A Simulation and Prediction Method for Autonomous Mobile Vehicle-Road Perception Module Based on Time Series Model

Abstract

This invention discloses a simulation prediction method for autonomous mobile vehicle-road perception modules based on a time-series model. The method includes: constructing an onboard perception information flow based on actual traffic scenarios; performing simulations and statistically analyzing the usage of perception modules based on the onboard perception information flow to obtain statistical data; determining ARIMA model parameters based on the statistical data to obtain the final model; and predicting the usage of perception modules based on the final model. By using this invention, the probability of perception device usage can be predicted, providing a reference for the allocation of perception modules in autonomous vehicles and solving the problem of information redundancy in autonomous driving perception modules in traffic simulation systems. This invention can be widely applied in the field of simulation prediction.

Description

Technical Field

[0001] This invention relates to the field of simulation prediction, and particularly to a simulation prediction method for an autonomous mobile vehicle-road perception module based on a time-series model. Background Technology

[0002] With the increasing prevalence of autonomous vehicles and the development of data acquisition technology, the problem of information redundancy in autonomous driving perception modules is becoming increasingly serious. Due to differences in the acquisition capabilities, speeds, and conditions of various acquisition devices in autonomous vehicles, using computer simulations to employ different acquisition devices in different traffic scenarios can effectively address the problem of information redundancy and reduce information processing time. However, this process requires predicting the usage of the perception modules before control can be implemented. Therefore, a method is urgently needed to predict the usage of perception modules under different environmental conditions, thereby reducing information redundancy. Summary of the Invention

[0003] In view of this, in order to solve the problem of redundant information collection in existing autonomous driving systems, this invention proposes a simulation prediction method for autonomous mobile vehicle-road perception modules based on a time-series model. The method includes the following steps:

[0004] Build an in-vehicle perception information process based on actual traffic scenarios;

[0005] Based on the vehicle-mounted perception information flow, simulations were performed and the usage of perception modules was statistically analyzed to obtain statistical data.

[0006] The ARIMA model parameters are determined based on statistical data to obtain the final model;

[0007] The usage of the perception module is predicted based on the final model.

[0008] The vehicle-mounted perception information process includes the processes of initiating perception, identifying acquisition devices, selecting devices, processing information, and storing information. The perception module includes a vehicle-mounted camera, lidar, ultrasonic radar, and infrared detector.

[0009] In some embodiments, the step of simulating and statistically analyzing the usage of the sensing module based on the vehicle-mounted perception information flow to obtain statistical data specifically includes:

[0010] Analyze the vehicle-mounted perception information flow and set parameters for the simulation model;

[0011] The parameters include data collection time, weather conditions, data collection distance, and camera field of view;

[0012] Based on the vehicle-mounted perception information flow and the set parameters, the usage of the perception module is statistically analyzed using the Assembler module in the simulation model at the same agent interval, and statistical data is obtained.

[0013] Through this optimization process, based on the characteristics of the four types of sensing and acquisition devices, and by judging through four main acquisition conditions, the main sensing and acquisition devices to be used are finally determined, and then the usage is statistically analyzed.

[0014] In some embodiments, the step of determining the ARIMA model parameters based on statistical data to obtain the prediction model specifically includes:

[0015] Convert statistical data into time series data to obtain the original series;

[0016] The original sequence is subjected to d-order differencing until a stationary sequence is obtained, and the parameters d of the ARIMA model are determined.

[0017] Based on stationary sequences, the parameters p and q of the ARIMA model are determined using the partial autocorrelation coefficient and the autocorrelation coefficient.

[0018] The final model is obtained based on the parameters d, p, and q of the ARIMA model.

[0019] This optimized step, which involves superimposing autocoherent holograms and then performing backpropagation, can increase the sample data, particularly useful when there are few samples.

[0020] In some embodiments, it also includes:

[0021] The original sequence was tested for stationarity.

[0022] In some embodiments, it also includes:

[0023] The residual sequence is tested. If the residual sequence is a white noise sequence, the model is considered to have been successfully established.

[0024] In some embodiments, the prediction formula of the final model is expressed as:

[0025]

[0026] In the above formula, c is a constant. p represents the autoregressive coefficient and its corresponding lag order, θ i and q are the moving average coefficients and their corresponding lag orders, respectively, and x is the moving average coefficient. t For time series, ε t It has a mean of zero and a variance of σ. 2 The residual sequence.

