Vehicle driving mode switching method, device, equipment and storage medium

Abstract

The application discloses a driving mode switching method and device of a vehicle, equipment and a storage medium, relates to the technical field of vehicle driving mode control, and discloses a driving mode switching method and device of a vehicle, equipment and a storage medium, which comprises the following steps: acquiring a perception signal and a vehicle body signal of the vehicle; determining a driving state of a driver according to the vehicle body signal, and determining a driving condition of the vehicle according to the perception signal; determining a target driving mode of the vehicle according to the driving state and the driving condition based on a driving mode recognition model; and switching the driving mode of the vehicle to the target driving mode, so that the driving state and the driving condition of the vehicle are determined from two dimensions of the perception signal and the vehicle body signal of the vehicle, the switching of the driving mode of the vehicle is judged from two angles, and the driving mode of the vehicle can be accurately switched.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle driving mode control, in particular to a driving mode switching method, device and equipment of a vehicle and a storage medium. BACKGROUND

[0002] The state is vigorously promoting the vigorous development of the intelligent automobile industry, at the same time, the public acceptance of intelligent driving system is also gradually rising, and the corresponding, the intelligent driving system also puts forward higher requirements. At present, the intelligent driving system is developing towards stronger adaptive ability, in order to adapt to different working conditions and different drivers. The driving state of the driver and the recognition accuracy of the working condition will strongly affect the performance of the system. For example, emergency braking, collision warning, lane keeping, automatic lane changing, adaptive cruise control and other functions, which need to consider different driving conditions and driver states to adjust the performance of the function. In the current technology, the driving mode of the vehicle is usually switched according to the macro driving mode or the visual detection of the driver alone to determine the driving state of the user, which may cause misjudgment and other situations, therefore, how to accurately switch the driving mode of the vehicle is the problem to be solved at present.

[0003] The above content is only used to assist in understanding the technical solutions of the present application, and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0004] The main purpose of the present application is to provide a driving mode switching method, device, equipment and storage medium of a vehicle, which aims to solve the technical problem of low accuracy of the driving mode switching of the vehicle in the prior art by visual detection of the driver to determine the driving state of the user.

[0005] To achieve the above purpose, the present application provides a driving mode switching method of a vehicle, which comprises:

[0006] Obtaining the perception signal and the body signal of the vehicle;

[0007] Determining the driving state of the driver according to the body signal, and determining the driving condition of the vehicle according to the perception signal;

[0008] According to the driving state and the driving condition, the target driving mode of the vehicle is determined based on a driving mode recognition model;

[0009] Switching the driving mode of the vehicle to the target driving mode.

[0010] In an embodiment, the step of obtaining the perception signal and the body signal of the vehicle comprises:

[0011] The perception CAN signal of the vehicle is obtained through the perception CAN of the vehicle.

[0012] The vehicle's body CAN signal is obtained through the vehicle's body CAN.

[0013] The sensing CAN signal and the vehicle body CAN signal are preprocessed to obtain the sensing signal and the vehicle body signal.

[0014] In one embodiment, the step of preprocessing the sensing CAN signal and the vehicle body CAN signal to obtain the sensing signal and the vehicle body signal includes:

[0015] The sensing CAN signal is filtered through a digital filter to remove high-frequency noise and obtain an initial sensing filtered signal.

[0016] The initial sensing filter signal is subjected to moving average filtering, and outlier processing is performed on the initial sensing filter signal based on the moving average filtering to obtain the sensing signal.

[0017] The vehicle CAN signal is synchronized in time to obtain a time-synchronized vehicle signal;

[0018] The time-synchronized vehicle body signal is verified. When the time-synchronized vehicle body signal data verification is successful, the time-synchronized vehicle body signal is de-jittered to obtain the vehicle body signal.

[0019] In one embodiment, the step of determining the target driving mode of the vehicle based on the driving mode recognition model and the driving state and driving conditions further includes:

[0020] The driving state at the observation time is obtained based on the sensing signal and the vehicle body signal at the observation time.

[0021] The driving state is marked to obtain the marked driving state;

[0022] The labeled driving states are sorted by time to obtain a time series of the driving states;

[0023] The initial driving pattern recognition model is trained based on the time series of the driving state to obtain the driving pattern recognition model.

[0024] In one embodiment, the step of training an initial driving pattern recognition model based on the time series of the driving state to obtain the driving pattern recognition model includes:

[0025] Determine the observation data and driving state in the time series of the driving state, and set the driving state as the hidden state of the initial driving mode recognition model. The observation data includes the sensing signal and the vehicle body signal.

[0026] The observed data is set as the observable state of the initial driving mode recognition model, and a mapping relationship is determined based on the observed data and the driving state;

[0027] Determine the probability distribution of the hidden state;

[0028] The transition probability matrix is ​​obtained based on the mapping relationship and the probability distribution.

[0029] The transition probability matrix is ​​analyzed based on a Gaussian mixture model to obtain the confidence level and density function. The initial driving mode recognition model is then obtained based on the confidence level and the density function.

[0030] The initial driving pattern recognition model is trained based on the observed data and the driving state to obtain the driving pattern recognition model.

