A Vehicle Braking Distance Prediction Method Based on Delay Compensation and Time-Varying Dynamics

By constructing a longitudinal prediction state vector and a historical command buffer, the vehicle braking process is predicted in segments, solving the problem of braking distance prediction under real-time on-board calculation, and achieving high-precision braking distance prediction and improved longitudinal control smoothness.

CN123104583APending Publication Date: 2026-11-17BEIJING SENIOR SMART DRIVING TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611318179.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-28
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately predict vehicle braking distances under real-time onboard computing constraints, especially under different operating conditions, leading to large errors in braking trigger timing and stopping position, which affects longitudinal control smoothness.

Method used

A method based on delay compensation and time-varying dynamics is adopted to construct a longitudinal prediction state vector. The delay of the actuator is compensated by the historical command buffer. The longitudinal response mode is adaptively switched in a segmented prediction stage. The time-varying external resistance acceleration is solved, and the longitudinal state vector is updated step by step until the vehicle stops.

Benefits of technology

It achieves high-precision online prediction of braking distance under all working conditions, meets the real-time calculation requirements of the on-board controller, and improves the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123104583A_ABST
    Figure CN123104583A_ABST
Patent Text Reader

Abstract

This application provides a vehicle braking distance prediction method based on delay compensation and time-varying dynamics, belonging to the field of longitudinal motion control and state prediction technology for intelligent vehicles. It constructs a longitudinal prediction state vector and achieves delay command storage and playback compensation through a historical command buffer that matches the pure delay duration of the actuator. Adaptive split calculations are implemented for the delay prediction stage and the stable braking prediction stage using a stage gating function. The time-varying external acceleration composed of rolling resistance, slope, and air resistance is simultaneously calculated in real time as an iterative disturbance input. The longitudinal prediction state vector is updated uniformly in each prediction step until a vehicle stopping event is triggered, obtaining a staged dynamic integral distance that closely matches real driving conditions. Finally, after filtering and limiting, a stable and reliable braking distance and braking duration are output. This meets the real-time calculation requirements of the onboard controller and improves the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application belongs to the field of longitudinal motion control and state prediction technology for intelligent vehicles. Specifically, it relates to a method for predicting vehicle braking distance based on delay compensation and time-varying dynamics. Background Technology

[0002] Intelligent mobile platform transport vehicles are commonly used in scenarios such as assembly line docking, cargo loading and unloading, trailer swapping, platform berthing, parking in narrow areas, and trajectory-based parking. In these scenarios, the vehicle not only needs to "eventually stop," but also needs to maintain a small longitudinal error between the mechanical interfaces, load-bearing surfaces, and locking mechanisms on the vehicle body or trailer and the target position. The longitudinal controller typically determines when to remove the driving force and apply the braking force based on the remaining distance between the vehicle and the target point and the predicted braking distance. Therefore, the accuracy of the predicted braking distance directly affects the braking trigger timing, stopping position error, and control smoothness.

[0003] From the moment the controller issues a deceleration command to the moment the vehicle speed changes accordingly, the actual vehicle typically goes through a process of communication, message scheduling, chassis controller processing, drive force removal, brake pressure establishment, and mechanical clearance elimination. This process also has the following characteristics: (1) Pure delay characteristic: The newly generated braking command at the current moment will not immediately act on the vehicle; during the delay time, the vehicle may still continue to execute the previously sent drive, coasting, or braking commands. (2) Inertial response characteristic: Even when the command takes effect, the vehicle's internal acceleration will not reach the target value instantaneously, but will change gradually according to the time constant of the actuator. (3) Asymmetric characteristics of drive and braking: The gain and time constant of the drive force removal, initial brake establishment, and stable braking stages are usually different and cannot be represented by the same fixed deceleration or the same set of first-order parameters. (4) Time-varying characteristics of external resistance: Rolling resistance, slope component, and air resistance all affect the vehicle acceleration, but as the speed decreases and the vehicle continues to move along the road, the effectiveness of each component in the prediction time domain is not the same. (5) State switching characteristics: When the vehicle brake pedal has a significant opening and the measured deceleration has been formed, starting the prediction from the complete delay stage will repeatedly calculate the brake establishment process, resulting in an overestimation of the predicted distance.

[0004] Existing technologies struggle to make the prediction process correspond to the vehicle's actual dynamic process of "historical command not yet in effect - braking gradually established - stable braking - stopping" while meeting the constraints of real-time onboard computing, and to continuously output braking distance and braking time that can be used for parking control under different operating conditions such as slope, rolling resistance, air resistance, and braking already established. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a vehicle braking distance prediction method based on delay compensation and time-varying dynamics, which can achieve high-precision online prediction of braking distance under all working conditions, meet the real-time calculation requirements of the vehicle controller, and improve the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.

