Vehicle emergency braking method and device

Abstract

The invention discloses a vehicle emergency braking method and device, and relates to the technical field of vehicle auxiliary driving. The vehicle emergency braking method comprises the steps that driving data of a vehicle and environment information of a traffic flow where the vehicle is located currently are collected, and the environment information comprises feature data of a plurality of VRUs; inputting the feature data of the plurality of VRUs into a pre-trained model, and predicting a prediction trajectory of each VRU; calculating the chaos degree of the traffic flow where the vehicle is located according to the driving data of the vehicle and the characteristic data of each VRU; according to the track of each VRU and the chaos degree of the traffic flow where the vehicle is located, a braking strategy is dynamically generated for the vehicle, so that the vehicle is braked through the braking strategy, the braking strategy can dynamically adapt to the vehicle driving scene, and the driving comfort of the vehicle is improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle driver assistance technology, and in particular to a vehicle emergency braking method and device. Background Technology

[0002] The Autonomous Emergency Braking (AEB) system is one of the core systems for realizing autonomous driving. Currently, AEB mainly formulates braking strategies for vehicles based on safe distance and pre-collision time. However, existing AEB systems primarily employ static strategies, meaning that the same braking strategy is used regardless of the driving scenario.

[0003] Existing static braking strategies for AEB are prone to mis-braking in complex vehicle driving scenarios, such as low-speed traffic flow scenarios dominated by vulnerable road users. They fail to guarantee vehicle ride comfort and lack dynamic adaptation to complex traffic flow scenarios. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a vehicle emergency braking method and apparatus. The method can dynamically generate a braking strategy for the vehicle according to the vehicle's current scenario, so that the braking strategy can dynamically adapt to the vehicle's driving scenario, thereby improving the vehicle's driving comfort.

[0005] To achieve the above objectives, in a first aspect, according to embodiments of the present invention, a vehicle emergency braking method is provided, comprising: The vehicle's driving data and the environmental information of the traffic flow in which the vehicle is currently located are collected, wherein the environmental information includes the characteristic data of multiple vulnerable road users (hereinafter referred to as VRUs); The feature data of multiple VRUs are input into a pre-trained model to predict the predicted trajectory of each VRU. The degree of chaos in the traffic flow where the vehicle is located is calculated based on the vehicle's driving data and the characteristic data of each VRU. Based on the trajectories of each VRU and the degree of chaos in the traffic flow where the vehicle is located, a braking strategy is dynamically generated for the vehicle to brake the vehicle using the braking strategy.

[0006] Optionally, the above-mentioned vehicle emergency braking method also includes: Using the historical location information of each historically vulnerable road user (h-VRU) in the continuous frame images of the collected historical traffic flow environmental historical data, a historical location vector sequence is constructed for each h-VRU; Based on the historical position vector sequences of each h-VRU, an adversarial network algorithm with an attention mechanism is iteratively trained, and during the iterative training process, a variety of loss functions are used to constrain the training to obtain a pre-trained model.

[0007] Optionally, the above-mentioned vehicle emergency braking method further includes: The historical position vector sequences of each h-VRU are input into the encoding module of the variational autoencoder to calculate the mean and variance of each historical position vector sequence. The mean and variance of each of the historical position vector sequences are input into the decoding module of the variational autoencoder to generate the input latent variables of each of the h-VRUs. The iterative training of the adversarial network algorithm with an attention mechanism includes: training the adversarial network algorithm with an attention mechanism using the input latent variables of each of the h-VRUs.

[0008] Optionally, the adversarial network algorithm that incorporates an attention mechanism includes a multi-layer LSTM, wherein a pooling layer is added to each adjacent LSTM in the space, and the pooling layer is used to collect hidden information of the neighboring h-VRU.

[0009] Optionally, the pooling layer collects hidden information of neighboring h-VRUs based on the following first calculation formula; First calculation formula:

[0010] in, express Are the sensors collecting environmental information in use? The range of grid data collected at any time Inside, Indicates the first The neighboring h-VRUs of each h-VRU, Indicates the LSTM output of the first... One h-VRU in The hidden state of the LSTM at time step.

