Method and system for correcting control signal delay in automatic driving

By using state prediction equations and deep learning algorithms in the autonomous driving system to correct the control signal, the deviation problem caused by the control signal delay is solved, the accurate evaluation of the bicycle state and real-time correction of the control signal are realized, and the automatic Control accuracy of driving system.

CN119002368APending Publication Date: 2024-11-22ANHUI JIANGHUAI AUTOMOBILE GRP CORP LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411220750.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-02
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In an autonomous driving system, due to the delay of the control signal, there are large deviations in the actual control results and the expected results. Especially when the vehicle or robot is driving at a fast speed, the problem of control signal delay is more prominent, and the prior art cannot effectively evaluate control. The bicycle state when the signal is sent makes it impossible to accurately correct it.

Method used

By receiving and processing the original data of the external environment, generating processing results and correcting the control signal based on the state prediction equation, real-time correction is performed using error state Kalman filtering algorithm and deep learning algorithm to ensure the accuracy and timeliness of the control signal.

Benefits of technology

Effective correction of the control signal delay is achieved, and the inevitable relationship between the control signal and the state prediction is formed. It can accurately evaluate the bicycle state when the control signal is sent, which improves the control accuracy of the vehicle or robot.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119002368A_ABST
    Figure CN119002368A_ABST
Patent Text Reader

Abstract

The invention discloses a method and system for correcting control signal delay in automatic driving, and belongs to the technical field of automatic driving, and the method comprises the steps: S1, receiving and processing original data from an external environment, and generating a processing result; s2, performing decision planning based on a processing result of the original data, and generating a state prediction equation; and S3, correcting the original control signal based on the state prediction equation, generating a corrected control signal, and issuing the corrected control signal. The control signal is corrected, the state of the vehicle can be evaluated when the control signal is sent, and accurate correction is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and more specifically, to a method and system for correcting control signal delay in autonomous driving. Background Art

[0002] In an autonomous driving system, it takes a certain amount of time for the sensor to acquire data, process the data, and then send it to the control layer, which then controls the specific operating status of the vehicle. During this period, the vehicle or robot itself continues to move. Therefore, the control signal received and sent by the control layer is actually the control signal required by the system status before the current moment, rather than the control signal required by the system at the current moment. Therefore, there will be a certain deviation between the actual control result and the expected control result, especially when the vehicle or robot is traveling at a relatively fast speed, the impact of the control deviation caused by the time lag will be greater.

[0003] However, even if the signal correction is done at the perception layer, there will still be a signal delay problem when the data is actually sent to the control layer; when the perception layer and the decision layer only predict the state of the vehicle but do not correct the control signal, there is no necessary relationship between the delay problem of the control signal and the state prediction, and the control signal delay still exists; if the predicted state of the vehicle is only a quantitative result, it is impossible to evaluate the state of the vehicle when the control signal is sent, and it is impossible to make accurate corrections. Summary of the invention

[0004] An object of the present invention is to provide a method and system for correcting control signal delay in automatic driving, so as to solve the problem of large difference between control result and expected result caused by the aforementioned control signal delay problem, and the problem of lagging vehicle state obtained when the control signal is sent.

[0005] According to a first aspect of the present invention, a method for correcting control signal delay in automatic driving is provided, comprising:

[0006] Step S1, receiving and processing raw data from the external environment to generate processing results;

[0007] Step S2, making decision planning based on the processing results of the original data and generating a state prediction equation;

[0008] Step S3, correcting the original control signal based on the state prediction equation, generating a corrected control signal, and sending the corrected control signal.

[0009] Optionally, the raw data from the external environment includes, but is not limited to, position information, speed information, acceleration information, obstacle position, motion trend and environmental characteristics, and the raw data are all the state of the external environment at time T0.

[0010] Optionally, in step S1 and step S2, the processing results of the original data are divided into a first output result and a second output result; the decision layer performs decision planning based on the first output result to generate an original control signal, performs state prediction based on the second output result to generate a state prediction equation, and corrects the original control signal based on the state prediction equation.

[0011] Optionally, the timing of correcting the original control signal in step S3 includes any one of the following methods:

[0012] Pre-correction method: before sending the original control signal to the control layer, first obtain the second output result of the perception layer, and correct the original control signal according to the second output result;

[0013] Real-time correction method: When the control layer is about to send a control signal, the current latest state prediction result is evaluated again, and the control signal to be sent is corrected in real time based on the latest prediction information.

[0014] Optionally, in the step of performing state prediction based on the second output result, performing state prediction based on an error state Kalman filter algorithm specifically includes the following sub-steps:

[0015] Using the time difference between the prediction result of the previous frame and the measurement result of the current frame, a preliminary estimate of the current frame is calculated;

[0016] Calculate the error between the initial estimate of the current frame and the measurement result of the current frame;

[0017] Calculate the Kalman gain based on the error;

[0018] The Kalman gain is used to update the state estimate to obtain the state prediction value, and the state prediction equation between the subsequent state and time is calculated based on the state prediction value.

