Cross-camber road spectrum driven lane keeping feedforward correction method, system and vehicle

Abstract

The application provides a kind of transverse slope road spectrum driven lane keeping feedforward correction method, system and vehicle, belongs to intelligent driving technical field.Vehicle is based on GNSS / odometer / two-dimensional IMU fusion positioning, forms road spectrum index key on road position.When steady-state gate is satisfied, the steady-state value of the sampled lateral deviation integral component is used as an observation, and is written into the current position entry.The mean and variance of the feedforward correction angle of the entry are recursively updated to achieve self-learning.During driving, according to the current position and the learned statistics of the next several along the program sequence, the low-frequency and limited feedforward correction angle is obtained by smoothing, low-pass filtering and amplitude / rate limiting.The correction is superimposed with the physical feedforward when the confidence meets, and the steering command is synthesized with the feedback control.The method does not need to add new sensors, and can make small-angle correction related to position before the vehicle enters the known road section, thereby reducing the burden of integral loop and improving the accuracy and stability of lane centering control.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and more specifically, to a lane keeping feedforward correction method, system, and vehicle driven by cross slope road spectrum. Background Technology

[0002] Lane Keeping Assist (LKA) and Lane Centering Control (LCC) are common features in intelligent driving systems. They are designed to keep the vehicle stable in the center of the lane by controlling the vehicle's steering, thereby improving driving safety and comfort.

[0003] Currently, the lateral control architecture based on road geometry and vehicle model, which is used to implement LKA and LCC, is commonly used. The feedforward part performs active steering compensation based on prior information such as road curvature to reduce tracking error, while the feedback part uses real-time measurement signals such as lateral distance deviation and heading deviation to perform closed-loop correction of deviation.

[0004] However, when a vehicle with this lateral control architecture enters a cross-slope section and the cross-slope information is unavailable or delayed, the physical feedforward will produce systematic errors. The actual centering mainly relies on the integral loop to slowly offset these errors. This can easily lead to problems such as the vehicle driving close to the edge, delayed centering, overshoot, and reduced phase margin, thus significantly affecting the stability, responsiveness, and ride comfort of the lane centering driving assistance function. Summary of the Invention

[0005] The purpose of this application is to provide a lane keeping feedforward correction method, system and vehicle driven by cross slope road spectrum, which can achieve the effect of performing small position-related angle corrections before the vehicle enters a known road segment without the need for additional sensors, thereby reducing the burden on the integral loop and improving the accuracy and stability of lane centering control.

[0006] The embodiments of this application are implemented as follows:

[0007] A first aspect of this application provides a lane keeping feedforward correction method driven by cross slope spectrum, the method comprising:

[0008] When a vehicle is traveling on a road, determine the road segment markers to which the vehicle belongs at its current location;

[0009] Based on the location key, the current location entry under the same road segment identifier and direction identifier, as well as the N preceding location entries along the route, are obtained; each location entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level; the location key consists of the road segment identifier, direction identifier, lane identifier, and the route index.

[0010] The feedforward correction angle is determined based on the average feedforward correction angle in the current position entry and the average feedforward correction angle in the N preceding position entries along the path.

[0011] Based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation, a steering angle control command is generated to control the vehicle's movement on the road.

[0012] Obtain the observation parameters of the vehicle at its current location, and update the current location entry based on the observation parameters.

[0013] As one possible implementation, determining the road segment identifier to which the vehicle belongs at its current location includes:

[0014] The integral component of the lateral deviation controller of the sampling vehicle is used as the observation sample;

[0015] The position key is determined by collecting parameters from at least one sensor on the vehicle.

[0016] Extract the road segment identifier from the location key to determine the road segment identifier to which the vehicle belongs at its current location.

[0017] As one possible implementation, the feedforward correction angle is determined based on the average feedforward correction angle in the current position entry and the average feedforward correction angle in the preceding N along-the-path position entries, including:

[0018] The average feedforward correction angle in the current position entry and the average feedforward correction angle in the preceding N position entries are aggregated and smoothed to obtain the initial feedforward correction angle.

[0019] The initial feedforward correction angle is corrected to obtain the feedforward correction angle.

[0020] As one possible implementation, the initial feedforward correction angle is corrected to obtain the feedforward correction angle, including:

[0021] Based on formula and formula The initial feedforward correction angle is corrected to obtain the feedforward correction angle.

[0022] in, Indicates the feedforward correction angle. This represents the initial feedforward correction angle. This represents the physical feedforward angle of the monorail vehicle model. This represents the absolute value of the feedforward correction angle. This represents the maximum allowable feedforward correction angle threshold for the vehicle. This represents the rate of change of the vehicle's feedforward correction angle. This represents the threshold for the maximum allowable rate of change of the vehicle's feedforward correction angle. and Indicates the calibration parameters.

[0023] As one possible implementation, vehicle centering deviation includes: lateral distance deviation feedback and azimuth deviation feedback. Based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the preceding N path entries, and the real-time collected vehicle centering deviation, the vehicle's movement on the road is controlled, including:

[0024] The path spectrum gating coefficients are determined based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, and the variance and confidence of the feedforward correction angle in the N preceding path position entries.

[0025] Based on the feedforward correction angle, road spectrum gating coefficient, lateral distance deviation feedback, and azimuth deviation feedback, the steering control command is generated.

[0026] The steering control command is sent to the vehicle's steering mechanism to control the vehicle's movement on the road.

[0027] As one possible implementation, the steering control command is generated based on the feedforward correction angle, road spectrum gating coefficient, lateral distance deviation feedback, and azimuth deviation feedback, including:

[0028] Based on formula , Generate corner control commands;

[0029] in, This represents the physical feedforward angle of the monorail vehicle model. Represents the spectrum gating coefficient. This indicates the feedforward correction angle of the vehicle at its current position. This indicates the amount of lateral distance deviation feedback for the vehicle. This indicates the vehicle's azimuth deviation feedback amount.

[0030] As one possible implementation, updating the current location entry based on observation parameters includes:

[0031] Based on the observed parameters, determine whether the vehicle meets the steady-state gating conditions;

[0032] If so, then determine the current quality weight;

[0033] Based on the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, update the mean of the feedforward correction angle in the current position entry to obtain the new mean of the feedforward correction angle.

[0034] Based on the mean of the new feedforward correction angle, the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, the variance of the feedforward correction angle is updated to obtain the new variance of the feedforward correction angle.

[0035] The confidence level is updated by adjusting the step size based on the location confidence level, the variance of the new feedforward correction angle, and the preset confidence level, thus obtaining a new confidence level.

[0036] As one possible implementation, determining whether the vehicle meets the steady-state gating condition based on observed parameters includes:

[0037] Obtain the observation parameters of the current location, including: road curvature, vehicle lateral distance deviation, and vehicle azimuth angle deviation;

[0038] The rate of change of the integral of the lateral distance deviation is determined based on the integral of the vehicle's lateral distance deviation.

[0039] The location reliability is determined based on the parameters collected by at least one sensor on the vehicle.

[0040] Based on the observed parameters, the rate of change of the integral of the lateral distance deviation, the positional confidence, the saturation state of the vehicle actuators, and the intervention state of the stability controller, it is determined whether the vehicle meets the steady-state gating conditions.

[0041] A second aspect of this application provides a lane keeping feedforward correction system driven by cross slope spectrum, the lane keeping feedforward correction system comprising:

[0042] When a vehicle is traveling on a road, the determination module is used to determine the road segment identifier to which the vehicle belongs at its current location;

[0043] The acquisition module is used to acquire the current location entry under the same road segment identifier and direction identifier, as well as the N preceding location entries based on the location key. Each location entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level. The location key consists of the road segment identifier, direction identifier, lane identifier, and along-the-road index.

[0044] The feedforward correction module is used to determine the feedforward correction angle based on the average feedforward correction angle in the current position entry and the average feedforward correction angle in the N preceding position entries along the path.

[0045] The steering control module is used to control the vehicle's movement on the road based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation.

