on the angle gain samples and speed gain samples in the loop gain information, include: Based on the loop gain information, the angle target quantity and the angle gain sample are processed by angle loop calculation to obtain the angle loop information. The angle loop information at each sampling position represents the angle loop output quantity after proportional adjustment of the angle target quantity based on the angle gain sample. Based on the loop target information, the angular velocity target quantity and the speed gain sample are processed by speed loop calculation to obtain speed loop information. The speed loop information at each sampling position represents the speed loop output quantity after proportional-integral adjustment of the angular velocity target quantity based on the speed gain sample. The angle loop output in the angle loop information and the speed loop output in the speed loop information are combined in a loop to obtain the angle control information. The angle control information is used as the outer loop target input of the inner loop current control structure at each sampling position.

[0052] In this embodiment, the angle loop calculation processing is performed on the target angle quantity and the angle gain sample based on the loop gain information. This is an outer loop angle loop operation completed under the established EPS angle closed loop structure. The target angle quantity given in the loop target information is paired with the aforementioned actual angle signal to form the angle deviation value (AngErr). The angle gain sample given in the loop gain information corresponds to the angle closed loop PID dynamic proportional parameter Kp1' obtained through the two-dimensional surface lookup table module. In specific implementation, those skilled in the art can perform proportional or proportional-integral adjustment in the angle loop based on the angle deviation value (AngErr) and the angle gain sample Kp1' to obtain the angle loop information. The angle loop information at each sampling position represents the angle loop output quantity calculated based on the angle deviation value (AngErr) and the dynamic gain Kp1' under the current operating conditions.

[0053] It should be noted that the speed loop calculation processing based on the target angular velocity and speed gain sample according to the loop target information is an outer loop speed loop operation completed under the existing speed closed-loop structure of EPS. The target angular velocity in the loop target information and the actual angular velocity constitute the speed error. The speed gain sample in the loop gain information corresponds to the speed closed-loop PID dynamic proportional gain Kp2' and integral gain Ki2' obtained by the two-dimensional surface lookup table module. In this embodiment, in the speed loop calculation processing, proportional-integral adjustment can be performed according to the speed error and the speed gain samples Kp2' and Ki2' to obtain the speed loop information. The speed loop information at each sampling position represents the speed loop output calculated based on the speed error and dynamic gain parameters under the current operating conditions. In this way, under different assisted driving conditions such as automatic parking and lane keeping, the response strength and integral compensation capability of the speed closed-loop PID can be automatically adjusted with the changes in vehicle speed signal and angle deviation value (AngErr).

[0054] Performing loop synthesis processing on the angle loop output in the angle loop information and the speed loop output in the speed loop information refers to the process of combining the aforementioned angle loop output and speed loop output into unified steering control information under a two-stage series outer loop structure. Those skilled in the art can combine the angle loop output and speed loop output in a predetermined superposition or weighting manner based on the existing EPS outer loop to inner loop interface design to form the outer loop target input quantity used to drive the inner loop current control structure at each sampling position, i.e., the steering control information. This is in contrast to the prior art which samples across the entire vehicle speed range and the entire angle deviation value (AngErr). This embodiment uses a two-stage PID outer loop operation with fixed Kp1, Kp2, and Ki2. By introducing dynamic angle gain samples and speed gain samples from the gain set information during the generation of angle loop information and speed loop information, and outputting steering angle control information after loop synthesis processing, the two-stage PID outer loops of steering angle closed loop and speed closed loop can adaptively select a more suitable gain combination for different vehicle speed signals and (AngErr) combinations without changing the EPS hardware structure and inner loop current closed loop control method. This helps to simultaneously take into account steering response speed and torque output smoothness in intelligent driving human-machine co-driving state.

[0055] To avoid "integral overload" caused by sudden errors or actuator limitations, the accumulation of integral terms is paused when the outer loop output approaches or reaches the preset upper limit, or when the error has entered a very small dead zone. Upper and lower limits are set for the integral accumulation value to avoid long-term accumulation. The outputs of the angle loop and speed loop are both subject to amplitude limits and appropriate rate of change limits before being fed into the inner loop to reduce jitter.