[0031] In one embodiment, the step of training the initial driving pattern recognition model based on the observed data and the driving state to obtain the driving pattern recognition model includes:

[0032] The observation sequence is determined based on the observation data, and the corresponding occurrence probability is determined based on the observation sequence.

[0033] Marginal probabilities are determined based on a forward-looking algorithm;

[0034] The joint probability is calculated based on the occurrence probability and the marginal probability;

[0035] The initial driving mode recognition model is trained based on the joint probability to obtain the target final state transition probability matrix and the observation state transition probability matrix.

[0036] The driving mode recognition model is based on the target final state transition probability matrix and the observation state transition probability matrix.

[0037] In one embodiment, the step of determining the target driving mode of the vehicle based on the driving mode recognition model and the driving condition includes:

[0038] The driving state and driving conditions are input into the driving mode recognition model. Based on the target final state transition probability matrix and the observation state transition probability matrix in the driving mode recognition model, the driving state and driving conditions are inferred, and the target driving mode is output.

[0039] Furthermore, to achieve the above objectives, this application also proposes a vehicle driving mode switching device, the vehicle driving mode switching device comprising:

[0040] The signal acquisition module is used to acquire the vehicle's sensing signals and body signals;

[0041] The signal processing module is used to determine the driver's driving status based on the vehicle body signals and to determine the vehicle's driving conditions based on the sensing signals.

[0042] The pattern recognition module is used to determine the target driving mode of the vehicle based on the driving state and the driving conditions, according to the driving pattern recognition model.

[0043] The mode switching module is used to switch the vehicle's driving mode to the target driving mode.

[0044] In addition, to achieve the above objectives, this application also proposes a vehicle driving mode switching device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the vehicle driving mode switching method described above.

[0045] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the vehicle driving mode switching method described above.

[0046] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the vehicle driving mode switching method described above.

[0047] One or more technical solutions proposed in this application have at least the following technical effects: acquiring vehicle perception signals and vehicle body signals; determining the driver's driving state based on the vehicle body signals and determining the vehicle's driving conditions based on the perception signals; determining the vehicle's target driving mode based on the driving mode recognition model and the driving state and driving conditions; switching the vehicle's driving mode to the target driving mode, thereby realizing the determination of the vehicle's driving state and driving conditions from two dimensions: vehicle perception signals and vehicle body signals, and judging whether the vehicle's driving mode should be switched from two perspectives, thus enabling accurate switching of the vehicle's driving mode. Attached Figure Description

[0048] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating the method for switching driving modes for the vehicle described in this application;

[0051] Figure 2 A flowchart illustrating the second embodiment of the driving mode switching method for the vehicle in this application;

[0052] Figure 3 A schematic diagram of a hidden Markov model provided for an embodiment of the vehicle driving mode switching method of this application;

[0053] Figure 4 A schematic diagram illustrating the forward algorithm derivation provided for an embodiment of the vehicle driving mode switching method of this application;

[0054] Figure 5 A simplified flowchart illustrating the vehicle driving mode switching method provided in Embodiment 2 of this application;

[0055] Figure 6 This is a schematic diagram of the module structure of the vehicle driving mode switching device according to an embodiment of this application;

[0056] Figure 7 This is a schematic diagram of the hardware operating environment involved in the vehicle driving mode switching method in this application embodiment.

[0057] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0058] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0059] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0060] The main solution of this application embodiment is: to acquire the vehicle's perception signals and body signals; to determine the driver's driving state based on the body signals and to determine the vehicle's driving condition based on the perception signals; to determine the vehicle's target driving mode based on the driving state and the driving condition using a driving mode recognition model; and to switch the vehicle's driving mode to the target driving mode.

[0061] In this embodiment, for ease of description, the following description will focus on the vehicle driving mode switching device as the execution subject.

[0062] Because existing technologies for vehicle driving mode recognition typically rely on macroscopic driving patterns or visual detection of the driver alone to determine the user's driving state and switch vehicle driving modes, misjudgments can occur. This application provides a solution that determines the vehicle's driving state and operating conditions from two dimensions: vehicle perception signals and vehicle body signals. This two-pronged approach allows for accurate judgment of whether the vehicle driving mode needs to be switched.

[0063] As can be seen from the above embodiments, this application acquires the vehicle's perception signals and body signals; determines the driver's driving state based on the body signals, and determines the vehicle's driving condition based on the perception signals; based on the driving mode recognition model, determines the vehicle's target driving mode based on the driving state and driving condition; and switches the vehicle's driving mode to the target driving mode. This achieves the determination of the vehicle's driving state and driving condition from two dimensions: the vehicle's perception signals and body signals. It also judges whether the vehicle's driving mode needs to be switched from two perspectives, enabling accurate switching of the vehicle's driving mode.

[0064] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a vehicle driving mode switching device. The following description uses a vehicle driving mode switching device as an example to illustrate this embodiment and the subsequent embodiments.

[0065] Based on this, embodiments of this application provide a method for switching vehicle driving modes, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the vehicle driving mode switching method of this application.