[0019] Optionally, the time difference between the prediction result of the previous frame and the measurement result of the current frame is used in combination with the system dynamic model to calculate a preliminary estimate of the current frame, wherein the system dynamic model is pre-established based on the physical characteristics and motion laws of the system and is used to describe the law of how the system state changes over time.

[0020] Optionally, in step S2, the processing results of the original data are processed using a deep learning algorithm, and the deep learning algorithm calculates the state equation between the subsequent control signal and time based on the processing result information of the input original data through a pre-trained network model.

[0021] Optionally, in step S3, when the control signal is sent, the original control signal is directly modified in real time according to the state equation between the control signal and time, and the real-time modification includes but is not limited to adjusting the amplitude, frequency, and timing parameters of the control signal.

[0022] Optionally, the deep learning algorithm includes but is not limited to convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory network (LSTM), Transformer (attention mechanism) model, which is used to capture the temporal dependency and spatial characteristics in the input raw data processing result information.

[0023] According to a second aspect of the present invention, a system for correcting control signal delay in automatic driving is provided, comprising:

[0024] The perception layer is used to receive and process raw data from the external environment and generate processing results;

[0025] The decision layer is used to make decision planning based on the processing results of the original data, generate a state prediction equation, and correct the original control signal based on the state prediction equation to generate a corrected control signal, and send the corrected control signal to the control layer;

[0026] The control layer is used to perform corresponding control operations based on the modified control signal.

[0027] The method and system for correcting control signal delay in autonomous driving disclosed herein have the following technical effects:

[0028] The decision layer of the present invention corrects the original control signal based on the state prediction equation, generates a corrected control signal, and sends the corrected control signal to the control layer to achieve effective control of the system. The correction of the control signal is achieved, and an inevitable relationship is formed between the delay problem of the control signal and the state prediction. The state of the vehicle when the control signal is sent can be evaluated to achieve accurate correction. In addition, the deep learning algorithm and the error state Kalman filter algorithm involved in the present invention have a clear, universal and reusable structure. The overall algorithm is highly cohesive, low-coupled and easy to maintain, which can achieve more accurate control of vehicles or robots.

[0029] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0031] Figure 1 A schematic flow chart of a method for correcting control signal delay in automatic driving provided by the first embodiment of the present invention;

[0032] Figure 2 A schematic diagram of the logic implementation of a method or system for correcting control signal delay in autonomous driving provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0033] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless otherwise specifically stated.

[0034] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0035] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0036] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0037] The present invention proposes a first embodiment of a method for correcting control signal delay in automatic driving. Specifically, Figure 1 As shown, the following steps are included:

[0038] Step S1, receiving and processing raw data from the external environment, generating processing results; it should be noted that the raw data from the external environment includes but is not limited to position information, speed information, acceleration information, obstacle position, motion trend and environmental characteristics, and the raw data are all the state of the external environment at time T0. The raw data comes from the external environment, including but not limited to physical sensors (such as GPS receivers, radars, laser radar (LiDAR), infrared sensors, cameras, etc.), network data (such as traffic flow information, weather forecasts), or any other source that can provide the required information, wherein the position information is provided by GPS or other positioning technologies; the speed information is obtained by measuring the change in the position of the object per unit time, including linear speed or angular speed; the obstacle position is detected and the precise position of the obstacle in the surrounding environment is determined by visual recognition, radar or laser radar and other technologies; the motion trend predicts the future movement direction or path of the object based on the current speed and acceleration; environmental characteristics include lighting conditions, temperature, humidity, air pressure, etc., which may affect the system performance or factors that need to be considered. The processing of raw data includes preprocessing the received raw data, including data cleaning (removing noise and outliers), data compression (reducing the amount of data), data conversion (unified format and unit), etc., to ensure that the data quality meets the requirements of subsequent processing, and extract key features from the preprocessed data. These key features are used as the processing results of the raw data.

[0039] Step S2, making decision planning based on the processing results of the original data and generating a state prediction equation; the state prediction equation is a mathematical model used to describe how the future state of the system changes over time, usually one or more mathematical expressions, linking the current state of the system, input, and the influence of the external environment.

[0040] Step S3, based on the state prediction equation, the original control signal is corrected to generate a corrected control signal, and the corrected control signal is issued. The generated state prediction equation is applied to the actual system, and specific control instructions or action plans are formulated according to the prediction results. At the same time, the system needs to continuously receive new original data to update and adjust the prediction results in real time.