[0046] The write module is used to obtain the observation parameters of the vehicle at its current location and update the current location entry based on the observation parameters.

[0047] A third aspect of this application provides a vehicle in which a lane keeping feedforward correction system driven by the cross slope spectrum described in the second aspect is deployed.

[0048] The beneficial effects of the embodiments of this application include:

[0049] In the lane keeping feedforward correction method driven by cross slope road spectrum provided in this application embodiment, when the vehicle travels to the target road, based on multi-source sensor fusion positioning, the road segment identifier to which the vehicle's current position belongs is first determined; accordingly, the current position entry is indexed using a position key (composed of road segment identifier, direction identifier, lane identifier, and traverse index), and N traverse position entries ahead are pre-fetched under the same road segment identifier and direction identifier. The average feedforward correction angle in the current position entry within the pre-fetching window and the average feedforward correction angle in the N traverse position entries ahead are aggregated and smoothed to obtain the initial feedforward correction angle; then, the variance and confidence of the feedforward correction angle in the current position entry and the variance and confidence of the feedforward correction angle in the N traverse position entries ahead (coverage and location confidence can be considered simultaneously) are combined to form a road spectrum gating coefficient, and the initial feedforward correction angle is low-pass and amplitude / rate of change is limited to obtain a low-frequency, restricted feedforward correction angle. Subsequently, the feedforward correction angle, along with the lateral position deviation feedback and azimuth deviation feedback in the parallel feedback structure, are used to generate the overall steering control command, which is then sent to the steering device to control the vehicle's movement on the road, achieving more stable lane centering. Simultaneously, upon successful steady-state gating and consistency verification, the vehicle uses the observed parameters at that moment as samples to update the statistics of the current position entry online, thereby improving the reliability and effectiveness of subsequent reuse.

[0050] In summary, this application can perform small-angle feedforward corrections based on the vehicle's position before it enters a known road segment, reducing the burden on the integral loop, suppressing overshoot and oscillation, and improving the accuracy, responsiveness, and stability of lane centering control. Moreover, it does not require additional sensors or external high-precision maps and has the ability to continuously learn and reuse across different applications.

[0051] Thus, this application can achieve the effect of making small position-related angle corrections before the vehicle enters a known road segment without the need for additional sensors, thereby reducing the burden on the integral loop and improving the accuracy and stability of lane centering control. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart of the first lane keeping feedforward correction method driven by cross slope road spectrum provided in the embodiments of this application;

[0054] Figure 2 A flowchart of the second lane keeping feedforward correction method driven by cross slope road spectrum provided in the embodiments of this application;

[0055] Figure 3 A flowchart of a third lane keeping feedforward correction method driven by cross slope spectrum provided in an embodiment of this application;

[0056] Figure 4 A flowchart of the fourth cross-slope spectrum-driven lane keeping feedforward correction method provided in the embodiments of this application;

[0057] Figure 5 This application provides a schematic diagram of a corner control command generation method.

[0058] Figure 6 A flowchart of the fifth cross-slope spectrum-driven lane keeping feedforward correction method provided in the embodiments of this application;

[0059] Figure 7 A flowchart of the sixth cross-slope road spectrum-driven lane keeping feedforward correction method provided in the embodiments of this application;

[0060] Figure 8 A flowchart illustrating a road spectrum data reading, writing, and pre-aiming method provided in this application embodiment;

[0061] Figure 9 A schematic diagram of a lane keeping feedforward correction system provided in an embodiment of this application;

[0062] Figure 10 This is a structural schematic diagram of a vehicle provided in an embodiment of this application. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0064] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0065] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0066] Currently, vehicles commonly employ a lateral control architecture based on road geometry and vehicle models, using a "feedforward + feedback correction" approach to achieve lane keeping assist and lane centering control. The feedforward component uses prior information such as road curvature to actively compensate for steering errors, while the feedback component relies on real-time measurements of lateral distance and heading deviations to perform closed-loop correction. However, vehicles built on this lateral control architecture experience systematic errors in the physical feedforward when entering cross-slope sections where cross-slope information is unavailable or delayed. Actual centering relies primarily on the integral loop to slowly offset these errors, easily leading to problems such as edge-grabbing, delayed centering, overshoot, and reduced phase margin. This severely impacts the stability of lane centering, significantly reducing passenger comfort.

[0067] To address this, this application provides a lane keeping feedforward correction method driven by cross slope road spectrum. The method involves acquiring parameters from multiple vehicle sensors in real-time and performing fusion positioning to determine the road segment identifier to which the vehicle's current position belongs. Based on a position key composed of the road segment identifier, direction identifier, lane identifier, and traverse index, the method indexes the current position entries under the same road segment identifier and direction identifier, and pre-fetches historical statistics of the preceding N traverse position entries according to the traverse index. The method determines the initial feedforward correction angle of the vehicle at the current moment based on the weighted aggregation and smoothing results of the mean feedforward correction angles in each position entry within the pre-aiming window. The initial feedforward correction angle is then subjected to low-pass filtering, amplitude limiting, and unit-time rate-of-change limitation to obtain a low-frequency, restricted feedforward correction angle. Based on the variance and confidence of the feedforward correction angle, the current position entry, and the forward correction angles in the preceding traverse entries, and combined with the lateral position deviation feedback and azimuth deviation feedback of the vehicle acquired in real-time by multiple sensors, a total steering angle control command is generated, and the vehicle's driving on the road is controlled accordingly. Therefore, the historical steady-state bias can be reused according to the vehicle's position on the road, and low-frequency restricted corrections can be made to the feedforward before the vehicle enters the known road section, thereby keeping the vehicle centered and improving the accuracy and stability of lane centering control.

[0068] Thus, this application can achieve the effect of making small position-related angle corrections before the vehicle enters a known road segment without the need for additional sensors, thereby reducing the burden on the integral loop and improving the accuracy and stability of lane centering control.

[0069] The following describes in detail the lane keeping feedforward correction method driven by cross slope spectrum provided in the embodiments of this application, with reference to the accompanying drawings.

[0070] Figure 1 A flowchart of a lane-keeping feedforward correction method driven by cross-slope road spectrum provided in this application is shown. This method can be applied to the controller in a vehicle. See also Figure 1 This application provides a lane keeping feedforward correction method driven by cross slope spectrum, including:

[0071] S101. When a vehicle is traveling on a road, determine the road segment marker to which the vehicle belongs at its current location.

[0072] Specifically, the local cross-slope road map deployed on the vehicle discretizes the road using a preset distance step size, and learns online based on the vehicle's historical observation parameters and location data on each road to form cross-slope road map entries indexed by location key (composed of road segment identifier, direction identifier, lane identifier, and distance index), so that the vehicle can quickly reuse the road map when it re-enters a known road. The preset distance step size can be 10 meters, 15 meters, 30 meters, etc., and this application does not limit it.

[0073] Optionally, "road" refers to the road on which the vehicle is currently traveling; "the road segment to which the current location belongs" refers to the road segment obtained by discretizing along the path based on the current location and the aforementioned step length.

[0074] It is worth noting that the roads are usually those that vehicles have previously traversed. After vehicles pass through multiple times, the local cross slope road spectrum learns online based on real-time observation data to obtain more reliable feedforward correction angle statistics, which can then be directly reused when vehicles pass through again, improving the accuracy of centering control.

[0075] Furthermore, if this is the first time the vehicle has passed through this road, there are no historical statistics available for reuse, and the vehicle can only rely on the traditional integral loop to slowly offset and achieve centering.

[0076] Specifically, when a vehicle is driving on the road, the current location of the vehicle is matched with the existing reference road segments based on the fusion positioning results to determine the road segment identifier to which the vehicle belongs at the current location; otherwise, if there is no matching reference road segment, a new road segment identifier is generated and registered for subsequent reuse.

[0077] Optionally, based on the multi-sensor fusion positioning results (such as satellite positioning, odometer, 2D inertial navigation, etc.), the road segment identifier to which the vehicle belongs at its current location is determined. This road segment identifier can be generated by lightweight clustering or hashing of historical positioning trajectories, or it can be determined by matching the base map with the real-time map; this application does not impose any limitation on this.