[0066] In this embodiment, the vehicle driving mode switching method includes steps S10 to S40:

[0067] Step S10: Acquire the vehicle's perception signals and body signals;

[0068] It should be noted that perception signals and vehicle body signals refer to signals collected from environmental sensors installed on the vehicle or from vehicle body sensors. These perception signals and vehicle body signals include, but are not limited to: signals related to following targets, target speed, target acceleration, target relative position, lane line polynomial coefficients, lane line type, lane width, lane line confidence level, turn signal, vehicle lateral speed, brake pedal opening, accelerator pedal opening, steering wheel angle, driver fatigue level, and driver proficiency level.

[0069] In one feasible implementation, step S10 may include the following steps:

[0070] The vehicle's sensing CAN signal is obtained through the vehicle's sensing CAN.

[0071] The vehicle's body CAN signal is obtained through the vehicle's body CAN.

[0072] The sensing CAN signal and the vehicle body CAN signal are preprocessed to obtain the sensing signal and the vehicle body signal.

[0073] In practical implementation, the data from environmental sensors and body sensors need to be transmitted through the CAN bus after acquisition. Therefore, the signals acquired by the sensors are CAN signals, namely the sensing CAN signal and the body CAN signal. After the vehicle's CAN bus processes the sensing CAN signal and the body CAN signal, it is possible to obtain the sensing CAN signal from the sensing CAN and the body CAN signal from the body CAN. Furthermore, preprocessing operations are performed on the sensing CAN signal and the body CAN signal to obtain the sensing signal and the body signal.

[0074] For example, the step of preprocessing the sensing CAN signal and the vehicle body CAN signal to obtain the sensing signal and the vehicle body signal includes:

[0075] The sensing CAN signal is filtered through a digital filter to remove high-frequency noise and obtain an initial sensing filtered signal.

[0076] The initial sensing filter signal is subjected to moving average filtering, and outlier processing is performed on the initial sensing filter signal based on the moving average filtering to obtain the sensing signal.

[0077] The vehicle CAN signal is synchronized in time to obtain a time-synchronized vehicle signal;

[0078] The time-synchronized vehicle body signal is verified. When the time-synchronized vehicle body signal data verification is successful, the time-synchronized vehicle body signal is de-jittered to obtain the vehicle body signal.

[0079] In practical implementation, the vehicle's sensing CAN signals and body CAN signals are acquired relatively easily on the vehicle, thus ensuring the feasibility of the solution. Since the acquired signals are noisy, basic preprocessing is required. A crucial step in preprocessing and filtering the sensing and body CAN signals is that their sources differ: one is based on sensors such as speed sensors and image sensors used to collect environmental information, while the other originates from the vehicle's control system, such as the braking and steering systems. When preprocessing and filtering the sensing CAN signal, denoising is first performed. During denoising, a digital filter is used to remove high-frequency noise from the signal, resulting in an initial sensing filtered signal. This noise may originate from electromagnetic interference, internal equipment noise, etc. The denoised signal is then smoothed using a moving average filter to reduce random fluctuations. The smoothed signal is then standardized or normalized, converting it to a fixed range, such as 0 to 1 or -1 to 1, reducing the signal strength for easier data analysis and faster signal processing. After processing the signal data, outliers are further processed to complete the filtering and obtain the sensing signal.

[0080] When filtering the vehicle body CAN signal, the vehicle body CAN signal can first be time-synchronized to ensure that all relevant vehicle body CAN signals are synchronized in time, resulting in a time-synchronized vehicle body signal. This facilitates subsequent data analysis and processing. After this, the data integrity of the vehicle body CAN signal can be checked to verify its integrity and correctness, including checking the length of the signal data, checksum, etc. After verifying the data of the time-synchronized vehicle body signal, if the verification is successful, the time-synchronized vehicle body signal is de-jittered to remove jitter signals caused by electromagnetic interference or mechanical vibration, thus obtaining the vehicle body signal.

[0081] Step S20: Determine the driver's driving status based on the vehicle body signal, and determine the vehicle's driving condition based on the sensing signal;

[0082] It should be noted that driving states mainly include acceleration, deceleration, constant speed, left turn and right turn, and driving conditions include straight driving, lane changing, turning, U-turn, reversing and other conditions.

[0083] Specifically, in determining the driver's state based on vehicle body signals, the driver's driving status can be judged based on the vehicle body signals. The driving status can be mainly divided into:

[0084] 1) Acceleration: Determined by the vehicle's acceleration / acceleration / throttle pedal opening;

[0085] 2) Deceleration: Determined by the vehicle's acceleration / speed / brake pedal opening;

[0086] 3) Constant speed: determined by the vehicle's acceleration / speed / brake pedal opening / accelerator pedal opening;

[0087] 4) Turning: Determined by the steering wheel angle / steering wheel angular velocity;

[0088] 5) Driver proficiency: Provided by the onboard DMS (Drive Monitor System) module;

[0089] 6) Driver fatigue level: provided by the onboard DMS module.

[0090] Driver status primarily characterizes the driver's actions, with signals derived from the vehicle's CAN bus, providing a basic description of the driver's state. This can be coupled with operating conditions for detailed classification, or used as pre-processing for driving modes.

[0091] Step S30: Based on the driving mode recognition model, determine the target driving mode of the vehicle according to the driving state and the driving conditions;

[0092] It should be noted that the driving pattern recognition model is trained based on the vehicle's perception signals and body signals. It is used to determine the current vehicle's form and driving conditions based on the vehicle's perception signals and body signals, and further determine the target driving state.