[0006] This application provides a vehicle braking distance prediction method based on delay compensation and time-varying dynamics, the method comprising the following steps: Acquire the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations; A longitudinal prediction state vector is constructed based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. A historical command buffer is constructed based on the pure delay time of the vehicle's longitudinal actuator, and the historical sequence of longitudinal control commands is stored in the historical command buffer. Based on the historical sequence of longitudinal control commands in the historical command buffer and the current longitudinal motion state of the vehicle, the braking distance prediction process is divided into a delayed prediction stage and a stable braking prediction stage. Based on the external resistance-related observations, the external acceleration, consisting of the equivalent disturbances corresponding to rolling resistance, ramp component, and air resistance, is determined step by step in the prediction process. Based on the braking state related observations, the stage gating result is determined, and according to the stage gating result, the option is to skip the delayed input replay and enter the stable braking prediction, or to replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain a valid input; Using the prediction stage, the effective input, the external acceleration, and the preset model parameter set as input, the longitudinal prediction state vector is updated step by step until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit. The braking distance and braking time are obtained based on the cumulative distance and cumulative time during the update process. The braking distance is filtered or limited, and the processed braking distance and braking time are output.

[0007] In some embodiments, the historical command buffer stores longitudinal control commands in chronological order, and the buffer length of the historical command buffer and the discrete delay duration it covers are as follows:

[0008] in, The number of vertical control commands stored in the historical command buffer. This is the pure delay calibration value for the longitudinal actuator. To control the cycle, The discrete delay duration covered by the historical command buffer, and When the number of vertical control commands stored in the historical command buffer is less than When this occurs, output an invalid flag or a safe default braking distance.

[0009] In some embodiments, the longitudinal predicted state vector at the j-th prediction step is represented as:

[0010] in, Let the vehicle speed be the value at the j-th prediction step. For the equivalent internal acceleration of the longitudinal actuator, To accumulate distance, To accumulate the prediction time, initially... , When the stage gating result indicates that delayed input replay is performed and At that time, the prediction phase is a delayed prediction phase; when the phase gating result indicates skipping delayed input replay or At that time, the prediction phase is the stable braking prediction phase.

[0011] The initial value of the equivalent internal acceleration of the longitudinal actuator is calculated according to the following formula:

[0012] in, For the filtered longitudinal acceleration, For rolling resistance acceleration, For ramp acceleration, This refers to acceleration due to air resistance.

[0013] In some embodiments, during the delayed prediction phase, non-negative historical control commands correspond to driving condition parameters, and negative historical control commands correspond to first-level braking condition parameters; during the stable braking prediction phase, a preset target deceleration is used as input and second-level braking condition parameters are employed; for any condition m, the equivalent internal acceleration of the longitudinal actuator satisfies a continuous first-order inertial model:

[0014] Using the implicit Euler method according to a fixed control period Discretization yields:

[0015]

[0016]

[0017] The first The j-predicted longitudinal actuator has equivalent internal acceleration. The effective control input is obtained after the stage gating function is determined in the j-th prediction step; m is the working condition identifier, representing three types of working conditions: driving working condition, first-level braking working condition, and second-level braking working condition; Let be the first-order inertial time constant corresponding to working condition m; This represents the first-order inertial static gain corresponding to operating condition m; For fixed control cycles; , These are the weighting coefficients for the iterative working conditions.

[0018] In some embodiments, the stage gating function is determined using a brake pedal opening threshold and a filtered acceleration threshold, and the gating identifier is:

[0019] in, For gating judgment, Iskip=1 means that the vehicle has passed the delay prediction stage and skipped the replay process, while Iskip=0 means that the delay prediction stage has not been passed and the cached instruction is executed for replay. These refer to the vehicle's real-time brake pedal opening and its threshold value. These are the vehicle's filtered acceleration and its threshold, respectively.

[0020] when =0 and At that time, the j-th pending historical command is read from the historical command buffer as... ;when =1 or At that time, skip the delayed input replay and enter the stable braking prediction, let = ,in, Deceleration to the preset target.

[0021] In some embodiments, the vehicle composite acceleration, vehicle speed, cumulative distance, and cumulative prediction time in the longitudinal prediction state vector are updated step-by-step, as shown in the following formula:

[0022]

[0023]

[0024]

[0025] in, Let J be the composite acceleration of the whole vehicle in the j-th prediction step. Let J be the external acceleration for the j-th prediction step. and For the vehicle speed in the adjacent prediction step, and The cumulative distance between adjacent prediction steps. and The cumulative prediction time of adjacent prediction steps; the recursion satisfies and and in or The recursion terminates at that time.

[0026] In some embodiments, a first-order low-pass filter is used to process the staged dynamic integral distance, as shown in the formula:

[0027] Based on the comparison between the remaining distance to the target point and the filtered output braking distance, and introducing a position control margin, a braking trigger flag is defined:

[0028] in, These are the filter coefficients. This is the original integral distance for this period. This is the output braking distance for this cycle. The remaining distance to the target point. For position control margin; =1 indicates that braking is triggered. =0 indicates that braking is not triggered.