[0027] Through this optimized step, the trained backpropagation hologram of the test image is processed to output a focused image. First, feature information is extracted from the input image through multiple convolutional layers. Then, the feature map of each convolutional layer is copied and merged into the decoded features of the corresponding deconvolutional layer.

[0028] In some embodiments, the ADF test is used to detect the stationarity of the original sequence, and the formula is expressed as follows:

[0029]

[0030] Where, {ε t} are independently distributed, and E(ε) t )=0, D(ε t )=σ 2 , E(ε 2 t If the actual value of the statistic t is less than the critical value in the ADF test, then the sequence can be considered a stationary sequence.

[0031] Based on the above scheme, this invention provides a simulation prediction method for autonomous mobile vehicle-road perception modules based on a time-series model. It simulates the perception modules and statistically analyzes the usage probability of the perception devices. The statistical results are relatively complete and accurate, which can effectively construct an ARIMA prediction model and ultimately predict the usage probability of the perception devices. This can provide a reference for the allocation of perception modules in autonomous vehicles and solve the problem of redundant information collected by autonomous driving perception modules in traffic simulation systems. Attached Figure Description

[0032] Figure 1 This is a flowchart of the steps of a simulation and prediction method for an autonomous mobile vehicle-road perception module based on a time-series model, according to the present invention.

[0033] Figure 2 This is a schematic diagram of the original sequence of a specific embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the autocorrelation coefficient in a specific embodiment of the present invention;

[0035] Figure 4 This is a schematic diagram of the partial autocorrelation coefficient in a specific embodiment of the present invention;

[0036] Figure 5 This is a schematic diagram illustrating a specific embodiment of the present invention;

[0037] Figure 6 This is a structural block diagram of an autonomous mobile vehicle-road perception module simulation and prediction system based on a time-series model, according to the present invention. Detailed Implementation

[0038] In current autonomous driving applications, perception modules need to be active at all times. However, in some situations, it is not necessary for all perception modules to be active to complete the driving task. This leads to the problem of overuse of perception modules in existing technologies.

[0039] To address the technical problem mentioned in the background art of not being able to use different data collection devices for different traffic scenarios, this invention uses the ARIMA model and simulation technology to statistically analyze historical data and construct a predictive model to predict the usage of the sensing module, thereby solving the problem of redundant collected information and reducing the time for information waiting to be processed.

[0040] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] It should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0042] It should be understood that the terms "system," "apparatus," "unit," and / or "module" used in this application are a method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0043] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "a," and / or "the" are not specifically singular and may include the plural. Generally, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements. An element defined by the phrase "comprising an..." does not exclude the presence of other identical elements in the process, method, product, or apparatus that includes the element.

[0044] In the description of the embodiments of this application, "a plurality of" refers to two or more. The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0045] Furthermore, flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, the steps can be processed in reverse order or simultaneously. Additionally, other operations can be added to these processes, or one or more steps can be removed from them.

[0046] Reference Figure 1 This is a flowchart illustrating an optional example of the simulation and prediction method for an autonomous mobile vehicle-road perception module based on a time-series model proposed in this invention. This method can be applied to computer equipment, and the prediction method proposed in this embodiment may include, but is not limited to, the following steps:

[0047] Step S1: Build the vehicle-mounted perception information process based on actual traffic scenarios;

[0048] Specifically, the autonomous driving traffic scenarios under different conditions are decoupled to build a traffic scenario information flow link. The link includes processes such as initiating perception, identifying acquisition devices, selecting main devices, information processing, and information storage. These processes are built into a complete vehicle perception information flow through modules in AnyLogic.

[0049] Step S2: Based on the vehicle-mounted perception information flow, perform simulation and statistical analysis of the usage of the perception module to obtain statistical data;

[0050] The model includes four sensing modules: a vehicle-mounted camera, a lidar, an ultrasonic radar, and an infrared detector. It should be noted that this invention is not limited to the sensing modules mentioned above; other devices with similar functions for sensing and data acquisition are also within the scope of protection of this solution.