[0093] For example, the step of determining the target driving mode of the vehicle based on the driving mode recognition model and the driving state and driving conditions includes:

[0094] The driving state and driving conditions are input into the driving mode recognition model. Based on the target final state transition probability matrix and the observation state transition probability matrix in the driving mode recognition model, the driving state and driving conditions are inferred, and the target driving mode is output.

[0095] In a specific implementation, the driving pattern recognition model can acquire data from M sensors on the vehicle and obtain the probability of the vehicle being in the current state by passing through the observation state transition probability matrix of the driving pattern recognition model. This probability can be obtained by a Gaussian model and takes a value between [-1, 1], which indicates the confidence level. This completes the estimation of the driver's state. After passing through the state transition matrix, the transition probability of another state in the current state can be obtained, completing the judgment of the transition process and outputting the target driving pattern.

[0096] Step S40: Switch the vehicle's driving mode to the target driving mode.

[0097] In the specific implementation, when the target driving mode is obtained from the output of the driving mode recognition model, the current vehicle driving mode can be switched to the target driving mode. Before switching the driving mode, it is necessary to compare the two driving modes. If the current driving mode is the same as the target driving mode, the current vehicle driving mode is maintained.

[0098] In this embodiment, the vehicle's sensing signals and body signals are acquired; the driver's driving state is determined based on the body signals, and the vehicle's driving condition is determined based on the sensing signals; based on a driving mode recognition model, the vehicle's target driving mode is determined based on the driving state and driving condition; the vehicle's driving mode is switched to the target driving mode, thereby determining the vehicle's driving state and driving condition from two dimensions: the vehicle's sensing signals and body signals. This allows for accurate switching of the vehicle's driving mode by judging whether to switch driving modes from two perspectives.

[0099] This embodiment provides a method for switching vehicle driving modes. It acquires vehicle sensing signals and body signals; determines the driver's driving state based on the body signals and the vehicle's driving condition based on the sensing signals; determines the vehicle's target driving mode based on a driving mode recognition model and the driving state and driving condition; and switches the vehicle's driving mode to the target driving mode. This method determines the vehicle's driving state and driving condition from two dimensions: sensing signals and body signals, and judges whether to switch the vehicle driving mode from two perspectives, enabling accurate switching of the vehicle's driving mode.

[0100] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 Before step S30, the vehicle driving mode switching method further includes steps S31 to S34:

[0101] S31: The driving state at the observation time is obtained based on the sensing signal and the vehicle signal at the observation time;

[0102] S32: Mark the driving state to obtain the marked driving state;

[0103] S33: Sort the labeled driving states by time to obtain the time sequence of the driving states;

[0104] S34: Train the initial driving pattern recognition model based on the time series of the driving state to obtain the driving pattern recognition model.

[0105] In practical implementation, the driver's state and driving conditions are evaluated together. The driver's driving state is generally categorized into acceleration, deceleration, constant speed, left turn, and right turn based on lateral and longitudinal directions. Driving conditions are generally categorized into straight driving, lane changing, turning, U-turn, and reversing. Because driving state and driving conditions are strongly correlated, they are combined. For subsequent model training, driving modes need to be determined, and flags need to be added to the data to form training data along with the input data. Data annotation is generally done by experienced personnel or by batch annotation of code based on simple rules. The following are the detailed rules for driving mode classification used in this solution, including but not limited to the following states:

[0106] 1) Follow the car going straight

[0107] Following the car in front is the most common driving mode, mainly characterized by minimal lateral movement. The corresponding driving mode is indicated by the following criteria:

[0108] a. Following a vehicle has a target.

[0109] b. There is a threshold for following distance; as vehicle speed increases, the following distance threshold also increases.

[0110] c. When following another vehicle, the absolute value of the steering wheel angle should be less than a threshold value; as the vehicle speed increases, the threshold value decreases.

[0111] d. When the vehicle is traveling within the solid line area, the confidence level for following the vehicle increases.

[0112] e. When there are vehicles on both sides of the vehicle, the confidence level of following the vehicle increases; the more vehicles around the vehicle, the greater the confidence level of following the vehicle.

[0113] 2) Go straight and change lanes to the left

[0114] The labeling of this driving mode mainly follows these rules:

[0115] a. If the vehicle has a target it is following, and the relative distance decreases, the confidence level for lane changing increases, and there is no target within the left-hand range.

[0116] b. When the steering wheel is turned to the left by the vehicle and the turn exceeds a threshold, this threshold decreases as the vehicle speed increases.

[0117] c. The vehicle turns on its left turn signal.

[0118] d. The left side is either a dashed line or a feasible region.

[0119] e. If the longitudinal direction of the vehicle being followed remains unchanged, and lane lines are present, the lane lines are fitted to a straight line, and the determination is based on the curvature of the lane lines.

[0120] 3) Go straight and change lanes to the right

[0121] This scenario is similar to the switching conditions on the left, but the corresponding signal direction is reversed.