[0029] In some embodiments, a vehicle braking distance prediction device based on delay compensation and time-varying dynamics is also provided, the device comprising: The acquisition module is used to acquire the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations. The construction module is used to construct a longitudinal prediction state vector based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. It also constructs a historical command buffer based on the pure delay duration of the vehicle's longitudinal actuator and stores the historical sequence of longitudinal control commands in the historical command buffer. The determination module is used to divide the braking distance prediction process into a delayed prediction stage and a stable braking prediction stage based on the historical sequence of longitudinal control commands in the historical command buffer and the current longitudinal motion state of the vehicle. The calculation module is used to determine, step by step, the external acceleration consisting of the equivalent disturbances corresponding to rolling resistance, ramp component and air resistance, based on the external resistance related observations. The judgment module is used to determine the stage gating result based on the braking state related observations, and select to skip the delayed input replay and enter the stable braking prediction according to the stage gating result, or replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain the effective input; The update module is used to update the longitudinal prediction state vector step by step, taking the prediction stage, the effective input, the external acceleration and the preset model parameter set as input, until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit, and to obtain the braking distance and braking time based on the cumulative distance and cumulative time during the update process. The supply module is used to filter or limit the braking distance and output the processed braking distance and braking time.

[0030] In some embodiments, an electronic device is also provided, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the steps of the vehicle braking distance prediction method based on delay compensation and time-varying dynamics described above are performed.

[0031] In some embodiments, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, performs the steps of the vehicle braking distance prediction method based on delay compensation and time-varying dynamics as described above.

[0032] This application describes a vehicle braking distance prediction method based on delay compensation and time-varying dynamics. It constructs a longitudinal prediction state vector and achieves delay command storage and playback compensation through a historical command buffer that matches the pure delay duration of the actuator. It utilizes a stage gating function to achieve adaptive split calculations for the delay prediction stage and the stable braking prediction stage, eliminating the need for full playback of historical commands and effectively reducing the iterative computing power consumption of the onboard controller. It synchronously calculates the time-varying external acceleration composed of rolling resistance, gradient, and air resistance as the iterative disturbance input, uniformly updating the longitudinal prediction state vector step-by-step until the vehicle stops. The system obtains a phased dynamic integral distance that closely matches real driving conditions. Finally, after filtering and limiting, it outputs a stable and reliable braking distance and braking duration, which can directly support functions such as vehicle stop triggering, docking control, trajectory screening, and safety control. The system integrates actuator delay, segmented dynamics, external disturbances, operating condition diversion, and stop event prediction into a complete online prediction link. It solves the problems of traditional braking distance prediction, such as uncompensated actuator lag, single dynamic model, large amount of calculation, large fluctuation of prediction results, and poor adaptability to operating conditions. It takes into account both the requirements of real-time on-board computing and the accuracy of braking distance prediction in all scenarios. Attached Figure Description

[0033] 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 the present invention 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.

[0034] Figure 1 A flowchart of a vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to an embodiment of this application is shown; Figure 2 A schematic diagram illustrating the timing relationship between the delay prediction stage and the stable braking prediction stage in an embodiment of this application is shown. Figure 3 A schematic diagram of the structure of the vehicle braking distance prediction device based on delay compensation and time-varying dynamics according to an embodiment of this application is shown; Figure 4 A schematic diagram of the structure of the electronic device described in an embodiment of this application is shown. Detailed Implementation

[0035] 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. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0036] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. 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.

[0037] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0038] In view of the technical problems raised in the background art, this application provides a vehicle braking distance prediction method based on delay compensation and time-varying dynamics, which can realize high-precision online prediction of braking distance under all working conditions, meet the real-time calculation requirements of the vehicle controller, and improve the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.

[0039] See the instruction manual appendix Figure 1 This application provides a vehicle braking distance prediction method based on delay compensation and time-varying dynamics, the method comprising the following steps: S1. Obtain the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations; S2. Construct a longitudinal prediction state vector based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. Construct a historical command buffer based on the pure delay time of the vehicle's longitudinal actuator and store the historical sequence of longitudinal control commands into the historical command buffer. S3. Based on the longitudinal control command history sequence in the historical command buffer and the current longitudinal motion state of the vehicle, the braking distance prediction process is divided into a delayed prediction stage and a stable braking prediction stage. S4. Based on the external resistance related observations, determine the external acceleration consisting of the equivalent disturbances corresponding to rolling resistance, ramp component and air resistance step by step. S5. Determine the stage gating result based on the braking state related observations, and select to skip delayed input replay and enter stable braking prediction according to the stage gating result, or replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain effective input; S6. Using the prediction stage, the effective input, the external acceleration, and the preset model parameter set as input, update the longitudinal prediction state vector step by step until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit, and obtain the braking distance and braking time based on the cumulative distance and cumulative time during the update process. S7. Filter or limit the braking distance, and output the processed braking distance and braking time.