[0056] Example 2 Please see Figure 2 As shown, based on the same inventive concept, this embodiment discloses a system for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions. For details not covered in this embodiment, please refer to the relevant sections of Embodiment 1. The system includes: Operating condition analysis module: used to acquire vehicle sensor data, perform operating condition analysis processing on the vehicle sensor data to obtain operating condition information. Among them, the vehicle sensor data includes vehicle speed signal and steering angle target information. The operating condition information is used to characterize the sampling state of the outer ring of the electric power steering system. The interval deviation module is used to perform interval division processing on the vehicle speed signal based on the operating condition information to obtain vehicle speed interval information. Based on the operating condition information, it performs deviation construction processing on the turning angle target information to obtain turning angle deviation information. The vehicle speed interval information and turning angle deviation information are used to describe the joint change characteristics of vehicle speed and turning angle target information under different assisted driving conditions. Gain lookup table module: It is used to perform outer loop gain mapping processing based on vehicle speed range information to obtain gain space information. It performs two-dimensional lookup table processing on gain space information based on steering angle deviation information to obtain gain set information. Among them, gain space information and gain set information are used to provide adjustable gain parameters for the two-level closed-loop outer loop without changing the hardware and inner loop structure of the electric power steering system. Closed-loop control module: Used to perform closed-loop control processing on the operating condition information based on the gain set information to obtain the angle control information.

[0057] The flowcharts in the accompanying drawings illustrate the functions that may be implemented by the methods according to various embodiments of the present invention. In this regard, each block in the flowchart may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the flowchart, and combinations of blocks in the flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0058] Those skilled in the art will understand that the features described in the various embodiments of the present invention can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention can be combined and / or combined in various ways, and all such combinations and / or combinations fall within the scope of the present invention.

[0059] The embodiments of the present invention have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of the invention is defined by the appended embodiments and their equivalents. Without departing from the scope of the invention, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of the invention.

Claims

1. A method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions, characterized in that, include: Acquire vehicle sensor data, perform operating condition analysis processing on the vehicle sensor data to obtain operating condition information. The vehicle sensor data includes vehicle speed signal and steering angle target information. The operating condition information is used to characterize the sampling state of the outer ring of the electric power steering system. Based on the operating condition information, the vehicle speed signal is divided into intervals to obtain vehicle speed interval information. Based on the operating condition information, the turning angle target information is deviated to obtain turning angle deviation information. The vehicle speed interval information and turning angle deviation information are used to describe the joint change characteristics of vehicle speed and turning angle target information under different assisted driving conditions. Based on the vehicle speed range information, outer loop gain mapping is performed to obtain gain space information. Based on the steering angle deviation information, two-dimensional lookup table processing is performed on the gain space information to obtain gain set information. The gain space information and gain set information are used to provide adjustable gain parameters for the two-level closed-loop outer loop without changing the hardware and inner loop structure of the electric power steering system. Based on the gain set information, closed-loop regulation is performed on the operating condition information to obtain the angle control information.

2. The method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions according to claim 1, characterized in that, Methods for obtaining corner deviation information by performing deviation construction processing on corner target information based on working condition information include: Based on the working condition information, the command angle information and actual angle information corresponding to the current working condition are selected from the angle target information. The command angle information and actual angle information are synchronously processed in chronological order to obtain the sampling sequence information. Extract the time position marker of each sampling moment from the sampling sequence information, and perform sequence alignment processing on the sampling sequence information according to the time position marker to obtain the aligned sequence information; Based on the alignment sequence information, the difference sample between the commanded angle information and the actual angle information is calculated at each sampling time. The difference sample is then aggregated and processed according to the time position mark to obtain the angle deviation information.

3. The method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions according to claim 2, characterized in that, Methods for performing aggregation calculations on difference samples according to time position markers to obtain corner deviation information include: Segmentation marking is performed on each sampling time based on the difference samples and time position markers to obtain segmentation marking information. Segmentation aggregation is then performed on the time position markers based on the segmentation marking information to obtain time segmentation information. The difference samples are segmented and aggregated based on the segmentation labeling information to obtain segmented sample information. The deviation statistics are then performed on each time segment based on the segmented sample information to obtain segmented deviation information. The segmented deviation information is weighted and aggregated based on the time segmentation information to obtain aggregated deviation information. The sampled sequence information is then subjected to deviation characterization processing based on the aggregated deviation information to obtain corner deviation information.

4. The method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions according to claim 3, characterized in that, Methods for performing weighted summarization processing on segmented deviation information based on time segmentation information to obtain summarized deviation information include: Based on the time segmentation information, the segmented deviation information is processed by sample expansion according to the time segmentation and sampling order to obtain a sample record set. Each record in the sample record set corresponds to a time index and the deviation amplitude corresponding to the time index. Based on the time segmentation information, the relative position calculation is performed on the time index in the sample record set to obtain a position sequence. Each position value in the position sequence is used to characterize the normalized time position of the corresponding sample record within its respective time segment. Perform weight generation processing on the position sequence to obtain a weight sequence. Each weight value in the weight sequence is a monotonic function output of the corresponding position value in the position sequence. The deviation amplitude in the sample record set is processed by performing sample product processing according to the weight value in the weight sequence to obtain a weighted record set. The weighted record set is then processed by performing segmented accumulation processing according to time segment information to obtain the summary deviation information.