[0122] 4) Do not change lanes when turning left

[0123] Turning differs from straight-line driving, making it more challenging to predict intent and trajectory. Different functions require specific constraints for turning to increase the stability of curve functions and reduce false triggering. Therefore, driving mode recognition is necessary for curves. Most common driver assistance functions are developed based on scene recognition, and the support of driving intent can greatly improve the accuracy of the functions. The labeling of left turns mainly follows these principles:

[0124] a. The steering wheel is turned to the left, with the angle exceeding the minimum threshold but less than the maximum threshold. The threshold decreases as the vehicle speed increases.

[0125] b. If lane lines are present, the lane lines are fitted as a smooth curve pointing to the left.

[0126] c. If there is a target, the target trajectory is fitted as a smooth curve pointing to the left.

[0127] d. If the left lane line is a solid line pointing to the left, or is a non-feasible region, or there is a target on the left, the confidence level is increased.

[0128] e. During the transition from straight ahead to a curve, the vehicle speed decreases to some extent.

[0129] 5) Do not change lanes when turning right

[0130] The judgment in this scenario is the same as for a left turn without changing lanes, but the positive or negative value of the relevant signals is changed. It should be noted that the difference between turning and changing lanes mainly comes from the judgment of the relative information of surrounding targets and lane lines. The lateral threshold for changing lanes on a curve is greater than that for not changing lanes on a curve.

[0131] 6) Changing lanes when turning left

[0132] a. Turn the steering wheel to the left, exceeding the threshold for not changing lanes in a curve; this threshold decreases as vehicle speed increases.

[0133] b. If lane markings are present, the lane markings should be fitted as a leftward curve, with the left side being a dashed line.

[0134] c. If there is a target, the trajectory of the vehicle following ahead is fitted as a curve pointing to the left, and there is no target on the left.

[0135] d. Left turn signal is on

[0136] e. The relative distance to the target ahead tends to decrease; during lane changing, the distance to the target in the original lane shortens, and the relative speed increases.

[0137] 7) Changing lanes when turning right

[0138] This driving mode is labeled the same as for left turn lane change, but the positive and negative values ​​of the relevant signals change. The conditions for changing lanes on a curve are more stringent than those for not changing lanes on a curve, and there are also higher requirements for the target and lane markings.

[0139] 8) Turn left at the intersection

[0140] a. The vehicle speed is less than the threshold; and the average speed of all identified valid targets must be less than the threshold.

[0141] b. Turn on the left turn signal.

[0142] c. Lane lines disappear in the immediate vicinity ahead during the steering process.

[0143] d. If there is a target, the trajectory of the target ahead is fitted as a curve pointing to the left, and the curvature is greater than a set threshold.

[0144] e. Before turning, increase the confidence level if there are stationary targets in the surrounding direction.

[0145] f. The steering wheel is turned to the left, and the angle exceeds the threshold, which increases with vehicle speed.

[0146] 9) Turn right at the intersection

[0147] The markings for this driving mode are the same as for left turn at an intersection, with the corresponding signals changing from positive to negative. Turning at an intersection is primarily about vehicle speed and steering wheel angle, and during the turn, lane markings may partially disappear. The lateral conditions for turning at an intersection are more demanding than those for turning on a curve.

[0148] 10) Free Driving. Other driving scenarios besides those defined above are defined as free driving. The free driving state can transition to any of the previously defined states. Here, free driving does not mean the driver can drive at will, but rather that driving modes outside the defined categories are defined as free states, such as reversing, parking, and other states. Driving modes can be further subdivided for different downstream intelligent driving functions. The above is only a basic classification of driving states.

[0149] During data preparation, the model's input and output need to be extracted. The model's input consists of M signals of interest acquired from the perception system and the vehicle's CAN bus. The output represents the driving mode states, specifically N states within the driving mode. The M observation signals can be directly acquired from the vehicle, while the N states are labeled after being judged according to the aforementioned rules. The data at each time step consists of these M+N data points. Spreading the state values ​​over time forms a time series, thus completing the labeling of the training data.

[0150] For example, the step of training the initial driving pattern recognition model based on the time series of the driving state to obtain the driving pattern recognition model includes:

[0151] Determine the observation data and driving state in the time series of the driving state, and set the driving state as the hidden state of the initial driving mode recognition model. The observation data includes the sensing signal and the vehicle body signal.

[0152] The observed data is set as the observable state of the initial driving mode recognition model, and a mapping relationship is determined based on the observed data and the driving state;

[0153] Determine the probability distribution of the hidden state;

[0154] The transition probability matrix is ​​obtained based on the mapping relationship and the probability distribution.

[0155] The transition probability matrix is ​​analyzed based on a Gaussian mixture model to obtain the confidence level and density function. The initial driving mode recognition model is then obtained based on the confidence level and the density function.

[0156] The initial driving pattern recognition model is trained based on the observed data and the driving state to obtain the driving pattern recognition model.

[0157] In its implementation, a Hidden Markov Model (HMM) mainly consists of hidden states, observed states, initial values, a state transition matrix, and an observed state probability matrix. The modeling process is as follows:

[0158] Hidden state: This refers to an intention that cannot be observed. In the driving mode determination process, the driving mode itself is a hidden state that cannot be directly observed. Therefore, hidden states are set based on the driving mode. The hidden states are a discrete set of N, and are modeled as follows:

[0159] S = {Go straight and follow other vehicles, go straight and change lanes left, go straight and change lanes right, turn left and don't change lanes, turn right and don't change lanes, turn left and change lanes, turn right at intersection, drive freely}

[0160] Where S represents the set of hidden states, and si represents the hidden state value at time i.