[0051] Step S3: Determine the ARIMA model parameters based on statistical data to obtain the final model;

[0052] Among them, the ARIMA model, the Integrated Moving Average Autoregressive Model, also known as the Integrated Moving Average Autoregressive Model (moving average can also be called sliding average), is one of the time series forecasting analysis methods. In ARIMA(p, d, q), AR stands for "autoregressive", p is the number of autoregressive terms; MA stands for "moving average", q is the number of moving average terms, and d is the number of differencing operations (order) to make it a stationary series.

[0053] In this specific embodiment, the desired final model can be obtained simply by determining the parameters d, p, and q.

[0054] Step S4: Predict the usage of the perception module based on the final model.

[0055] Based on the established ARIMA model, the usage status of each data acquisition device in the vehicle-mounted perception module is predicted. In this step, the prediction of perception module usage should match the perception module data from the aforementioned statistical data. See the prediction diagram below. Figure 5 .

[0056] In some embodiments, the step of simulating and statistically analyzing the usage of the sensing module based on the vehicle-mounted perception information flow to obtain statistical data specifically includes:

[0057] Step S2.1: Analyze the vehicle-mounted perception information flow and set parameters for the simulation model. The parameters include acquisition time, acquisition weather, acquisition distance, and camera field of view.

[0058] The parameters of the SelectOut module in the simulation model are set, i.e., the acquisition condition discrimination is performed. The four main acquisition conditions are acquisition time, acquisition weather, acquisition distance, and camera field of view. The acquisition device is discriminated by setting the parameters of the SelectOut module in AnyLogic.

[0059] It should be noted that the acquisition conditions correspond to the sensing modules. When different sensing modules are selected, the acquisition conditions should also be adapted accordingly, and the replaced content should also be included in the protection scope of this invention.

[0060] Step S2.2: Based on the vehicle-mounted perception information flow and the set parameters, use the Assembler module in the simulation model to statistically analyze the usage of the perception module at the same agent interval and obtain statistical data.

[0061] In some embodiments, the step of determining the ARIMA model parameters based on statistical data to obtain the prediction model specifically includes:

[0062] S3.1. Convert the statistical data into time series data to obtain the original series. Refer to the diagram of the original series. Figure 2 ;

[0063] S3.2. Perform d-order differencing on the original sequence until a stationary sequence is obtained, and determine the parameters d of the ARIMA model.

[0064] S3.3. Based on stationary sequences, determine the parameters p and q of the ARIMA model using the partial autocorrelation coefficient and the autocorrelation coefficient.

[0065] according to Figure 3 and Figure 4 The parameters p and q are determined using the PACF method and the ACF method, respectively.

[0066] S3.4. Based on the parameters d, p, and q of the ARIMA model, the final model is obtained.

[0067] Where d is the number of differencing operations (order) performed to make it a stationary sequence, p is the number of autoregressive terms, and q is the number of moving average terms.

[0068] The prediction formula of the ARIMA model is expressed as:

[0069]

[0070] Where c is a constant, p represents the autoregressive coefficient and its corresponding lag order, θ i and q are the moving average coefficients and their corresponding lag orders, respectively, and x is the moving average coefficient. t For time series, ε t It has a mean of zero and a variance of σ. 2 The residual sequence.

[0071] Based on this, this embodiment also includes a stationarity test on the original sequence. Determining a stationary sequence requires the use of the ADF test, the principle of which is as follows:

[0072]

[0073] Where, {ε t} are independently distributed, and E(ε) t )=0, D(ε t )=σ 2 , E(ε 2 t If the actual value of the statistic t is less than the critical value, then the sequence can be considered a stationary sequence.

[0074] By first performing a stationarity test, and determining that the original sequence is non-stationary, subsequent processing is then performed to determine the parameters of the ARIMA model.

[0075] In some embodiments, the residual sequence is further tested, and if the residual sequence is a white noise sequence, the model is determined to be successfully established.

[0076] Adding a judgment step to determine whether the model has been successfully established, i.e., a verification step, can improve the practical application effect of this method.

[0077] like Figure 6 As shown, an autonomous mobile vehicle-road perception module simulation and prediction system based on a time-series model includes:

[0078] The process building module is used to build vehicle perception information processes based on actual traffic scenarios;

[0079] The information statistics module simulates and statistically analyzes the usage of the sensing module based on the vehicle-mounted sensing information flow, and obtains statistical data.

[0080] The model determination module is used to determine the ARIMA model parameters based on statistical data to obtain the final model.