[0041] In addition, the correction method for control signal delay in autonomous driving also includes a feedback mechanism, which is used to collect actual system state information after the correction control signal is executed, compare it with the predicted state, evaluate the effect of the control signal correction, and adjust the model parameters or optimization strategy of the state prediction equation according to the evaluation results to continuously improve the accuracy and adaptability of the control signal correction. Through the feedback mechanism, the system can continuously optimize its decision-making planning and state prediction capabilities to adapt to the complex and changing external environment.

[0042] In addition, the method for correcting control signal delay in autonomous driving also includes a monitoring and evaluation step, which is used to monitor the operating status and performance of the system in real time, evaluate the effects of state prediction and control signal correction, and adjust the parameters of decision planning or state prediction based on the evaluation results to optimize the performance of the overall control system.

[0043] In a second embodiment of the present invention, Figure 2 The figure shows a logic implementation diagram of a method or system for correcting control signal delay in autonomous driving. In step S1 and step S2, the processing results of the original data are divided into a first output result and a second output result. The first output result is directly used for decision planning to provide basic information for subsequent control behaviors. The second output is used for state prediction. The state prediction process can be designed to be directly embedded in the decision layer algorithm or implemented separately as an independent module to predict the future state of the system or environment. The decision layer performs decision planning based on the first output result to generate the original control signal, performs state prediction based on the second output result to generate a state prediction equation, and corrects the original control signal based on the state prediction equation.

[0044] In the second embodiment of the present invention, the timing of modifying the original control signal in step S3 includes any one of the following methods:

[0045] Pre-correction method: Before sending the original control signal to the control layer, the second output result of the perception layer is obtained first, and the original control signal is corrected according to the second output result to better adapt to the predicted future state;

[0046] Real-time correction method: When the control layer is about to send a control signal, it re-evaluates the current latest state prediction result, and makes real-time corrections to the control signal to be sent based on the latest prediction information to ensure the timeliness and accuracy of the control behavior.

[0047] In the second embodiment of the present invention, in the step of performing state prediction based on the second output result, the state prediction is performed based on the error state Kalman filter algorithm, specifically including the following sub-steps:

[0048] The time difference between the prediction result of the previous frame and the measurement result of the current frame (T0) is used to calculate the preliminary estimate of the current frame. The prediction process includes the integration or iterative solution of the system dynamic equation.

[0049] Calculate the error between the initial estimate of the current frame and the measurement result of the current frame;

[0050] The Kalman gain is calculated based on the error, which reflects the degree of confidence in the measurement result relative to the predicted value;

[0051] The Kalman gain is used to update the state estimate to obtain the state prediction value, and the state prediction equation between the subsequent state and time is calculated based on the state prediction value. Among them, the initial estimate is added to the measurement error (weighted by the Kalman gain) to obtain the updated state estimate, that is, the state prediction value.

[0052] In a second embodiment of the present invention, a preliminary estimate of the current frame is calculated using the time difference between the prediction result of the previous frame and the measurement result of the current frame in combination with a system dynamic model, wherein the system dynamic model is pre-established based on the physical characteristics and motion laws of the system and is used to describe the law of how the system state changes over time.

[0053] In a third embodiment of the present invention, in step S2, the processing results of the original data are processed using a deep learning algorithm. The deep learning algorithm calculates the state equation between the subsequent control signal and time based on the processing result information of the input original data through a pre-trained network model.

[0054] In the third embodiment of the present invention, in step S3, when the control signal is sent, the original control signal is directly corrected in real time according to the state equation between the control signal and time, and the real-time correction content includes but is not limited to adjusting the amplitude, frequency, and timing parameters of the control signal. The implementation method of real-time correction of the control signal usually relies on advanced control algorithms and real-time computing technology. These algorithms can quickly calculate the correction amount of the control signal based on the system state equation and real-time feedback information, and apply it to the original control signal. At the same time, real-time computing technology ensures that these calculation processes can be completed in a very short time to meet the needs of real-time control.

[0055] In a third embodiment of the present invention, the deep learning algorithm includes but is not limited to a convolutional neural network (CNN), a recurrent neural network (RNN), a long short-term memory network (LSTM), and a Transformer (attention mechanism) model, which is used to capture the temporal dependency and spatial characteristics of the input raw data processing result information.

[0056] The present invention also provides an embodiment of a system for correcting control signal delay in automatic driving, comprising:

[0057] The perception layer is used to receive and process raw data from the external environment and generate processing results;

[0058] The decision layer is used to make decision planning based on the processing results of the original data, generate a state prediction equation, and correct the original control signal based on the state prediction equation to generate a corrected control signal, and send the corrected control signal to the control layer;

[0059] The control layer is used to perform corresponding control operations based on the modified control signal.