[0161] Observable state: Information that can be acquired or estimated by the perception system and has a certain relationship with the hidden state, i.e., the system's input signal. In this model, it includes three parts: lane line-related information acquired by perception (including but not limited to lane line polynomial coefficients, lane width, lane line confidence, lane line type), target information acquired by perception (following target, target relative position, speed, acceleration, etc., target type), and vehicle CAN information (vehicle steering wheel angle, yaw rate, accelerator and brake pedal opening, vehicle speed and acceleration, etc.). The number of signals required by the perception system is M, mathematically represented as follows:

[0162] O={single1,single2,single3,…,singleM}

[0163] Where O is the set of observable states, oi is the observed state at time i, and Single is the name of the extracted input signal.

[0164] Initial state distribution: Characterizes the probability distribution of hidden states given the initial state. It indicates the different probabilities of the N hidden states at the initial time step, mathematically modeled as follows:

[0165] p0 = P(s = s) i ), 1≤i≤N

[0166] Where p0 is the probability distribution of the initial state, that is, the probability that the initial hidden state s is at position s. i The probability of a state.

[0167] State transition probability matrix: This matrix represents the probability of transitioning from the current hidden state to one of the N hidden states at the next time step. It is an N*N matrix, and its mathematical representation is as follows:

[0168] a ij =P(s) i+1 =S j |s i =S i ), 1≤i≤N, 1≤j≤N

[0169] A = [a ij ] N×N

[0170] Where a ij This indicates the current state s. i For S i At that moment, the next moment s i+1 Transfer to S j The probability of.

[0171] Observation state transition probability matrix: This matrix represents the probability of a state occurring in the hidden state N under the current observation state. It is an N*M matrix, mathematically represented as follows:

[0172] b ij =P(s) i+1 =S i |o j =O j ), 1≤i≤N, 1≤j≤M

[0173] B = [b] ij ] N×M

[0174] Where b ij This indicates the current observation status. j For O j At that moment, the next moment s i+1 Transfer to S i The probability of.

[0175] Gaussian Mixture Model. The output of a Gaussian model is [-1, 1], which can highly adapt to probabilistic models. However, the hidden states are discrete, lacking confidence levels to represent the signal's reliability. Therefore, a Gaussian Mixture Model is introduced to provide continuous output, representing the confidence level of the output signal. A Gaussian Mixture Model can be understood as K Gaussian distributions, each assigned a weight. Its mathematical model is as follows:

[0176]

[0177] Where μ is the mean, ∑ is the covariance, d is the data dimension, x is the independent variable, T denotes transpose, and i represents the i-th Gaussian distribution. The marginal probability of the sample is:

[0178]

[0179] Where φ is the density function, and the parameter θ contains the probability distribution of the dependent variable x. Solving this model requires an iterative method, specifically the EM algorithm. First, an E-step is performed to calculate the probability that each data point j comes from model k; then, an M-step is performed to calculate the unknown parameters of the iterative model, until the iteration error between the two steps is less than the minimum value. The parameters obtained at this point are the solved values. The detailed mathematical derivation is quite complex; the derivation result is presented here as follows:

[0180]

[0181]

[0182]

[0183] Therefore, the representations of each Gaussian distribution are obtained. Wherein, γ ij For probability distribution, Let represent the parameter of the j-th Gaussian distribution.

[0184] Therefore, the problem of solving the hidden states of driving modes can be abstracted into a Hidden Markov Model, which, given M perceptible signals, provides the predicted probabilities and transition state probability matrices for M formal modes. The specific model structure diagram is shown below. Figure 3 As shown.

[0185] For example, the step of training the initial driving pattern recognition model based on the observed data and the driving state to obtain the driving pattern recognition model includes:

[0186] The observation sequence is determined based on the observation data, and the corresponding occurrence probability is determined based on the observation sequence.

[0187] Marginal probabilities are determined based on a forward-looking algorithm;

[0188] The joint probability is calculated based on the occurrence probability and the marginal probability;

[0189] The initial driving mode recognition model is trained based on the joint probability to obtain the target final state transition probability matrix and the observation state transition probability matrix.

[0190] The driving mode recognition model is based on the target final state transition probability matrix and the observation state transition probability matrix.

[0191] In the specific implementation, during model training, a set of labeled data is provided, where the observed state contains M signals and the hidden state contains N signals. The hidden values ​​are obtained through logical judgments in the previous steps and data labeling. The Hidden Markov Model can be solved using a forward-backward algorithm. The basic derivation of the forward algorithm is given here, and the structural diagram is as follows. Figure 4 As shown. The derivation is as follows: Given a Hidden Markov Model λ=(A,B,p0), and an observation sequence O={O1,O2,O3,…,O2,P0 ... M Then calculate the probability of a set of observed sequences occurring:

[0192]

[0193] Based on the above, calculate the next joint probability:

[0194]

[0195] Based on the above derivation, when calculating probability problems, it is necessary to input the observation sequence and the initial probability, calculate the forward probability of the next time step, then derive the forward probability of the subsequent time steps according to the above formula, and finally sum them up to obtain the probability of the observation sequence.