[0078] S102. Based on the location key, obtain the current location entry and the N preceding location entries. Each location entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level. The location key consists of the road segment identifier, direction identifier, lane identifier, and along-the-road index.

[0079] Optionally, the local cross-slope road spectrum determines the location key based on the vehicle's current location and indexes the matching road spectrum file in local storage, from which the current location entry is read. The current location entry refers to a record indexed by the aforementioned location key, and the record content is the statistics at that location: the mean, variance, and confidence level of the feedforward correction angle.

[0080] Specifically, the mean of the feedforward correction angle is the average of typical small DC angle corrections obtained from steady-state bias observations of the vehicle passing through the position multiple times; the variance of the feedforward correction angle characterizes the uncertainty / dispersion of this mean; the confidence score is a comprehensive rating of the reliability of this item, and its value ranges from [value missing]. It can be initialized to When the write conditions for steady-state gating and consistency verification are met, updates are performed using a recursive formula: each write operation first obtains the sample quality score based on the location reliability and the feedforward correction angle variance of the current entry. Then from the old confidence level and sample quality score The new confidence level is calculated. For details, please refer to formulas (17) and (18) below.

[0081] Optionally, after locating the current location entry, the system previews along the same road segment markers and direction markers using the traverse index, pre-retrieving N traverse location entries ahead (e.g., N is 3, 4, 5, etc., determined by the preview window size and the traverse step size). These "N traverse entries ahead" are continuous with the current location entry along the traverse, and their contents also include the mean, variance, and confidence level.

[0082] S103. Determine the feedforward correction angle based on the average of the feedforward correction angles in the current position entry and the average of the feedforward correction angles in the preceding N along-the-path position entries.

[0083] Among them, the feedforward correction angle is the steering compensation amount provided in advance based on learned statistics when the vehicle passes through the road segment again.

[0084] Optionally, the average value of the feedforward correction angle in the current position entry and the average value of the feedforward correction angle in the N forward position entries within the preview window are weighted, aggregated, and smoothed to obtain a smooth and stable feedforward correction angle, so as to ensure the continuity and controllability of the steering device output.

[0085] S104. Based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation, control the vehicle's movement on the road.

[0086] Optionally, the average confidence level is first determined based on the confidence levels of the current position entry and the N preceding position entries (which can be combined with the preview coverage and positioning confidence). Then, the road spectrum gating coefficient is calculated by combining the variance of the feedforward correction angle in the current position entry with other gating conditions. Next, based on the feedforward correction angle, the road spectrum gating coefficient, and the lateral position deviation feedback and azimuth deviation feedback obtained from the parallel feedback structure, a total steering control command is generated and sent to the steering device to ensure the vehicle stays centered. The deviation provided by the parallel feedback structure is used to correct the feedforward correction, improving the accuracy and robustness of lateral control.

[0087] S105. Obtain the observation parameters of the vehicle at its current location, and update the current location entry based on the observation parameters.

[0088] Optionally, the observation parameters are parameters collected in real time by multiple sensors during vehicle operation, including at least: mileage information, positioning information, inertial measurement information, etc.; the types of multiple sensors include precision positioning devices, odometers, two-dimensional inertial measurement units, etc., which are not limited in this application.

[0089] Optionally, when the steady-state gating and consistency check are satisfied, the bias observations corresponding to the observation parameters at this moment are used as valid samples to recursively update the mean, variance and confidence of the feedforward correction angle of the current position entry online, thereby improving the reliability and effectiveness of subsequent reuse.

[0090] Specifically, the current position entry is updated based on the observation parameters of the vehicle at the current position. When the steady-state gating and consistency verification are satisfied, the observation samples are used to recursively update the feedforward correction angle mean, variance and confidence level in the current position entry.

[0091] In this embodiment, when a vehicle travels to the target road, based on multi-source sensor fusion positioning, the road segment identifier to which the vehicle's current location belongs is first determined. Based on this, a location key (composed of road segment identifier, direction identifier, lane identifier, and traverse index) is used to index the current location entry, and N traverse location entries ahead are pre-fetched under the same road segment identifier and direction identifier. The average feedforward correction angle in the current location entry within the pre-fetching window and the average feedforward correction angle in the N traverse location entries ahead are aggregated and smoothed to obtain the initial feedforward correction angle. Then, the variance and confidence of the feedforward correction angle in the current location entry and the variance and confidence of the feedforward correction angle in the N traverse location entries ahead (coverage and location confidence can be considered simultaneously) are combined to form a road spectrum gating coefficient. The initial feedforward correction angle is then subjected to low-pass and amplitude / rate-of-change limitations to obtain a low-frequency, restricted feedforward correction angle. Subsequently, the feedforward correction angle, along with the lateral position deviation feedback and azimuth deviation feedback in the parallel feedback structure, are used to generate the overall steering control command, which is then sent to the steering device to control the vehicle's movement on the road, achieving more stable lane centering. Simultaneously, upon successful steady-state gating and consistency verification, the vehicle uses the observed parameters at that moment as samples to update the statistics of the current position entry online, thereby improving the reliability and effectiveness of subsequent reuse.

[0092] In summary, this application can perform small-angle feedforward corrections based on the vehicle's position before it enters a known road segment, reducing the burden on the integral loop, suppressing overshoot and oscillation, and improving the accuracy, responsiveness, and stability of lane centering control. Moreover, it does not require additional sensors or external high-precision maps and has the ability to continuously learn and reuse across different applications.

[0093] Thus, this application can achieve the effect of making small position-related angle corrections before the vehicle enters a known road segment without the need for additional sensors, thereby reducing the burden on the integral loop and improving the accuracy and stability of lane centering control.

[0094] In one alternative implementation, see [link to implementation details]. Figure 2 The operation of "determining the road segment identifier to which the current location belongs" in step S101 above can be specifically as follows:

[0095] S201, The integral component of the lateral deviation controller of the sampling vehicle is used as the observation sample.

[0096] Optionally, the integral component (steady-state sampling) of the lateral deviation controller is sampled as an offset observation for consistency verification with the road spectrum data.

[0097] Alternatively, the integral component of the vehicle's lateral deviation controller can be determined according to the following formula (1), which is as follows:

[0098] (1)

[0099] in, Used to represent the integral component of the vehicle's lateral deviation controller. Used to indicate the lateral distance deviation of a vehicle. Used to indicate the preset integral gain.

[0100] Specifically, the integral component of the lateral deviation controller refers to the integral component output by the PID controller in the parallel feedback structure of the vehicle. The preset integral gain is a calibration parameter used to determine the strength of the integral action.

[0101] Alternatively, the road spectrum label can be determined according to the following formula (2), which is as follows:

[0102] (2)

[0103] in, Road map labels used to indicate the road segment where a vehicle is located. Used to represent the integral component of the vehicle's lateral deviation controller.

[0104] Specifically, the road spectrum label is the sampled value of the integral of the lateral distance deviation of the lateral PID controller in the parallel feedback structure under the steady-state gating condition.

[0105] S202. Determine the position key based on the parameters collected by at least one sensor on the vehicle.

[0106] Alternatively, the position key can be determined according to the following formula (3), which is as follows:

[0107] (3)

[0108] in, This represents the location key, used for indexing within the local cross slope road map; The road segment identifier, which represents the road segment, can be obtained by segmenting the vehicle's historical GNSS trajectory by road segment and then performing lightweight clustering or hashing, ensuring that the same road segment is mapped to a stable and consistent unique identifier when passed multiple times; This indicates the directional value for that road segment, used to distinguish between two-way roads; This indicates that lane markings on this road segment can be disabled or set to the default value when only meter-level positioning is available. It is the total distance along the route with the starting point of the road segment as zero. Used to represent the mileage raster index (mileage raster number) within this road segment.