[0040] This application abandons the traditional, rudimentary method of calculating braking distance through empirical numerical accumulation. Instead, it reconstructs braking distance prediction into a longitudinal dynamics problem that includes input pure delay, segmented mode switching, external disturbance fields, stage gating, and stopping event constraints. The vehicle's longitudinal prediction state vector is defined as follows:

[0041] in For longitudinal velocity, For the equivalent internal acceleration of the implementing agency, This represents the cumulative distance traveled from the predicted starting point. The predicted distance of the vehicle from the current moment to the moment of stopping can be represented by the distance calculation model corresponding to the stopping event:

[0042]

[0043] in, This represents the available historical vertical command sequence at the current moment. This represents the external resistance field, such as rolling resistance, slope, and air resistance. This represents the set of equivalent model parameters for the vehicle and actuators. The core of this scheme lies in how to construct... , The segmented dynamics enable online calculations to correspond to the actual braking process while also meeting the real-time calculation requirements of the onboard controller.

[0044] The chassis actuators, communication links, and brake hydraulic system establishment process have inherent pure delays. The brake commands issued in real time cannot immediately affect the vehicle's dynamic response. Therefore, a historical command buffer is built according to the equivalent delay time of the actuators. The discrete historical sequence is used to perform equivalent online approximation processing on the delayed inputs that take effect later. In the delay prediction stage, the vehicle's motion state is not uniform gliding. It will be continuously driven and evolved by the historical driving, gliding, and braking commands in the buffer.

[0045] The dynamic characteristics of the longitudinal actuator vary significantly in different working ranges. Segmented mode variables are used to distinguish the vehicle operating conditions: drive removal mode, initial braking establishment mode, stable braking mode, and other equivalent longitudinal response modes.

[0046] The longitudinal state of the vehicle in any mode satisfies a unified evolution equation:

[0047] in, For external resistance vector, These are the equivalent parameters under the current mode. They can be implemented using first-order inertia, second-order response, lookup table models, state-space models, or piecewise linear models. The onboard controller selects the appropriate equivalent response model based on the response characteristics of the vehicle's actuators and maintains consistent updates of state variables, effective inputs, external drag, and mode parameters during the prediction process.

[0048] At the online computing level, this application divides the prediction process from the current moment to the vehicle's stop into a delayed prediction stage and a stable braking prediction stage. The delayed prediction stage replays historical instructions from the buffer and calculates the vehicle state at the end of the delay. The stable braking prediction stage uses this state as the initial value for iteration and continuously recursively calculates it under the target deceleration / equivalent braking input until a stop event is triggered. Both stages share the same state space and uniformly update four types of state variables: internal acceleration, external resistance, vehicle speed, and cumulative distance.

[0049] To ensure that the predicted initial values ​​are consistent with the actual operating state of the vehicle, a stage gating function is introduced:

[0050] in, This is an observation vector composed of braking feedback, measured acceleration, chassis condition, or other observable measurements. When When the vehicle has passed the pure delay and initial braking establishment phase, it is considered to have directly entered the stable braking prediction phase in its current state; when At that time, the historical command replay is executed. This gating function can be implemented using threshold logic, a state machine, or a classification function.

[0051] After the phased dynamics iterative calculation is completed, the predicted distance is output as the parking control prediction quantity. The predicted distance is compared with the target remaining distance to complete the braking trigger and vehicle parking control.

[0052] Compared with existing technologies, this application unifies actuator delay, historical input, piecewise time-varying dynamics, external resistance field, stage gating, and stopping events into a single predictive framework. In practical implementation, the onboard controller can read state variables and historical commands according to the control cycle, call the calibrated states, modes, gating, and recursive functions, and calculate the stopping distance step by step.

[0053] Based on the above overall design concept, dynamic modeling framework, segmented prediction architecture and stage gating discrimination mechanism, this application completes the entire set of online braking distance calculations based on the execution process of steps S1-S7.

[0054] Step S1 primarily involves acquiring observation data for multiple vehicle types. Specifically, the longitudinal command sequence from the previous cycle is synchronously acquired during each discrete control cycle. Current vehicle speed Longitudinal acceleration after filtering Rolling resistance acceleration ramp acceleration air resistance acceleration Vehicle planar position Brake pedal opening And whether it has come to a standstill.

[0055] Furthermore, a state clearing protection logic is added: when the vehicle is stationary and the speed is less than the clearing threshold, the command buffer, predicted trajectory, and predicted distance of the previous cycle are cleared to prevent the state of the previous parking process from being carried over to the next start, ensuring that the initial state of the current prediction is true and valid.

[0056] Step S2 mainly involves constructing a longitudinal prediction state vector, uniformly encapsulating three core state variables: vehicle speed, equivalent internal acceleration, and cumulative distance. This provides a standardized and unified calculation carrier for subsequent full-process dynamics iterations, ensuring consistent data formats and seamless state transfer between the two prediction stages. It also establishes a command buffer that matches the delay duration of the actuators, specifically compensating for the physical latency caused by the chassis actuators and communication links, storing past control commands to enable delayed command replay calculations. Simultaneously, it calculates the initial value of the equivalent internal acceleration, providing the dynamics iteration with a valid initial state value that closely matches the actual operating conditions of the vehicle.

[0057] The constructed vehicle longitudinal prediction state vector includes, in addition to the aforementioned longitudinal velocity, the longitudinal velocity. Equivalent internal acceleration of the execution agency The cumulative driving distance from the predicted starting point In addition, it can be expanded to include braking pressure state, actuator hysteresis state, vehicle load, and road adhesion coefficient estimation values ​​to adapt to complex road surfaces and variable load conditions.