[0196] The parameters are iteratively optimized using the EM algorithm. At a certain iteration, when the error is less than a set value, the final state transition probability matrix and the observed state transition probability matrix are obtained. These two matrices represent the probability of the driver being in the current perceived state and the probability of transitioning to another state from the current state, thus solving the problem of estimating the driver's state and modeling the transition process.

[0197] For example, to help understand the implementation process of the vehicle driving mode switching method obtained by combining this embodiment with the above embodiment one, please refer to... Figure 5 , Figure 5 A simplified flowchart of a vehicle driving mode switching method is provided. Specifically, the method involves preprocessing the acquired sensing signals and vehicle CAN signals, logically judging the driver's state and identifying the driving mode, and then labeling the data segments to provide a training basis for the driving mode recognition model. A driving mode recognition model based on GMM-HMM is built and then trained to obtain the GMM-HMM driving mode recognition model, and finally outputting the driving mode and confidence level.

[0198] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the driving mode switching method of the vehicle in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0199] This application also provides a vehicle driving mode switching device, please refer to... Figure 6 The vehicle driving mode switching device includes:

[0200] The signal acquisition module 10 is used to acquire the vehicle's sensing signals and body signals;

[0201] The signal processing module 20 is used to determine the driver's driving status based on the vehicle body signals and to determine the vehicle's driving conditions based on the sensing signals.

[0202] The pattern recognition module 30 is used to determine the target driving mode of the vehicle based on the driving state and the driving conditions, according to the driving pattern recognition model.

[0203] The mode switching module 40 is used to switch the vehicle's driving mode to the target driving mode. The vehicle driving mode switching device provided in this application, employing the vehicle driving mode switching method in the above embodiments, can solve the technical problem of low accuracy in the prior art of using visual detection of the driver to determine the user's driving state for vehicle driving mode switching. Compared with the prior art, the beneficial effects of the vehicle driving mode switching device provided in this application are the same as those of the vehicle driving mode switching method provided in the above embodiments, and other technical features in the vehicle driving mode switching device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0204] This application provides a vehicle driving mode switching device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the vehicle driving mode switching method in the above embodiment 1.

[0205] The following is for reference. Figure 7 The diagram illustrates a structural schematic of a vehicle driving mode switching device suitable for implementing embodiments of this application. The vehicle driving mode switching device in embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The driving mode switching device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0206] like Figure 7As shown, the vehicle's driving mode switching device may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the vehicle's driving mode switching device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the vehicle's driving mode switching device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a vehicle driving mode switching device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented or possessed alternatively.

[0207] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0208] The vehicle driving mode switching device provided in this application, employing the vehicle driving mode switching method in the above embodiments, can solve the technical problem of low accuracy in the prior art of using visual detection of the driver to determine the user's driving state for vehicle driving mode switching. Compared with the prior art, the beneficial effects of the vehicle driving mode switching device provided in this application are the same as those of the vehicle driving mode switching method provided in the above embodiments, and other technical features of this vehicle driving mode switching device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0209] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0210] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0211] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the vehicle driving mode switching method in the above embodiments.

[0212] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0213] The aforementioned computer-readable storage medium may be included in the vehicle's driving mode switching device; or it may exist independently and not installed in the vehicle's driving mode switching device.

[0214] The aforementioned computer-readable storage medium carries one or more programs that, when executed by the vehicle's driving mode switching device, cause the vehicle's driving mode switching device to:

[0215] Acquire vehicle perception signals and body signals;

[0216] The driver's driving status is determined based on the vehicle body signals, and the vehicle's driving condition is determined based on the sensing signals.

[0217] Based on the driving mode recognition model, the target driving mode of the vehicle is determined according to the driving state and the driving conditions;

[0218] Switch the vehicle's driving mode to the target driving mode.

[0219] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0220] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0221] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0222] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described vehicle driving mode switching method. This solves the technical problem in the prior art where the accuracy of visual detection of the driver to determine the user's driving state for vehicle driving mode switching is low. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the vehicle driving mode switching method provided in the above embodiments, and will not be repeated here.

[0223] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the vehicle driving mode switching method described above.

[0224] The computer program product provided in this application can solve the technical problem of low accuracy in the prior art of using visual detection of the driver to determine the user's driving state and switch vehicle driving modes. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the vehicle driving mode switching method provided in the above embodiments, and will not be repeated here.