[0109] Specifically, the location key represents the inherent spatial position within the road segment: based on the location key, the current location entry can be quickly located, and several entries ahead can be previewed in the order of the along-the-road index to support the aggregation, smoothing and gating of the feedforward correction angle.

[0110] Optionally, road segment identifiers can be generated from vehicle historical positioning trajectories through lightweight clustering or hashing, or they can be provided by a combination of a base map and real-time map matching results. The goal is to establish a stable and consistent unique identifier for the same road segment in the absence of a high-precision map.

[0111] It should be noted that when a vehicle passes a target road for the first time, there are no historical statistics for the corresponding location key, and only basic control can be relied upon. After passing through multiple times and accumulating samples, the mean, variance, and confidence of the items will gradually converge and can be reused in subsequent driving to improve the control effect.

[0112] It should also be noted that lane markings are optional labels. When there are no lane markings on the road segment where the vehicle is located, the lane marking will default to 1. When there are lane markings on the road segment where the vehicle is located, the corresponding lane marking will be selected according to the lane where the vehicle is located. This application does not make specific restrictions on this.

[0113] Optionally, when a base map is available, road segment identifiers can be obtained by linear map matching based on the target road's ID and driving direction. When a base map is unavailable, stable road segment hashes can be generated by lightweight clustering or hash calculation based on the target road's historical positioning trajectory, thus forming stable road segment identifiers. This application primarily uses a map-free solution, but this does not imply that this application can only establish road segment identifiers using a map-free solution; this application does not impose specific limitations in this regard.

[0114] It is worth noting that road segment signs only need to be created when a vehicle first passes through a certain road. When a vehicle passes through a known road again, it can directly reuse the road segment signs of each segment on that road.

[0115] Optionally, the driving direction is determined based on the vehicle's heading or speed direction on its current road segment. Based on the vehicle's driving direction, each road segment can be divided into two segments, one for travel in each direction, thereby avoiding confusion in the lateral offset control data.

[0116] Alternatively, the integral of the distance traveled can be determined by the following formula (4), which is as follows:

[0117] (4)

[0118] in, Used to represent the integral of the distance traveled along the route at the current moment. Used to represent the integral of the distance traveled along the route at the previous moment. Used to represent the longitudinal speed of the vehicle at the current moment. Used to indicate the sampling time interval.

[0119] Specifically, the mileage integral refers to the cumulative distance the vehicle has traveled along the centerline of the road from the starting point of the target road to the current position; the mileage integral of the previous moment refers to the mileage value of the vehicle in the previous calculation cycle; the vehicle's longitudinal speed refers to the instantaneous speed of the vehicle in its forward direction, usually provided by the wheel speed odometer; and the sampling time interval refers to the time difference between two consecutive calculation cycles.

[0120] It should be noted that the mileage integral along the route is determined by using wheel speed odometer data as the primary data, performing global correction through the positioning cycle of the positioning device (GNSS), and using IMU for short-term dead reckoning.

[0121] Optionally, the discretized mileage grid corresponding to the road segment can be determined by the following formulas (5) and (6), which are as follows:

[0122] (5)

[0123] (6)

[0124] in, The raster index (raster number) representing the distance traveled within this road segment is used, and s represents the distance traveled at the current time. Used to indicate the preset grid length, The standard deviation used to represent positioning error.

[0125] Specifically, the above formula (6) is a constraint condition of the mileage grid along the local cross slope spectrum, which is used to weigh the engineering design of the mileage grid along the route from multiple perspectives such as spatial accuracy, positioning error and system stability.

[0126] It should be noted that the standard deviation of the positioning error represents the uncertainty of the vehicle's positioning in the lateral direction. The standard deviation of the positioning error is given by the fusion positioning algorithm of multi-sensor data sampled from GNSS, wheel speed odometer and IMU.

[0127] It is worth noting that, It is the minimum empirical length set by the local cross slope road spectrum, thereby avoiding the target road being over-fragmented; The positioning error is set to six times the local cross slope road spectrum, which ensures that even under the worst-case positioning error, the calculated position key remains stable and will not jump to adjacent grids. This engineering design is only an example and does not mean that this application can only set the grid length according to this constraint; this application does not make any specific limitations in this regard.

[0128] S203. Extract the road segment identifier from the location key to determine the road segment identifier to which the vehicle belongs at its current location.

[0129] In one alternative implementation, see [link to implementation details]. Figure 3 The specific operation of step S103 above can be as follows:

[0130] S301. Aggregate and smooth the mean of the feedforward correction angle in the current position entry and the mean of the feedforward correction angle in the preceding N path entries to obtain the initial feedforward correction angle.

[0131] Optionally, the initial feedforward correction angle refers to the result obtained by aggregating and smoothing the mean of the feedforward correction angle of the current position entry and the mean of the feedforward correction angle of the preceding path entry in the local cross slope spectrum.

[0132] Among them, the aggregation process is used to merge the mean values ​​of the feedforward correction angles in multiple discrete mileage grids pre-fetched by the local cross slope road spectrum in the aiming window to obtain a more statistically significant value, thereby fusing multiple discrete point information into continuous segment information, which can effectively suppress the influence of individual abnormal point information.

[0133] Specifically, the following methods can be used to aggregate the mean feedforward correction angle of the current position entry with the mean feedforward correction angle of multiple preceding mileage entries: The first method is to take the arithmetic mean of the mean feedforward correction angles of multiple mileage gratings pre-fetched within the preview window to obtain the aggregation result; the second method is to average the mean feedforward correction angles of multiple mileage gratings pre-fetched within the preview window using the confidence level as the weight to obtain the aggregation result; the third method is to interpolate and complete the mileage gratings without data within the preview window using the mileage gratings with data before and after them, and then perform an arithmetic mean to obtain the aggregation result.

[0134] In addition, smoothing is used to filter the aggregated results obtained from the local cross slope road spectrum to preserve the low-frequency variation trend of the mean of the feedforward correction angle, thereby suppressing high-frequency fluctuations and noise in the aggregated results.

[0135] Specifically, the aggregation results can be smoothed using any of the following methods: The first method is to smooth using a moving average / sliding window, such as sliding a fixed-length window across the aggregation results and calculating the average value of the data within the window each time as the smoothed value for that point; the second method is to use a low-pass filter to filter out signal components in the aggregation results that are higher than the cutoff frequency; the third method is to smooth using spline interpolation, such as fitting discrete aggregation points with a smooth curve.

[0136] S302. Correct the initial feedforward correction angle to obtain the feedforward correction angle.

[0137] Optionally, the feedforward correction angle is the feedforward control quantity that participates in the generation of the total steering command from the local cross slope spectrum output.

[0138] In one optional implementation, step S302 can specifically be performed as follows:

[0139] The initial feedforward correction angle is low-pass filtered, limited, and its rate of change is limited to obtain the feedforward correction angle.

[0140] Optionally, by performing low-pass filtering and amplitude limiting on the initial feedforward correction angle, the final feedforward correction angle participating in vehicle steering control can be obtained.

[0141] In one optional implementation, the initial feedforward correction angle is low-pass filtered, limited, and its rate of change is limited based on the following formulas (7) and (8) to obtain the feedforward correction angle. The specific formulas (7) and (8) are as follows:

[0142] (7)

[0143] (8)

[0144] in, Indicates the feedforward correction angle. This represents the initial feedforward correction angle. This represents the physical feedforward angle of the monorail vehicle model. This represents the absolute value of the feedforward correction angle. This represents the maximum allowable feedforward correction angle threshold for the vehicle. This represents the rate of change of the vehicle's feedforward correction angle. This represents the threshold for the maximum allowable rate of change of the vehicle's feedforward correction angle. and Indicates the calibration parameters.

[0145] Among them, the physical feedforward angle of the monorail vehicle model is the theoretical steering angle calculated by the monorail vehicle dynamics model based on the curvature of the road, the vehicle speed and the inherent parameters of the vehicle; the maximum feedforward correction angle is the physical safety red line pre-set according to the execution capability of the vehicle's electric power steering system and the stability of the vehicle.