[0011] Optionally, LSTM is powered by a forget gate. Input gate and output gate Composed of, among which, the Gate of Oblivion Input gate and output gate The following set of calculation formulas must be satisfied: Calculation formula set:

[0012] in, This indicates a deviation from the forget gate; Indicates the deviation of the input gate; Indicates the deviation of the output gate; Indicates the weight of the forget gate; Indicates the weights of the input gates; Indicates the weight of the output gate; This represents the input value at the previous time step; This represents the input value at the current moment. This represents the activation function used for the output of hidden layer neurons.

[0013] Optionally, the diversification loss function is calculated using the following second formula:

[0014] in, Indicates hyperparameters, This represents the actual trajectory points of the h-VRU; This represents the predicted trajectory point of h-VRU.

[0015] Optionally, the calculation of the degree of chaos in the traffic flow where the vehicle is located includes: For each VRU, determine the relative distance and relative speed between the VRU and the vehicle, as well as the azimuth angle of the VRU relative to the vehicle's direction of travel; The degree of chaos in the traffic flow where the vehicle is located is calculated using the relative distance and relative speed to the vehicle, as well as the directional angle of the VRU relative to the vehicle's direction of travel.

[0016] Optionally, the degree of chaos in the traffic flow where the vehicle is located is calculated, including: Assuming the motion of each VRU is a spherical system, for the first VRU... Each VRU, and the principal axis length of its corresponding ellipsoidal system. The three states correspond to the first The relative distance between the VRU and the vehicle, the first The relative speed between the VRU and the vehicle and the first The direction angle of each VRU relative to the vehicle's direction of travel; Input the principal axis length of the ellipsoidal system corresponding to the motion of each VRU into the third calculation formula below to determine the degree of chaos in the traffic flow where the vehicle is located. Third calculation formula:

[0017] in, Indicates the degree of chaos in the traffic flow in which the vehicle is located; This indicates the total number of VRUs included in the traffic flow in which the vehicle is located; , Indicates the first The type weight coefficient of the type to which each VRU belongs.

[0018] Optionally, the vehicle emergency braking method also includes: Calculate the trajectory overlap rate between the predicted trajectory of each VRU and the driving trajectory of the vehicle; Select target VRUs whose trajectory overlap rate is not lower than the preset trajectory overlap threshold; The calculation of the degree of chaos in the traffic flow where the vehicle is located includes: calculating the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each target VRU. The dynamic generation of braking strategy for the vehicle includes: dynamically generating a braking strategy for the vehicle corresponding to the target VRU.

[0019] Optionally, dynamically generating a braking strategy corresponding to the target VRU for the vehicle includes: The degree of chaos Under these conditions, the maximum braking deceleration is set to 6 m / s². 2 ; And / or, When the degree of chaos is greater than or equal to 0.8 and less than 1.3, the maximum braking deceleration is set to 3 m / s². 2 .

[0020] Optionally, if the level of chaos is greater than or equal to 1.3, when the distance to the target VRU exceeds 0.15m, no active braking is applied, and a warning message indicating chaotic traffic flow is given.

[0021] Optionally, dynamically generating a braking strategy for the vehicle corresponding to the target VRU further includes: When the degree of chaos is less than 1.3, Using the following safety distance correction model, a braking safety distance and a warning safety distance are dynamically generated for the vehicle; Safe distance correction model:

[0022] in, This indicates the safe braking distance of the vehicle relative to the target VRU; This indicates the safe warning distance between the vehicle and the target VRU. Indicates the vehicle's speed; This indicates the relative speed of the target VRU with respect to the vehicle; Indicates the vehicle's maximum braking deceleration; This indicates the maximum braking deceleration of the target VRU; Indicates system delay time; Indicates the driver's reaction time; This represents the minimum distance between the vehicle and the target VRU after the vehicle reaches the safe braking distance relative to the target VRU, from the start of braking. Indicates the shortest warning time; This represents the correction factor for low-speed road conditions. The recommended default value is 0.7; Represents the traffic flow chaos correction coefficient, where, , Indicates the scaling factor. ; Indicates the degree of chaos.

[0023] Optionally, a braking strategy corresponding to the target VRU can be dynamically generated for the vehicle, including: displaying warning information corresponding to the level of chaos on the central control screen.