[0225] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for switching driving modes of a vehicle, characterized in that, The method includes: Acquire vehicle perception signals and body signals; The driver's driving status is determined based on the vehicle body signals, and the vehicle's driving condition is determined based on the sensing signals. The driving mode recognition model based on the fusion of Hidden Markov Model and Gaussian Mixture Model uses the driving state as the hidden state and the driving condition and vehicle body signal as the observed state. The target driving mode of the vehicle is determined according to the state transition probability matrix and the observation probability matrix. Switch the vehicle's driving mode to the target driving mode; The driving mode recognition model based on the fusion of Hidden Markov Model and Gaussian Mixture Model, which uses the driving state as the hidden state and the driving condition and vehicle signals as the observed state, further includes the following steps before determining the target driving mode of the vehicle based on the state transition probability matrix and the observation probability matrix: The driving state at the observation time is obtained based on the sensing signal and the vehicle body signal at the observation time. The driving state is marked to obtain the marked driving state; The labeled driving states are sorted by time to obtain a time series of the driving states; Determine the observation data and driving state in the time series of the driving state, and set the driving state as the hidden state of the initial driving mode recognition model. The observation data includes the sensing signal and the vehicle body signal. The observed data is set as the observable state of the initial driving mode recognition model, and a mapping relationship is determined based on the observed data and the driving state; Determine the probability distribution of the hidden state; The transition probability matrix is ​​obtained based on the mapping relationship and the probability distribution. The transition probability matrix is ​​analyzed based on a Gaussian mixture model to obtain the confidence level and density function. The initial driving mode recognition model is then obtained based on the confidence level and the density function. The initial driving pattern recognition model is trained based on the observed data and the driving state to obtain the driving pattern recognition model.

2. The method as described in claim 1, characterized in that, The steps for acquiring the vehicle's sensing signals and body signals include: The vehicle's sensing CAN signal is obtained through the vehicle's sensing CAN. The vehicle's body CAN signal is obtained through the vehicle's body CAN. The sensing CAN signal and the vehicle body CAN signal are preprocessed to obtain the sensing signal and the vehicle body signal.

3. The method as described in claim 2, characterized in that, The step of preprocessing the sensing CAN signal and the vehicle CAN signal to obtain the sensing signal and the vehicle signal includes: The sensing CAN signal is filtered through a digital filter to remove high-frequency noise and obtain an initial sensing filtered signal. The initial sensing filter signal is subjected to moving average filtering, and outlier processing is performed on the initial sensing filter signal based on the moving average filtering to obtain the sensing signal. The vehicle CAN signal is synchronized in time to obtain a time-synchronized vehicle signal; The time-synchronized vehicle body signal is verified. When the time-synchronized vehicle body signal data verification is successful, the time-synchronized vehicle body signal is de-jittered to obtain the vehicle body signal.

4. The method as described in claim 1, characterized in that, The step of training the initial driving pattern recognition model based on the observed data and the driving state to obtain the driving pattern recognition model includes: The observation sequence is determined based on the observation data, and the corresponding occurrence probability is determined based on the observation sequence. Marginal probabilities are determined based on a forward-looking algorithm; The joint probability is calculated based on the occurrence probability and the marginal probability; The initial driving mode recognition model is trained based on the joint probability to obtain the target final state transition probability matrix and the observation state transition probability matrix. The driving mode recognition model is obtained based on the target final state transition probability matrix and the observation state transition probability matrix.

5. The method as described in claim 1, characterized in that, The step of determining the target driving mode of the vehicle based on the driving mode recognition model and the driving state and driving conditions includes: The driving state and driving conditions are input into the driving mode recognition model. Based on the target final state transition probability matrix and the observation state transition probability matrix in the driving mode recognition model, the driving state and driving conditions are inferred, and the target driving mode is output.

6. A vehicle driving mode switching device, characterized in that, The device includes: The signal acquisition module is used to acquire the vehicle's sensing signals and body signals; The signal processing module is used to determine the driver's driving status based on the vehicle body signals and to determine the vehicle's driving conditions based on the sensing signals. The pattern recognition module is used to determine the target driving mode of the vehicle based on the driving mode recognition model that is based on the fusion of Hidden Markov Model and Gaussian Mixture Model. The driving state is used as the hidden state, the driving condition and the vehicle body signal are used as the observed state, and the target driving mode of the vehicle is determined according to the state transition probability matrix and the observation probability matrix. A mode switching module is used to switch the vehicle's driving mode to the target driving mode; The driving mode recognition model based on the fusion of Hidden Markov Model and Gaussian Mixture Model, which uses the driving state as the hidden state and the driving condition and vehicle signals as the observed state, further includes the following steps before determining the target driving mode of the vehicle based on the state transition probability matrix and the observation probability matrix: The driving state at the observation time is obtained based on the sensing signal and the vehicle body signal at the observation time. The driving state is marked to obtain the marked driving state; The labeled driving states are sorted by time to obtain a time series of the driving states; Determine the observation data and driving state in the time series of the driving state, and set the driving state as the hidden state of the initial driving mode recognition model. The observation data includes the sensing signal and the vehicle body signal. The observed data is set as the observable state of the initial driving mode recognition model, and a mapping relationship is determined based on the observed data and the driving state; Determine the probability distribution of the hidden state; The transition probability matrix is ​​obtained based on the mapping relationship and the probability distribution. The transition probability matrix is ​​analyzed based on a Gaussian mixture model to obtain the confidence level and density function. The initial driving mode recognition model is then obtained based on the confidence level and the density function. The initial driving pattern recognition model is trained based on the observed data and the driving state to obtain the driving pattern recognition model.

7. A vehicle driving mode switching device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the driving mode switching method for a vehicle as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the vehicle driving mode switching method as described in any one of claims 1 to 5.

Images