[0146] Specifically, the above formula (8) is used to limit the amplitude of the feedforward correction angle. It is a dynamic limitation on the feedforward correction angle, which can ensure that the road spectrum correction amount matches the current driving environment of the vehicle; It is an absolute upper limit on the feedforward correction angle; the feedforward correction angle cannot exceed the maximum allowable feedforward correction angle value.

[0147] It should be noted that the rate of change of the feedforward correction angle is specifically determined by the derivative of the feedforward correction angle (which can be achieved by first-order difference plus low-pass filtering), and the rate of change of the maximum feedforward correction angle is specifically determined by the derivative of the maximum feedforward correction angle. This application does not make specific limitations on this.

[0148] Optionally, calibration parameters and These are standard parameters and calibration parameters preset by the user. The range of values ​​is Calibration parameters The range of values ​​is This application does not impose specific limitations on this.

[0149] In one alternative implementation, see [link to implementation details]. Figure 4 The specific operation of step S104 above can be as follows:

[0150] S401. Determine the path spectrum gating coefficients based on the feedforward correction angle, the variance and confidence level of the feedforward correction angle in the current position entry, and the variance and confidence level of the feedforward correction angle in the preceding N path position entries.

[0151] Specifically, under the same road segment identification / direction, based on the coverage of the preview window, the mean confidence level, the variance of the feedforward correction angle of the current location entry, and the location confidence level, the road spectrum gating coefficient can be determined by the following formula (9) or formula (10), as follows:

[0152] , , , , , (9)

[0153] , (10)

[0154] in, Used to represent the road spectrum gating coefficient. The confidence level of the trail mileage raster prefetched by the aiming window to represent the local cross slope road spectrum. The number of grid cells along the route. The total number of mileage grid cells pre-fetched by the preview window to represent the local cross slope road spectrum. Used to indicate coverage. The average confidence level of all along-distance mileage gratings pre-fetched by the preview window to represent the local cross-slope road spectrum. The variance used to represent the feedforward correction angle in the current location entry (segment identifier / direction / alongline index). The variance used to represent the maximum feedforward correction angle. Used to indicate the confidence level in the current location entry. Used to represent minimum confidence level Used to represent the range of values ​​for the minimum confidence level. Used to represent the range of values ​​for the path spectrum gating coefficient. Used to represent location reliability Used to represent minimum location confidence.

[0155] Optionally, the road spectrum gating coefficient, as a dynamic weighting coefficient, is used to precisely control the degree of influence of the road spectrum experience learned from the local cross slope road spectrum on the actual vehicle steering. This indicates that the local cross slope road spectrum fully trusts the current cross slope road spectrum experience and uses the complete feedforward correction angle in steering control; when This indicates that the local cross-slope road spectrum completely disregards the current cross-slope road spectrum experience, disables the road spectrum correction function, and relies on the basic control structure; when This indicates that the local cross slope road spectrum part relies on the current cross slope road spectrum experience and reduces the influence of the road spectrum correction amount proportionally.

[0156] Optionally, coverage is used to evaluate the completeness and continuity of information from multiple road segments pre-fetched within the preview window, and location reliability is used to evaluate the credibility of the current location information provided by the vehicle positioning system.

[0157] S402. Generate steering control commands based on feedforward correction angle, road spectrum gating coefficient, lateral distance deviation feedback, and azimuth deviation feedback.

[0158] Among them, the steering angle control command is the control command that the lane centering control system in the vehicle ultimately issues to the vehicle's steering device. The vehicle's steering device controls the vehicle's steering under the action of the steering control command.

[0159] Optionally, the lateral distance deviation feedback in the vehicle centering deviation can be determined by the following formula (11), and the direction angle deviation feedback in the vehicle centering deviation can be determined by the following formula (12). Formulas (11) and (12) are as follows:

[0160] (11)

[0161] (12)

[0162] in, Used to represent the lateral distance deviation feedback amount. Used to represent the gain of lateral distance deviation. Used to represent the integral gain of lateral distance deviation. Used to represent the differential gain of lateral distance deviation. Used to indicate the actual lateral distance deviation of a vehicle. Used to represent the rate of change of the actual lateral distance deviation of a vehicle. Used to represent the azimuth angle deviation feedback amount. Used to represent the azimuth deviation proportional gain. Used to indicate the actual azimuth deviation of a vehicle. Used to represent the differential gain of azimuth angle deviation Used to represent the rate of change of the actual azimuth angle deviation of a vehicle.

[0163] Among them, the lateral distance deviation proportional gain, lateral distance deviation integral gain, and lateral distance deviation derivative gain are all determined by the PID controller in the parallel feedback structure. These parameters together determine how the lane centering system converts the actual lateral distance deviation of the vehicle into steering correction commands.

[0164] Specifically, the lateral distance deviation proportional gain is used to control the actual lateral distance deviation of the vehicle at the current moment, the lateral distance deviation integral gain is used to control the historical cumulative lateral distance deviation of the vehicle, and the lateral distance deviation differential gain is used to control the trend of the lateral distance deviation of the vehicle.

[0165] Furthermore, both the azimuth deviation proportional gain and the azimuth deviation differential gain are determined by the PD controller in the parallel feedback structure. The azimuth deviation proportional gain is used to generate a steering angle proportional to the actual azimuth deviation of the vehicle, thereby turning the front of the vehicle back to the correct direction. The azimuth deviation differential gain, as the anti-sway damper of the PD controller, is used to suppress yaw oscillation and generate a damping torque proportional to the rate of change of azimuth, thereby preventing overshoot when the front of the vehicle returns to center quickly.

[0166] Optionally, the lane centering control system in the vehicle dynamically generates the total steering control command for the vehicle based on the feedforward correction angle given by the local cross slope spectrum and the lateral distance deviation feedback and azimuth deviation feedback given by the parallel feedback structure, so as to control the steering action of the vehicle.

[0167] S403. Send the steering angle control command to the vehicle steering device to control the vehicle's movement on the road.

[0168] Optionally, the steering device in the vehicle can be any of the front wheels, steering wheel, or other devices of the vehicle, and this application does not specifically limit it.

[0169] In one optional implementation, the steering control command can be determined by the following formula (13), which is as follows:

[0170] , (13)

[0171] in, This represents the physical feedforward angle of the monorail vehicle model. Represents the spectrum gating coefficient. This indicates the feedforward correction angle of the vehicle at its current position. This indicates the amount of lateral distance deviation feedback for the vehicle. This indicates the vehicle's azimuth deviation feedback amount.

[0172] In one alternative approach, see Figure 5The specific process of generating steering control commands in the lane keeping feedforward correction method driven by cross slope road spectrum provided in this application embodiment is as follows: S1. The real-time lateral distance deviation and real-time azimuth deviation of the vehicle at the current moment are obtained through the vehicle's front-facing camera sensor; S2. The PID controller in the parallel feedback structure determines the lateral distance deviation feedback amount of the vehicle at the current moment based on the real-time lateral distance deviation; at the same time, the PD controller in the parallel feedback structure determines the azimuth deviation feedback amount of the vehicle at the current moment based on the real-time azimuth deviation, and determines the total lateral distance deviation feedback amount of the vehicle at the current moment based on the lateral distance deviation feedback amount and the azimuth deviation feedback amount of the vehicle at the current moment; S3. The total steering control command is generated based on the physical feedforward angle calculated by the monorail vehicle model, the feedforward correction angle provided by the local cross slope road spectrum, and the total vehicle centering deviation feedback amount (lateral distance deviation feedback amount and azimuth deviation feedback amount of the vehicle) output by the parallel feedback structure.

[0173] In one alternative implementation, see [link to implementation details]. Figure 6 The specific operation of step S105 above can be as follows:

[0174] S601. Based on the observed parameters, determine whether the vehicle meets the steady-state gating conditions.

[0175] Optionally, the observation parameters are data collected in real time by various sensors deployed in the vehicle. The observation parameters include: the real-time lateral distance deviation of the vehicle at the current moment, the real-time azimuth deviation of the vehicle at the current moment, the actuator saturation state of the vehicle at the current moment, the lateral control state of the vehicle at the current moment, and the positioning reliability of the vehicle.