[0058] The length of the constructed historical command buffer is:

[0059] in, This is the pure delay calibration value for the longitudinal actuator. To control the cycle, This represents the actual prediction delay after discretization.

[0060] Command history is stored in the command history buffer in ascending order of age.

[0061] When the historical command buffer is not fully filled, an invalid flag or a safe default braking distance is output to avoid the risk of incorrect prediction; data overflow automatically removes the oldest command.

[0062] Simultaneously calculate the initial value of the equivalent internal acceleration of the longitudinal actuator, using the following formula:

[0063] in, For the filtered longitudinal acceleration, For rolling resistance acceleration, For ramp acceleration, This represents air resistance acceleration. It accurately characterizes the current dynamic response state of the actuator, serving as the initial state for subsequent iterative predictions.

[0064] After constructing the longitudinal prediction state vector and storing historical commands, step S3 breaks down the calculation time sequence according to the actual physical process of vehicle braking, dividing the whole vehicle braking prediction process into two core time sequence intervals: the delayed prediction stage and the stable braking prediction stage, accurately matching the actual dynamic response characteristics of the vehicle. Specifically, the early stage of vehicle braking is continuously affected by historical residual driving, coasting, and braking commands, corresponding to the delayed prediction process; after braking is fully established, the vehicle continuously follows the fixed target deceleration for stable deceleration, corresponding to the stable braking process. Simultaneously, both the delayed prediction stage and the stable braking prediction stage contain multiple longitudinal response modes, specifically including drive removal, initial braking establishment, stable braking, low adhesion protection, brake pressure ramp-up, and load adaptive mode. By combining different working condition-specific dynamic parameters, the model fitting accuracy and full-scene adaptability are significantly improved. Thus, by combining the effective input determined by the stage gating function, braking feedback, longitudinal control state, and vehicle operating state, adaptive switching of multiple longitudinal response modes is achieved, building a segmented, adaptive dynamic prediction framework, providing effective input and accurate working condition basis for subsequent state recursive calculations.

[0065] When the stage gating function does not trigger skipping the delayed prediction stage, the first... The delayed input that actually takes effect in each prediction step can be directly retrieved from the buffer:

[0066] in This indicates reading the first command from the history command buffer. A command that is about to take effect. If communication timestamp alignment, command hold, or resampling is required, preprocessing can be performed before writing to the buffer. Prediction is still performed by reading point by point according to the same formula.

[0067] set up For the phased gating results, This indicates that the vehicle is currently in an effective braking state. This indicates that the delayed input still needs to be replayed. Prediction mode. It can be determined by the following formula:

[0068] in Indicates the driver or non-negative command response mode. This indicates the initial braking response mode during the delayed prediction phase. This indicates a stable braking mode. Accordingly, the effective input for each prediction step is:

[0069] Therefore, the historical drive, coasting, or braking commands that continue to act during the pure delay period, as well as the target deceleration input during the stable braking prediction phase, are all clearly distinguished in the same prediction sequence.

[0070] To achieve continuous, non-jumping iterative updates in the delayed prediction phase and the stable braking prediction phase, a unified discrete state transition framework is established to update the complete longitudinal state of the vehicle step by step with a fixed time step.

[0071] Let the first The state of each prediction step is: In the current mode Next, take the corresponding time constant and gain. And update the internal acceleration according to the first-order response:

[0072] The external resistance is expressed as follows within this prediction step:

[0073] in, and These are the attenuation coefficients of the ramp component and the air resistance component at the prediction time, respectively.

[0074] Total acceleration, velocity, distance, and prediction time are derived recursively using the following formula:

[0075]

[0076]

[0077]

[0078] The above recursion can be written in a unified state transition form to achieve a one-step recursive update of the vehicle's longitudinal state:

[0079] in, These are parameters such as time constant, gain, and limiting in the current mode.

[0080] The stopping event is defined in the discrete domain as:

[0081] The phased dynamic integral distance can be directly taken as the cumulative distance corresponding to the stopping step:

[0082] It can also be equivalently written as the sum of the distance increments at each prediction step:

[0083] Therefore, this scheme uses "delayed input replay - mode transition - stop event - distance calculation" as the online calculation structure.

[0084] Among these, rolling resistance acceleration remains approximately constant within the short-term parking prediction; the slope resistance component decreases as the vehicle moves along the road and the effective gradient range changes; and the air resistance component decreases as the vehicle speed rapidly decreases. Under other vehicle or road conditions, updates can also be made based on road longitudinal profile, load estimation, or the speed square term. The various components in it.

[0085] Step S4 primarily quantifies the external time-varying disturbance field caused by rolling resistance, slope, and air resistance during the driving process. The vehicle braking distance is not only determined by the braking force, but road and air resistance also continuously consume the vehicle's kinetic energy. This step separately calculates the external acceleration as a fixed disturbance input term for the dynamic iteration, corrects the calculation deviation of the pure actuator dynamic model, and restores the real vehicle force environment.