5. The method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions according to claim 2, characterized in that, Methods for obtaining gain set information by performing two-dimensional lookup table processing on gain spatial information based on rotation deviation information include: Based on the rotation deviation information, perform mesh selection processing on the two-dimensional nodes in the gain space information to obtain lookup table mesh information. Based on the lookup table mesh information, perform coordinate mapping processing on the rotation deviation information to obtain mesh index information. Based on the grid index information, interpolation calculation is performed on the node gain values ​​in the lookup table grid information to obtain the initial gain information. Based on the corner deviation information, the initial gain information is then processed by dimension adjustment to obtain the gain set information.

6. The method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions according to claim 5, characterized in that, Methods for obtaining initial gain information by performing interpolation calculations on the node gain values ​​in the lookup table grid information based on the grid index information include: Based on the grid index information and the lookup table grid information, a neighbor filtering process is performed on the position of the two-dimensional grid node corresponding to each grid index to obtain the neighbor index information. Each record in the neighbor index information associates a grid index with the position indexes of several adjacent nodes. Based on the lookup grid information, the node gain values ​​corresponding to the neighbor index information are subjected to gain aggregation processing to obtain neighbor gain information. The neighbor gain information is aggregated according to each grid index to form an ordered gain set information by aggregating the node gain values ​​of adjacent nodes. Based on the neighbor gain information and grid index information, interpolation weight calculation is performed on each grid index to obtain weight set information. The relative position of each set of weights in the weight set information and the corresponding grid index in the two-dimensional grid maintains a monotonic relationship. The node gain values ​​in the neighbor gain information are weighted and combined according to the weights in the weight set information to obtain the initial gain information. The initial gain information provides the interpolated gain value after a two-dimensional lookup table at each grid index.

7. A method for improving torque smoothness in human-machine co-driving state for EPS-responsive intelligent driving according to claim 6, characterized in that, Methods for obtaining initial gain information by performing weighted combination processing on the node gain values ​​in the neighbor gain information according to the weights in the weight set information include: For each node gain value in the neighbor gain information, perform sample expansion processing according to the record order to obtain the gain sample set; for each weight value in the weight set information, perform sample expansion processing according to the record order to obtain the weight sample set. The gain value of each node in the gain sample set is multiplied with the corresponding weight value in the weight sample set to obtain the product sample set. The product sample set represents the product result of the gain value of each node and the corresponding weight value at the sample level. The initial gain information is obtained by performing a summation calculation on all product results in the product sample set.

8. A method for improving torque smoothness in EPS-responsive intelligent driving under human-machine co-driving conditions, as described in claim 7, characterized in that... Methods for obtaining steering angle control information by performing closed-loop regulation processing on operating condition information based on gain set information include: The rotation angle target information and angular velocity target information in the working condition information are processed synchronously to obtain the loop target information. The loop target information provides the corresponding rotation angle target quantity and angular velocity target quantity at each sampling position. The angle gain sample and speed gain sample in the gain set information are processed by loop allocation according to the sampling position to obtain loop gain information. The loop gain information associates the angle gain sample and speed gain sample with the rotation angle target quantity and angular velocity target quantity in the loop target information at each sampling position. For the target values ​​of rotation angle and angular velocity in the loop target information, closed-loop calculation processing is performed according to the angle gain sample and speed gain sample in the loop gain information to obtain the rotation angle control information.

9. A method for improving torque smoothness in human-machine co-driving state for EPS-responsive intelligent driving according to claim 8, characterized in that, The methods for obtaining steering control information by performing closed-loop calculations on the steering angle and angular velocity target values ​​in the loop target information, based on the angle gain samples and speed gain samples in the loop gain information, include: Based on the loop gain information, the angle target quantity and the angle gain sample are processed by angle loop calculation to obtain the angle loop information. The angle loop information at each sampling position represents the angle loop output quantity after proportional adjustment of the angle target quantity based on the angle gain sample. Based on the loop target information, the angular velocity target quantity and the speed gain sample are processed by speed loop calculation to obtain speed loop information. The speed loop information at each sampling position represents the speed loop output quantity after proportional-integral adjustment of the angular velocity target quantity based on the speed gain sample. The angle loop output in the angle loop information and the speed loop output in the speed loop information are combined in a loop to obtain the angle control information. The angle control information is used as the outer loop target input of the inner loop current control structure at each sampling position.

Images