[0176] It should be noted that the steady-state gating condition, as the plotting condition of the lane keeping feedforward correction method driven by the cross slope road spectrum provided in this application, can only write the integral of the lateral distance deviation of the vehicle at the current moment into the discretized mileage grid corresponding to the road segment when the vehicle meets the steady-state gating condition.

[0177] S602. If so, determine the current quality weight.

[0178] Optionally, the quality weight corresponding to the current sample data can be determined by the following formula (14), which is as follows:

[0179] (14)

[0180] in, Used to represent the current quality weight, Used to represent the location reliability quality of a vehicle. Used to represent the dynamic state quality of a vehicle. Used to indicate vehicle speed and mass. Used to indicate the saturation mass of the actuators of a vehicle.

[0181] Among them, the positioning reliability quality can be obtained by normalizing the vehicle's positioning information or position covariance; the dynamic state quality can be obtained by the absolute value of the integral of the lateral distance deviation of near-straight line or small error; the vehicle speed quality can be obtained by whether the vehicle's current driving speed is within the safe speed range; and the actuator saturation quality can be determined by the control state of the actuator.

[0182] Specifically, positional reliability reflects the accuracy and reliability of the vehicle's current position information and is used to evaluate the quality of the fusion positioning results from GNSS, wheel speed odometer, and IMU; dynamic state quality reflects the balance and purity of the vehicle's lateral motion at the current moment and is used to evaluate whether the vehicle is in an ideal cruising state unaffected by cross slopes; vehicle speed quality reflects whether the vehicle's current speed is within the vehicle's effective speed range; and actuator saturation quality reflects the linear operating state of the steering actuator and lane centering control system to detect whether the steering actuator (EPS) is saturated or whether the stability control system (ESC / ESP) intervenes in the vehicle's lateral control.

[0183] S603. Based on the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, update the mean value of the feedforward correction angle in the current position entry to obtain the new mean value of the feedforward correction angle.

[0184] Alternatively, the mean value of the new feedforward correction angle can be obtained by the following formula (15), which is as follows:

[0185] , (15)

[0186] in, Used to represent the mean of the new feedforward correction angle. Used to represent the preset weighted recursive coefficients, This is used to represent the mean of the original feedforward correction angles in the current location entry for this road segment. Used to represent the current quality weight, The road spectrum label used to indicate the road segment where the vehicle is located is the integral component of the lateral deviation controller obtained from steady-state sampling.

[0187] Optionally, the preset weighted recursion coefficients are user-defined basic weighted recursion coefficients, and the range of values ​​for the preset weighted recursion coefficients is [range missing]. .

[0188] S604. Based on the mean of the new feedforward correction angle, the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, update the variance of the feedforward correction angle to obtain the new variance of the feedforward correction angle.

[0189] Alternatively, the variance of the new feedforward correction angle can be obtained through the following formula (16), which is as follows:

[0190] , (16)

[0191] in, The variance used to represent the new feedforward correction angle Used to represent the preset weighted recursive coefficients, This is used to represent the variance of the original feedforward correction angle in the current location entry for this road segment. Used to represent the current quality weight, Road map labels used to indicate the road segment where a vehicle is located. Used to represent the mean of the feedforward correction angle.

[0192] S605. Adjust the step size based on the location confidence, the variance of the new feedforward correction angle, and the preset confidence, update the confidence, and obtain the new confidence.

[0193] Alternatively, a new confidence level can be determined using the following formulas (17) and (18), which are as follows:

[0194] , , ,

[0195] (18)

[0196] in, Used to represent a new confidence level. Used to indicate the preset confidence level adjustment step size. Used to indicate the original confidence level in the current location entry of this road segment. Used to represent the quality score of this sample data Used to represent location reliability The variance used to represent the new feedforward correction angle Used to represent a preset adjustable reference scale constant.

[0197] Optionally, the preset confidence adjustment step size is the confidence adjustment step size preset by the user, and the preset adjustable reference scale constant is the variance value of the reference feedforward correction angle preset by the user. This application does not specifically limit this.

[0198] It is worth noting that the smaller the variance of the feedforward correction angle obtained from the observation parameters collected by the vehicle at the current moment, the more reliable it is to use the observation parameters collected by the vehicle at the current moment as valid sample data.

[0199] It should also be noted that if the vehicle does not meet the conditions for writing the map at the current moment, the current location entry of that segment of the target road where the vehicle is located does not need to be updated.

[0200] In an optional implementation, the lane keeping feedforward correction method driven by cross slope road spectrum provided in this application also needs to prevent expired data from continuing to participate in road spectrum correction control. The expired data may be caused by factors such as slow changes in road and vehicle status over time, road repaving, shoulder deformation, seasonal changes, and tire pressure changes. The expiration of road segment information corresponding to each road segment can be determined by the following formula (19). Formula (19) is as follows:

[0201] (19)

[0202] in, The current timestamp used to represent road segment information. The timestamp used to represent the last updated data of road segment information. Used to indicate the longest period for updating road segment information.

[0203] Optionally, if the time difference between the current timestamp of the road segment information and the timestamp of the last data update exceeds the preset maximum period, then when reading the road segment information, the confidence level in the road segment information needs to be set to zero, and the data in the road segment information will not be started by default.

[0204] In one alternative implementation, see [link to implementation details]. Figure 7 The specific operation of step S601 above can be as follows:

[0205] S701. Obtain the observation parameters of the current location, including: road curvature, vehicle lateral distance deviation, and vehicle azimuth angle deviation.

[0206] Optionally, various sensors deployed in the vehicle can be used to acquire the road curvature, real-time lateral distance deviation, and real-time azimuth deviation of the road where the vehicle is currently located. Road curvature indicates the degree of road curvature in the segment where the vehicle is located; real-time lateral distance deviation indicates the distance deviation between the vehicle and the centerline of its lane at the current moment; and real-time azimuth deviation is the deviation between the vehicle's current direction of travel and the tangent direction of the centerline of its lane.

[0207] S702. Determine the rate of change of the lateral distance deviation integral based on the integral of the vehicle's lateral distance deviation.

[0208] Optionally, the rate of change of the vehicle lateral distance deviation integral can be obtained by first-order difference plus low-pass calculation, which is used to characterize the rate of change of the lateral distance deviation integral.

[0209] S703. Determine the location reliability based on the parameters collected by at least one sensor on the vehicle.

[0210] Optionally, the collected parameters refer to the parameters actually collected by GNSS, wheel speed odometer and IMU on the vehicle, such as positioning data, mileage along the route, wheel speed and other data. The positioning reliability of the vehicle is determined based on the fusion positioning results of the collected parameters from multiple sensors on the vehicle.

[0211] S704. Based on the observed parameters, the rate of change of the integral of the lateral distance deviation, the positional confidence, the saturation state of the vehicle actuator, and the intervention state of the stability controller, determine whether the vehicle meets the steady-state gating conditions.

[0212] Optionally, the vehicle's steady-state gating condition can be determined according to the following formulas (20), (21), (22), and (23), which are as follows:

[0213] , (20)

[0214] , , (twenty one)

[0215] (twenty two)

[0216] (twenty three)

[0217] in, Used to indicate the road curvature of the current road segment where the vehicle is located. Used to indicate the rate of change of road curvature in the current road segment where the vehicle is located. Used to represent the preset threshold for the rate of change of road curvature. Used to represent a preset road curvature threshold. Used to indicate the vehicle's current actual lateral distance deviation. Used to indicate the rate of change of the vehicle's current actual lateral distance deviation. Used to indicate the vehicle's current actual azimuth deviation. Used to represent the preset maximum lateral distance deviation threshold. Used to represent the preset threshold for the rate of change of lateral distance deviation. Used to represent the preset maximum azimuth deviation threshold. Used to represent the absolute value of the rate of change of the integral of the vehicle's current lateral distance deviation. Used to represent the preset threshold for the rate of change of the integral of the lateral distance deviation. Used to represent the current location confidence of the vehicle. Used to represent the preset location confidence benchmark value.