[0086] Predicted time External acceleration is defined as:

[0087] in, , These are the attenuated equivalent acceleration of the ramp and the equivalent acceleration due to air resistance, respectively. The correction coefficients are obtained through... Figure it out.

[0088] Step S5 mainly involves completing the quantitative calculation of the vehicle's time-varying external resistance and establishing dual-stage multi-mode dynamic input rules. To avoid invalid delayed command playback, adapt to the vehicle's actual braking state, and reduce the redundant computing power overhead of the onboard controller, adaptive branch determination of the prediction process is achieved by constructing a stage gating function.

[0089] Specifically, the stage gating function uses the brake pedal opening threshold and the filtered acceleration threshold for determination:

[0090] in, For gating judgment, Iskip=1 means that the vehicle has passed the delay prediction stage and skipped the replay process, while Iskip=0 means that the delay prediction stage has not been passed and the cached instruction is executed for replay. These refer to the vehicle's real-time brake pedal opening and its threshold value. These are the vehicle's filtered acceleration and its threshold, respectively.

[0091] Step S6 primarily relies on the piecewise dynamics model to iteratively update the vehicle's longitudinal state frame by frame, and solves for the precise braking distance through integration using the stopping event criterion. The delayed prediction stage and the stable braking prediction stage have sequential temporal logic; the vehicle's final state calculated in the delayed prediction stage serves as the initial value for iteration in the stable braking prediction stage. The temporal flow relationship between the two prediction stages can be found in the appendix of the instruction manual. Figure 3 .

[0092] For any dynamic mode The internal acceleration adopts a first-order inertial model:

[0093] The following is obtained using implicit Euler discretization:

[0094] in, The first The j-predicted longitudinal actuator has equivalent internal acceleration. The effective control input is obtained after the stage gating function is determined in the j-th prediction step; m is the working condition identifier, representing three types of working conditions: driving working condition, first-level braking working condition, and second-level braking working condition; Let be the first-order inertial time constant corresponding to working condition m; This represents the first-order inertial static gain corresponding to operating condition m; For fixed control cycles; , These are the weighting coefficients for the iterative working conditions.

[0095] During the delayed prediction phase, the mode is selected based on the buffer command symbol:

[0096] That is, when the historical command is a driving or non-negative command. Use when the historical command is a braking command. During the stable braking prediction phase, using and order This segmented approach allows the residual driving force, initial braking establishment, and subsequent stabilizing braking that may occur during the pure delay period to be described by different parameters.

[0097] Overall vehicle acceleration:

[0098] Iterative speed calculation with non-negative constraints, and formula for integral distance using uniform acceleration:

[0099]

[0100] when The prediction terminates when the time is reached. A maximum prediction time is also set to prevent abnormal parameters from causing an infinite loop. This forms a dual termination criterion, i.e., the formula:

[0101] The current prediction step iteration ends immediately if either the vehicle speed is below the stopping threshold or the cumulative prediction time exceeds the upper limit. At the end, the staged numerical prediction distance is obtained. and predicted time .

[0102] The original integral braking distance obtained by the phased dynamic recursion has single-cycle calculation fluctuations and cannot directly meet the requirements of vehicle smooth control. Step S7 performs smoothing optimization and effectiveness limiting processing on the original prediction results, and connects the final output to the vehicle parking control function to realize the engineering application of the algorithm.

[0103] The phased dynamic integral distance is processed using a first-order low-pass filter, and the formula is as follows:

[0104] in, These are the filter coefficients. This is the original integral distance for this period. Output braking distance for this cycle; and through Determine if braking is triggered. The remaining distance to the target point. This provides a position control margin.

[0105] The parking controller can be based on Determine whether braking is triggered, whereby, The remaining distance to the target point. This represents the position control margin. The output can be used as a parking point prediction value for braking triggering, candidate trajectory selection, parking point accessibility assessment, or safety envelope calculation.

[0106] This application provides a vehicle braking distance prediction method based on delay compensation and time-varying dynamics. It collects vehicle motion, control commands, resistance, and braking feedback data to construct a longitudinal prediction state vector and a command buffer that matches actuator delay. The method divides the prediction into two main stages: delay and stable braking, and adaptively switches between multiple longitudinal response modes. It solves for time-varying external resistance acceleration and selectively skips command playback using a dual-threshold gating function. The vehicle state is updated iteratively in segments, and the braking distance is obtained by integrating with vehicle speed or maximum duration as the termination condition. The filtered output is used for vehicle parking, docking, trajectory, and safety control. This application compensates for the pure delay of the actuator, adapts to multiple operating conditions using segmented dynamics, and balances onboard real-time computing power with braking distance prediction accuracy.