[0060] It should be noted that the time lag correction algorithm can be applied in a variety of similar scenarios. As long as it involves a complete process that is continuous in time but discrete in actual processing, the signal correction scheme of the present invention can be used for continuity correction.

[0061] It should be noted that the information interaction, execution process and other contents between the above-mentioned hierarchical structures are based on the same concept as the method embodiment of the present application, and are a system corresponding to the method for correcting the control signal delay in the above-mentioned autonomous driving. All implementation methods in the above-mentioned method embodiment are applicable to the embodiments of the device. Its specific functions and the technical effects brought about can be found in the method embodiment part, which will not be repeated here.

[0062] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings, but the above is only a preferred embodiment of the present invention. It should be noted that the technical features involved in the above embodiments and their preferred methods can be reasonably combined and matched into a variety of equivalent schemes by those skilled in the art without departing from or changing the design ideas and technical effects of the present invention; therefore, the present invention is not limited to the scope of implementation shown in the drawings, and all changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the specification and drawings, should be within the protection scope of the present invention.

Claims

1. A method for correcting control signal delay in automatic driving, characterized in that: include: Step S1, receiving and processing raw data from the external environment to generate processing results; Step S2, making decision planning based on the processing results of the original data and generating a state prediction equation; Step S3, correcting the original control signal based on the state prediction equation, generating a corrected control signal, and sending the corrected control signal.

2. The method for correcting control signal delay in automatic driving according to claim 1, characterized in that: The raw data from the external environment includes but is not limited to position information, speed information, acceleration information, obstacle location, motion trend and environmental characteristics, and the raw data are all the status of the external environment at time T0.

3. The method for correcting control signal delay in automatic driving according to claim 2, characterized in that: In step S1 and step S2, the processing results of the original data are divided into a first output result and a second output result; the decision layer performs decision planning based on the first output result to generate an original control signal, performs state prediction based on the second output result to generate a state prediction equation, and corrects the original control signal based on the state prediction equation.

4. The method for correcting control signal delay in automatic driving according to claim 3, characterized in that: The timing of correcting the original control signal in step S3 includes any one of the following methods: Pre-correction method: before sending the original control signal to the control layer, first obtain the second output result of the perception layer, and correct the original control signal according to the second output result; Real-time correction method: When the control layer is about to send a control signal, the current latest state prediction result is evaluated again, and the control signal to be sent is corrected in real time based on the latest prediction information.

5. The method for correcting control signal delay in automatic driving according to claim 3, characterized in that: In the step of performing state prediction based on the second output result, the state prediction is performed based on the error state Kalman filter algorithm, which specifically includes the following sub-steps: Using the time difference between the prediction result of the previous frame and the measurement result of the current frame, a preliminary estimate of the current frame is calculated; Calculate the error between the initial estimate of the current frame and the measurement result of the current frame; Calculate the Kalman gain based on the error; The Kalman gain is used to update the state estimate to obtain the state prediction value, and the state prediction equation between the subsequent state and time is calculated based on the state prediction value.

6. The method for correcting control signal delay in automatic driving according to claim 5, characterized in that: The time difference between the prediction result of the previous frame and the measurement result of the current frame is used to calculate the preliminary estimate of the current frame in combination with the system dynamic model, wherein the system dynamic model is pre-established according to the physical characteristics and motion laws of the system and is used to describe the law of system state change over time.

7. The method for correcting control signal delay in automatic driving according to claim 2, characterized in that: In step S2, the processing results of the original data are processed using a deep learning algorithm. The deep learning algorithm calculates the state equation between the subsequent control signal and time based on the processing result information of the input original data through a pre-trained network model.

8. The method for correcting control signal delay in automatic driving according to claim 7, characterized in that: In step S3, when the control signal is sent, the original control signal is directly modified in real time according to the state equation between the control signal and time. The real-time modification includes but is not limited to adjusting the amplitude, frequency and timing parameters of the control signal.

9. The method for correcting control signal delay in automatic driving according to claim 7, characterized in that: The deep learning algorithms include but are not limited to convolutional neural networks (CNN), recurrent neural networks (RNN), long short-term memory networks (LSTM), and Transformer (attention mechanism) models, which are used to capture the temporal dependency and spatial features in the input raw data processing result information.

10. A control signal delay correction system in automatic driving, characterized in that: include: The perception layer is used to receive and process raw data from the external environment and generate processing results; The decision layer is used to make decision planning based on the processing results of the original data, generate a state prediction equation, and correct the original control signal based on the state prediction equation to generate a corrected control signal, and send the corrected control signal to the control layer; The control layer is used to perform corresponding control operations based on the modified control signal.

Citation Information

Cited By

  • Performance detection method of automatic driving system, electronic equipment and program product

    CN122126296A

  • Performance testing methods, electronic devices and software products for autonomous driving systems

    CN122126296B