[0218] Optionally, the vehicle’s current location can be determined based on the above formula (20) whether it is a near-straight road.

[0219] Optionally, if the absolute value of the road curvature is less than a preset road curvature threshold, and the absolute value of the road curvature change rate is less than a preset road curvature change rate threshold, then the current road segment where the vehicle is located is determined to be a near-straight road segment; and the absolute value of the vehicle's actual lateral distance deviation is less than a preset maximum lateral distance deviation threshold, the absolute value of the vehicle's actual lateral distance deviation change rate is less than a preset lateral distance deviation change rate threshold, and the absolute value of the vehicle's actual azimuth deviation is less than a preset maximum azimuth deviation threshold; and the absolute value of the vehicle's lateral distance deviation integral change rate is less than a preset lateral distance deviation integral change rate threshold, and the duration reaches a preset value; and the saturation state of the vehicle's actuator is unsaturated, the connection state of the stability controller is not connected to lane centering control; and the vehicle's current positioning confidence is greater than a preset positioning confidence benchmark value, then the vehicle is determined to meet the steady-state gating conditions at the current moment.

[0220] In one optional implementation, after the lane keeping feedforward correction method driven by cross slope road spectrum provided in this application selects the road spectrum entry corresponding to the current position in the local cross slope road spectrum, it is also necessary to perform a consistency check with the integral component (steady-state sampling) of the vehicle's real-time lateral deviation controller to determine whether to update the current position entry of the road segment as a valid sample using the vehicle's observation data.

[0221] Specifically, the consistency of the vehicle's real-time lateral distance deviation integral is verified using the following formula (24), which is as follows:

[0222] (twenty four)

[0223] in, This is used to represent the mean of the vehicle's real-time feedforward correction angle. Used to represent the integral of the real-time lateral distance deviation of a vehicle. Used to indicate the preset tolerance value.

[0224] Optionally, when the mean of the real-time feedforward correction angle of the vehicle and the integral of the real-time lateral distance deviation satisfy formula (24), the observation parameters of the vehicle at the current moment are used as valid samples to update the current position entry of the road segment; when the mean of the real-time feedforward correction angle of the vehicle and the integral of the real-time lateral distance deviation do not satisfy formula (24), the road spectrum gating coefficient of the vehicle is quickly set to zero, and the update process of the current position entry of the road segment is paused.

[0225] In one optional implementation, when the lane keeping feedforward correction method driven by the cross slope road spectrum provided in this application enables road spectrum feedforward correction, it is necessary to implement leakage of the integral channel of the lateral distance deviation PID controller in the parallel feedback structure (such as reducing the lateral distance deviation integral gain, or freezing the integral on straight lines) to avoid phase drag caused by overlapping with the low-frequency part of the road spectrum gating coefficient.

[0226] In one alternative implementation, see [link to implementation details]. Figure 8 The lane keeping feedforward correction method driven by cross slope road spectrum provided in this application includes a map writing process and a road spectrum correction process. The specific implementation process is the same as the above method embodiment, and will not be described in detail here.

[0227] The following describes the system and vehicle used to implement the lane keeping feedforward correction method driven by the cross slope spectrum provided in this application. The specific implementation process and technical effects are described above and will not be repeated below.

[0228] Figure 9 This is a schematic diagram of a lane keeping feedforward correction system provided in this application. See also: Figure 9 The lane keeping feedforward correction system 10 includes:

[0229] When a vehicle is traveling on a road, the determination module 901 is used to determine the road segment identifier to which the vehicle belongs at its current location.

[0230] Specifically, the determination module 901 uses GNSS / odometer / 2D IMU fusion positioning to match the vehicle's current location with the existing set of reference road segments to determine the road segment identifier to which the current location belongs; when there is no matching road segment, it generates and registers a new road segment identifier for subsequent reuse.

[0231] The acquisition module 902 is used to acquire the current position entry under the same road segment identifier and direction identifier, as well as the N forward position entries based on the position key. Each position entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level. The position key consists of the road segment identifier, direction identifier, lane identifier, and along the route index.

[0232] Specifically, the acquisition module 902 is used to obtain the current location entry of the road segment based on the road segment identifier, and to obtain several entries ahead under the same road segment identifier and direction according to the traverse index. Each entry includes the mean, variance, and confidence level of the feedforward correction angle;

[0233] The feedforward correction module 903 is used to determine the feedforward correction angle based on the average of the feedforward correction angles in the current position entry and the average of the feedforward correction angles in the preceding N along-the-path position entries.

[0234] Specifically, the feedforward correction module 903 is used to perform window aggregation / interpolation and smoothing processing based on the average feedforward correction angle in the current position entry and the average feedforward correction angle in the previous entry to determine the initial feedforward correction angle; and to perform low-pass and amplitude / rate limiting on the initial feedforward correction angle to obtain the final feedforward correction angle.

[0235] The steering control module 904 is used to control the vehicle's movement on the road based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation.

[0236] Specifically, the steering control module 904 is used to determine the road spectrum gating coefficient based on the feedforward correction angle, the current position, and the variance and confidence of the preceding entry; and to generate steering angle control commands based on the feedforward correction angle, the road spectrum gating coefficient, the vehicle lateral distance deviation, and the vehicle azimuth angle deviation, in order to control the vehicle's driving on this road segment.

[0237] The write module 905 is used to obtain the observation parameters of the vehicle at its current location and update the current location entry based on the observation parameters.

[0238] Specifically, the write module 905 is used to sample the integral component (steady-state sampling) of the lateral deviation controller as an observation sample when the consistency check and steady-state gating are satisfied, and to perform quality-weighted recursive updates on the mean, variance and confidence of the feedforward correction angle of the current position entry.

[0239] Optionally, the determination module 901 is specifically used for: sampling the integral component of the lateral deviation controller of the vehicle as an observation sample; determining the position key based on parameters collected by at least one sensor on the vehicle; and extracting the road segment identifier from the position key to determine the road segment identifier to which the vehicle belongs at its current position.

[0240] Specifically, the determination module 901 determines the location key based on the collected parameters of at least one sensor on the vehicle, which is used to index the road spectrum; when there is no matching road segment identifier, a new road segment identifier is generated and registered, and a segment-level directional representative value and a zero point along the route are established for it to support the positioning of subsequent current location entries.

[0241] Optionally, the feedforward correction module 903 is specifically used to: aggregate and smooth the average value of the feedforward correction angle in the current position entry and the average value of the feedforward correction angle in the preceding N along-the-path position entries to obtain the initial feedforward correction angle; and correct the initial feedforward correction angle to obtain the feedforward correction angle.

[0242] Specifically, the feedforward correction module 903 performs window aggregation and sliding window smoothing / interpolation on the average feedforward correction angle of the current position entry and the N forward path entries to obtain the initial feedforward correction angle; and performs low-pass and amplitude / rate limiting on the initial feedforward correction angle to obtain the feedforward correction angle.

[0243] Optionally, the steering control module 904 is specifically used to: determine the road spectrum gating coefficient based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, and the variance and confidence of the feedforward correction angle in the N forward position entries; generate a steering angle control command based on the feedforward correction angle, the road spectrum gating coefficient, the lateral distance deviation feedback, and the azimuth angle deviation feedback; and send the steering angle control command to the steering device in the vehicle to control the vehicle's movement on the road.

[0244] Specifically, the steering control module 904 determines the road spectrum gating coefficient based on the variance and confidence level of the current position and the items ahead; it synthesizes the steering angle control command based on the feedforward correction angle, the road spectrum gating coefficient, the vehicle lateral distance deviation feedback, and the vehicle azimuth angle deviation feedback; and sends the steering angle control command to the vehicle steering device to control the vehicle's driving on the road.