[0107] Based on the same inventive concept, this application also provides a vehicle braking distance prediction device based on delay compensation and time-varying dynamics. Since the principle of the device in this application is similar to the vehicle braking distance prediction method based on delay compensation and time-varying dynamics described above, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0108] As per the instruction manual Figure 3 As shown in the figure, this application embodiment also provides a vehicle braking distance prediction device based on delay compensation and time-varying dynamics, the device comprising: The acquisition module 301 is used to acquire the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations; The construction module 302 is used to construct a longitudinal prediction state vector based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. It also constructs a historical command buffer based on the pure delay time of the vehicle's longitudinal actuator and stores the historical sequence of longitudinal control commands into the historical command buffer. The determination module 303 is used to divide the braking distance prediction process into a delayed prediction stage and a stable braking prediction stage based on the historical sequence of longitudinal control commands in the historical command buffer and the current longitudinal motion state of the vehicle. Calculation module 304 is used to determine, step by step, the external acceleration consisting of the equivalent disturbances corresponding to rolling resistance, ramp component and air resistance based on the external resistance related observations; The judgment module 305 is used to determine the stage gating result based on the braking state related observations, and select to skip the delayed input replay and enter the stable braking prediction according to the stage gating result, or replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain valid input; The update module 306 is used to update the longitudinal prediction state vector step by step with the prediction stage, the effective input, the external acceleration and the preset model parameter set as input, until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit, and to obtain the braking distance and braking time based on the cumulative distance and cumulative time during the update process. The supply module 307 is used to filter or limit the braking distance and output the processed braking distance and braking time.

[0109] The vehicle braking distance prediction device based on delay compensation and time-varying dynamics described in this application can achieve high-precision online prediction of braking distance under all working conditions, meet the real-time calculation requirements of the on-board controller, and improve the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.

[0110] Based on the same concept of the present invention, as shown in the appendix to the specification. Figure 4 As shown in the figure, an embodiment of this application provides the structure of an electronic device 400, which includes: at least one processor 401, at least one network interface 404 or other user interface 403, memory 405, and at least one communication bus 402. The communication bus 402 is used to realize the connection and communication between these components. The electronic device 400 may optionally include a user interface 403, including a display (e.g., touch screen, LCD, CRT, holographic imaging, or projector, etc.), a keyboard, or a clicking device (e.g., mouse, trackball, touchpad, or touch screen, etc.).

[0111] Memory 405 may include read-only memory and random access memory, and provides instructions and data to processor 401. A portion of memory 405 may also include non-volatile random access memory (NVRAM).

[0112] In some implementations, memory 405 stores executable modules or data structures, or subsets thereof, or extended sets thereof: The 4051 operating system contains various system programs used to implement various basic business functions and handle hardware-based tasks. Application module 4052 contains various applications, such as desktop (launcher), media player (MediaPlayer), browser (Browser), etc., to implement various application services.

[0113] In this embodiment of the application, by calling the program or instructions stored in the memory 405, the processor 401 is used to execute steps such as a vehicle braking distance prediction method based on delay compensation and time-varying dynamics.

[0114] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs steps such as those in a vehicle braking distance prediction method based on delay compensation and time-varying dynamics.

[0115] Specifically, the storage medium can be a general-purpose storage medium, such as a portable disk or hard disk. When the computer program on the storage medium is run, it can achieve high-precision online prediction of braking distance under all working conditions, meet the real-time calculation requirements of the vehicle controller, and improve the parking positioning accuracy and longitudinal control smoothness of autonomous vehicles.

[0116] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, and the indirect coupling or communication connection of the apparatus or units may be electrical, mechanical, or other forms.

[0117] 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.

[0118] In addition, the functional units in the embodiments provided in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0119] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. 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.

[0120] Finally, it should be noted that the above embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A vehicle braking distance prediction method based on delay compensation and time-varying dynamics, characterized in that, The method includes the following steps: Acquire the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations; A longitudinal prediction state vector is constructed based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. A historical command buffer is constructed based on the pure delay time of the vehicle's longitudinal actuator, and the historical sequence of longitudinal control commands is stored in the historical command buffer. Based on the historical sequence of longitudinal control commands in the historical command buffer and the current longitudinal motion state of the vehicle, the braking distance prediction process is divided into a delayed prediction stage and a stable braking prediction stage. Based on the external resistance-related observations, the external acceleration, consisting of the equivalent disturbances corresponding to rolling resistance, ramp component, and air resistance, is determined step by step in the prediction process. Based on the braking state related observations, the stage gating result is determined, and according to the stage gating result, the option is to skip the delayed input replay and enter the stable braking prediction, or to replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain a valid input; Using the prediction stage, the effective input, the external acceleration, and the preset model parameter set as input, the longitudinal prediction state vector is updated step by step until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit. The braking distance and braking time are obtained based on the cumulative distance and cumulative time during the update process. The braking distance is filtered or limited, and the processed braking distance and braking time are output.

2. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 1, characterized in that, The historical command buffer stores longitudinal control commands in chronological order. The buffer length of the historical command buffer and the discrete delay duration it covers are as follows: in, The number of vertical control commands stored in the historical command buffer. This is the pure delay calibration value for the longitudinal actuator. To control the cycle, The discrete delay duration covered by the historical command buffer, and When the number of vertical control commands stored in the historical command buffer is less than When this occurs, output an invalid flag or a safe default braking distance.

3. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 1, characterized in that, The longitudinal prediction state vector at the j-th prediction step is represented as: in, Let the vehicle speed be the value at the j-th prediction step. For the equivalent internal acceleration of the longitudinal actuator, To accumulate distance, To accumulate the prediction time, initially... , When the stage gating result indicates that delayed input replay is performed and At that time, the prediction phase is a delayed prediction phase; when the phase gating result indicates skipping delayed input replay or At that time, the prediction phase is the stable braking prediction phase. The initial value of the equivalent internal acceleration of the longitudinal actuator is calculated according to the following formula: in, For the filtered longitudinal acceleration, For rolling resistance acceleration, For ramp acceleration, This refers to acceleration due to air resistance.

4. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 3, characterized in that, During the delayed prediction phase, non-negative historical control commands correspond to drive condition parameters, and negative historical control commands correspond to first-level braking condition parameters. During the stable braking prediction phase, a preset target deceleration is used as input, and second-level braking condition parameters are employed. For any condition m, the equivalent internal acceleration of the longitudinal actuator satisfies a continuous first-order inertial model. Using the implicit Euler method according to a fixed control period Discretization yields: The first The j-predicted longitudinal actuator has equivalent internal acceleration. The effective control input is obtained after the stage gating function is determined in the j-th prediction step; m is the working condition identifier, representing three types of working conditions: driving working condition, first-level braking working condition, and second-level braking working condition; Let be the first-order inertial time constant corresponding to working condition m; This represents the first-order inertial static gain corresponding to operating condition m; For fixed control cycles; , These are the weighting coefficients for the iterative working conditions.

5. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 4, characterized in that, The stage gating function is determined using a brake pedal opening threshold and a filtered acceleration threshold, and the gating identifier is: in, For gating judgment, Iskip=1 means that the vehicle has passed the delay prediction stage and skipped the replay process, while Iskip=0 means that the delay prediction stage has not been passed and the cached instruction is executed for replay. These refer to the vehicle's real-time brake pedal opening and its threshold value. These are the vehicle's filtered acceleration and its threshold, respectively. when =0 and At that time, the j-th pending historical command is read from the historical command buffer as... ;when =1 or At that time, skip the delayed input replay and enter the stable braking prediction, let = ,in, Deceleration to the preset target.

6. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 5, characterized in that, The vehicle composite acceleration, vehicle speed, cumulative distance, and cumulative prediction time in the longitudinal prediction state vector are updated step by step, as shown in the following formula: in, Let J be the composite acceleration of the whole vehicle in the j-th prediction step. Let J be the external acceleration for the j-th prediction step. and For the vehicle speed in the adjacent prediction step, and The cumulative distance between adjacent prediction steps. and The cumulative prediction time of adjacent prediction steps; the recursion satisfies and and in or The recursion terminates at that time.

7. The vehicle braking distance prediction method based on delay compensation and time-varying dynamics according to claim 6, characterized in that, The phased dynamic integral distance is processed using a first-order low-pass filter, and the formula is as follows: Based on the comparison between the remaining distance to the target point and the filtered output braking distance, and introducing a position control margin, a braking trigger flag is defined: in, These are the filter coefficients. This is the original integral distance for this period. This is the output braking distance for this cycle. The remaining distance to the target point. For position control margin; =1 indicates that braking is triggered. =0 indicates that braking is not triggered.

8. A vehicle braking distance prediction device based on delay compensation and time-varying dynamics, characterized in that, The device includes: The acquisition module is used to acquire the vehicle's current longitudinal motion state, the historical sequence of longitudinal control commands, external resistance-related observations, and braking state-related observations. The construction module is used to construct a longitudinal prediction state vector based on the current longitudinal motion state of the vehicle. The longitudinal prediction state vector includes the vehicle speed, the equivalent internal acceleration of the longitudinal actuator, and the cumulative distance. It also constructs a historical command buffer based on the pure delay duration of the vehicle's longitudinal actuator and stores the historical sequence of longitudinal control commands in the historical command buffer. The determination module is used to divide the braking distance prediction process into a delayed prediction stage and a stable braking prediction stage based on the historical sequence of longitudinal control commands in the historical command buffer and the current longitudinal motion state of the vehicle. The calculation module is used to determine, step by step, the external acceleration consisting of the equivalent disturbances corresponding to rolling resistance, ramp component and air resistance, based on the external resistance related observations. The judgment module is used to determine the stage gating result based on the braking state related observations, and select to skip the delayed input replay and enter the stable braking prediction according to the stage gating result, or replay the historical commands to be effective in the delayed time domain from the historical command buffer to obtain the effective input; The update module is used to update the longitudinal prediction state vector step by step, taking the prediction stage, the effective input, the external acceleration and the preset model parameter set as input, until the vehicle speed is lower than the stopping threshold or the prediction time reaches the set upper limit, and to obtain the braking distance and braking time based on the cumulative distance and cumulative time during the update process. The supply module is used to filter or limit the braking distance and output the processed braking distance and braking time.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of a vehicle braking distance prediction method based on delay compensation and time-varying dynamics as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of a vehicle braking distance prediction method based on delay compensation and time-varying dynamics as described in any one of claims 1 to 7.