[0081] The prediction module predicts the usage of the perception module based on the final model.

[0082] In some embodiments, the information statistics module further includes:

[0083] The parameter setting unit is used to analyze the vehicle perception information flow and set parameters for the simulation model. The parameters include acquisition time, acquisition weather, acquisition distance, and camera field of view.

[0084] The statistics unit is used to statistically analyze the usage of the perception module under the same agent interval, based on the vehicle perception information flow and the set parameters, using the Assembler module in the simulation model;

[0085] In some embodiments, the model determination module further includes:

[0086] The transformation unit is used to convert statistical data into time series data to obtain the original sequence.

[0087] The first parameter unit is used to perform d-order differencing on the original sequence until a stationary sequence is obtained, thereby determining the parameter d of the ARIMA model.

[0088] The second parameter unit, based on stationary sequences, uses partial autocorrelation coefficients and autocorrelation coefficients to determine the parameters p and q of the ARIMA model;

[0089] The integration unit is used to obtain the final model based on the parameters d, p, and q of the ARIMA model.

[0090] In some embodiments, it also includes:

[0091] The stationarity detection module is used to test the stationarity of the original sequence.

[0092] In some embodiments, it also includes:

[0093] The model judgment module is used to test the residual sequence. If the residual sequence is a white noise sequence, the model is judged to have been successfully established.

[0094] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0095] An autonomous mobile vehicle-road perception module simulation and prediction device based on a time-series model:

[0096] At least one processor;

[0097] At least one memory for storing at least one program;

[0098] When the at least one program is executed by the at least one processor, the at least one processor implements the simulation prediction method for an autonomous mobile vehicle-road perception module based on a time-series model as described above.

[0099] The content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0100] A storage medium storing processor-executable instructions, which, when executed by a processor, are used to implement the simulation prediction method for an autonomous mobile vehicle-road perception module based on a time-series model as described above.

[0101] The content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0102] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A simulation and prediction method for an autonomous mobile vehicle-road perception module based on a time-series model, characterized in that, Includes the following steps: Build an in-vehicle perception information process based on actual traffic scenarios; Based on the vehicle-mounted perception information flow, simulations were performed and the usage of perception modules was statistically analyzed to obtain statistical data. The ARIMA model parameters are determined based on statistical data to obtain the final model; The usage of the sensing module is predicted based on the final model; The vehicle-mounted perception information process includes the processes of initiating perception, identifying acquisition devices, selecting devices, processing information, and storing information. The perception module includes a vehicle-mounted camera, lidar, ultrasonic radar, and infrared detector. The step of simulating and statistically analyzing the usage of the sensing modules based on the vehicle-mounted perception information flow to obtain statistical data specifically includes: Analyze the vehicle-mounted perception information flow and set parameters for the simulation model; The parameters include data collection time, weather conditions, data collection distance, and camera field of view; Based on the vehicle-mounted perception information flow and the set parameters, the usage of the perception module is statistically analyzed using the Assembler module in the simulation model at the same agent interval, and statistical data is obtained. The step of determining the ARIMA model parameters based on statistical data to obtain the prediction model specifically includes: Convert statistical data into time series data to obtain the original series; The original sequence is subjected to d-order differencing until a stationary sequence is obtained, and the parameters d of the ARIMA model are determined. Based on stationary sequences, the parameters p and q of the ARIMA model are determined using the partial autocorrelation coefficient and the autocorrelation coefficient. The final model is obtained based on the parameters d, p, and q of the ARIMA model. The final prediction formula of the model is expressed as: In the above formula, c is a constant. p represents the autoregressive coefficient and its corresponding lag order, θ i and q are the moving average coefficients and their corresponding lag orders, respectively, and x is the moving average coefficient. t For time series, ε t It has a mean of zero and a variance of σ. 2 The residual sequence.

2. The simulation and prediction method for an autonomous mobile vehicle-road perception module based on a time-series model according to claim 1, characterized in that, The ADF test is used to detect the stationarity of the original sequence. The formula is as follows: Where, {ε t } are independently distributed, and E(ε) t )=0, D(ε t )=σ 2 , E(ε 2 t If the actual value of the statistic t is less than the critical value in the ADF test, then the sequence can be considered a stationary sequence. t Let represent a time series, where α and y represent different unit root test process variables.

Images