[0245] Optionally, the writing module 905 is specifically used for: determining whether the vehicle meets the steady-state gating conditions based on the observed parameters; if so, determining the current mass weight; updating the mean of the feedforward correction angle in this entry based on the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, to obtain a new mean of the feedforward correction angle; updating the variance of the feedforward correction angle based on the new mean of the feedforward correction angle, the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, to obtain a new variance of the feedforward correction angle; and updating the confidence level based on the positional confidence level, the variance of the new feedforward correction angle, and the preset confidence level adjustment step size, to obtain a new confidence level.

[0246] Optionally, the writing module 905 is further configured to: acquire the observation parameters corresponding to the current position, including road curvature, vehicle lateral distance deviation, vehicle azimuth angle deviation, etc.; calculate the rate of change of the integral component of the lateral deviation controller; determine the positioning reliability based on the parameters collected by at least one sensor on the vehicle; and determine whether the vehicle meets the steady-state gating conditions based on the observation parameters, the rate of change of the integral component, the positioning reliability, the saturation state of the vehicle actuator, and the intervention state of the stability controller.

[0247] The lane keeping feedforward correction system described above is used to execute the method provided in the foregoing embodiments. Its implementation principle and technical effect are similar, and will not be described again here.

[0248] These modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more microprocessors, or one or more Field Programmable Gate Arrays (FPGAs). Alternatively, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together as a System on Chip (SOC).

[0249] Figure 10 This is a structural schematic diagram of a vehicle provided in this application. See also: Figure 10 The vehicle 20 is equipped with the aforementioned lane keeping feedforward correction system 10.

[0250] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the various method embodiments described above.

[0251] Optionally, this application also provides a program product, such as a computer-readable storage medium, including a program that, when executed by a processor, is used to perform any of the above-described embodiments of the lane keeping feedforward correction method driven by cross slope spectrum.

[0252] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0253] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0254] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0255] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute certain steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0256] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0257] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A lane keeping feedforward correction method driven by cross slope road spectrum, characterized in that, The method includes: When a vehicle is traveling on a road, determine the road segment markers to which the vehicle belongs at its current location; Based on the location key, the current location entry under the same road segment identifier and direction identifier, as well as the N preceding location entries along the route, are obtained. Each location entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level. The location key is composed of the road segment identifier, direction identifier, lane identifier, and route index. The feedforward correction angle is determined based on the average of the feedforward correction angles in the current position entry and the average of the feedforward correction angles in the N preceding position entries. The vehicle's movement on the road is controlled based on the feedforward correction angle, the variance and confidence level of the feedforward correction angle in the current position entry, the variance and confidence level of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation. Obtain the observation parameters of the vehicle at the current location, and update the current location entry based on the observation parameters.

2. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 1, characterized in that, The determination of the road segment identifier to which the vehicle belongs at its current location includes: The integral component of the lateral deviation controller of the sampling vehicle is used as the observation sample; The position key is determined based on parameters collected by at least one sensor on the vehicle. Extract the road segment identifier from the location key to determine the road segment identifier to which the vehicle belongs at its current location.

3. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 1, characterized in that, The step of determining the feedforward correction angle based on the average of the feedforward correction angles in the current position entry and the average of the feedforward correction angles in the preceding N forward position entries includes: The average value of the feedforward correction angle in the current position entry and the average value of the feedforward correction angle in the preceding N forward position entries are aggregated and smoothed to obtain the initial feedforward correction angle. The initial feedforward correction angle is corrected to obtain the feedforward correction angle.

4. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 3, characterized in that, The step of correcting the initial feedforward correction angle to obtain the feedforward correction angle includes: Based on formula and formula The initial feedforward correction angle is corrected to obtain the feedforward correction angle; in, Indicates the feedforward correction angle. This represents the initial feedforward correction angle. This represents the physical feedforward angle of the monorail vehicle model. This represents the absolute value of the feedforward correction angle. This represents the maximum allowable feedforward correction angle threshold for the vehicle. This represents the rate of change of the vehicle's feedforward correction angle. This represents the threshold for the maximum allowable rate of change of the vehicle's feedforward correction angle. and Indicates the calibration parameters.

5. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 1, characterized in that, The vehicle centering deviation includes: lateral distance deviation feedback and azimuth angle deviation feedback. Controlling the vehicle's movement on the road based on the feedforward correction angle, the variance and confidence level of the feedforward correction angle in the current position entry, the variance and confidence level of the feedforward correction angle in the N preceding position entries, and the real-time collected vehicle centering deviation includes: The path spectrum gating coefficient is determined based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, and the variance and confidence of the feedforward correction angle in the N preceding path position entries. Angle control commands are generated based on the feedforward correction angle, the road spectrum gating coefficient, the lateral distance deviation feedback, and the azimuth deviation feedback. The turning angle control command is sent to the vehicle steering device to control the vehicle's movement on the road.

6. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 5, characterized in that, The step of generating a steering control command based on the feedforward correction angle, the road spectrum gating coefficient, the lateral distance deviation feedback, and the azimuth deviation feedback includes: Based on formula , Generate the corner control command; in, This represents the physical feedforward angle of the monorail vehicle model. Represents the spectrum gating coefficient. This indicates the feedforward correction angle of the vehicle at its current position. This indicates the amount of lateral distance deviation feedback for the vehicle. This indicates the vehicle's azimuth deviation feedback amount. Used to indicate the road curvature of the current road segment where the vehicle is located. Used to indicate the rate of change of road curvature in the current road segment where the vehicle is located.

7. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 2, characterized in that, Updating the current location entry based on the observation parameters includes: Based on the observed parameters, determine whether the vehicle meets the steady-state gating conditions; If so, then determine the current quality weight; Based on the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, the mean value of the feedforward correction angle in the current position entry is updated to obtain a new mean value of the feedforward correction angle. The variance of the feedforward correction angle is updated based on the mean of the new feedforward correction angle, the mass weight, the preset weighted recursive coefficient, and the integral component of the lateral deviation controller, to obtain the variance of the new feedforward correction angle. The confidence level is updated based on the location confidence level, the variance of the new feedforward correction angle, and the preset confidence level adjustment step size to obtain a new confidence level.

8. The lane keeping feedforward correction method driven by cross slope spectrum according to claim 7, characterized in that, The step of determining whether the vehicle meets the steady-state gating condition based on the observed parameters includes: The observation parameters of the current location are obtained, including: road curvature, vehicle lateral distance deviation, and vehicle azimuth angle deviation. The rate of change of the lateral distance deviation integral is determined based on the integral of the lateral distance deviation of the vehicle; Based on the parameters collected by at least one sensor on the vehicle, the location reliability is determined; Based on the observed parameters, the rate of change of the integral of the lateral distance deviation, the positional confidence, the saturation state of the vehicle actuator, and the intervention state of the stability controller, it is determined whether the vehicle meets the steady-state gating conditions.

9. A lane keeping feedforward correction system driven by cross slope spectrum, characterized in that, The lane keeping feedforward correction system includes: When a vehicle is traveling on a road, the determination module is used to determine the road segment identifier to which the vehicle belongs at its current location; The acquisition module is used to acquire the current location entry and the N preceding location entries under the same road segment identifier and direction identifier based on the location key. Each location entry includes: the mean of the feedforward correction angle, the variance of the feedforward correction angle, and the confidence level. The location key is composed of the road segment identifier, direction identifier, lane identifier, and along the route index. The feedforward correction module is used to determine the feedforward correction angle based on the average of the feedforward correction angles in the current position entry and the average of the feedforward correction angles in the preceding N along-the-path position entries. The steering control module is used to control the vehicle's movement on the road based on the feedforward correction angle, the variance and confidence of the feedforward correction angle in the current position entry, the variance and confidence of the feedforward correction angle in the N forward position entries, and the real-time collected vehicle centering deviation. The writing module is used to obtain the observation parameters of the vehicle at the current position and update the current position entry according to the observation parameters.

10. A vehicle, characterized in that, The vehicle is equipped with the lane keeping feedforward correction system driven by the cross slope spectrum as described in claim 9.

Images