[0024] Optionally, before dynamically generating a braking strategy for the vehicle, the method further includes: Determine whether the vehicle meets the following conditions, and if the vehicle meets the conditions, execute a dynamically generated braking strategy for the vehicle; Conditions: The vehicle is in an outdoor parking environment, in forward gear, and its speed is between 0 km / h and 45 km / h.

[0025] Optionally, the vehicle emergency braking method further includes: sequentially transforming, fitting and smoothing multiple future trajectory points included in the predicted trajectory of each VRU to obtain a trajectory curve for each VRU, and visualizing the trajectory curve of each VRU.

[0026] Optionally, the vehicle emergency braking method is applied to the vehicle emergency braking system. The vehicle emergency braking method further includes: when the dynamic braking mode of the vehicle emergency braking system is activated, performing the step of dynamically generating a braking strategy for the vehicle.

[0027] Optionally, when the dynamic braking mode of the vehicle emergency braking system is turned off, other braking strategies of the vehicle emergency braking system are used to brake the vehicle.

[0028] Optionally, the dynamic braking mode is activated if the environmental information of the traffic flow in which the vehicle is currently located meets the activation conditions of the dynamic braking mode.

[0029] Secondly, embodiments of the present invention provide a vehicle emergency braking device, comprising: a data acquisition module, a processing module, a dynamic strategy generation module, and a braking module, wherein... The data acquisition module is used to collect vehicle driving data and environmental information of the traffic flow in which the vehicle is currently located, wherein the environmental information includes characteristic data of multiple vulnerable road users (VRUs). The processing module is used to input the feature data of multiple VRUs into a pre-trained model, predict the predicted trajectory of each VRU, and calculate the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each VRU. The dynamic strategy generation module is used to dynamically generate a braking strategy for the vehicle based on the trajectory of each VRU and the degree of chaos in the traffic flow where the vehicle is located. A braking module for braking the vehicle using the braking strategy.

[0030] Thirdly, embodiments of the present invention provide an electronic device for emergency braking of a vehicle, the electronic device comprising: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle emergency braking method provided in the first aspect embodiments and related embodiments described above.

[0031] Fourthly, embodiments of the present invention provide a vehicle, characterized in that it includes the vehicle emergency braking device provided in the second aspect embodiment or the vehicle emergency braking electronic device provided in the third aspect embodiment.

[0032] Fifthly, embodiments of the present invention provide a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the vehicle emergency braking method as described in the above embodiments.

[0033] One embodiment of the above invention has the following advantages or beneficial effects: by combining the predicted trajectory of the VRU and the degree of chaos of the traffic flow in which the vehicle is located, a braking strategy is dynamically generated for the vehicle, so that the braking strategy can match the actual scene of the traffic flow in which the vehicle is located. By introducing the predicted trajectory of the VRU, the braking strategy can better ensure the safety of the VRU and can dynamically adapt to the vehicle driving scene, thereby improving the driving comfort of the vehicle.

[0034] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0035] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein: Figure 1This is a schematic diagram of the main process of a vehicle emergency braking method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of the training model according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the trajectory change of the ellipsoidal system abstracted from the VRU according to an embodiment of the present invention from time t0 to time t; Figure 4 This is a schematic diagram of the main components included in a vehicle emergency braking method according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the main process of a vehicle emergency braking method according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the main modules of a vehicle emergency braking device according to an embodiment of the present invention; Figure 7 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied; Figure 8 This is a schematic diagram of the structure of a computer system suitable for implementing embodiments of the present invention. Detailed Implementation

[0036] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0037] It should be noted that, unless otherwise specified, the embodiments of the present invention and the technical features thereof can be combined with each other.

[0038] The vehicle emergency braking method provided in this embodiment of the invention is mainly aimed at road sections with vulnerable road users (VRUs), such as scenarios with large pedestrian traffic in front of shopping malls and complex VRU types, road sections with VRUs on urban roads, and road intersections.

[0039] In this embodiment of the invention, the VRU does not follow a standardized driving route; its driving or movement trajectory is random and complex. This includes pedestrians, children, two-wheeled electric vehicles, three-wheeled electric vehicles, and bicycles in the lane. There are potential for sudden speed changes by children and two-wheeled electric vehicles, as well as potential diagonal and "S-shaped" weaving behaviors by two-wheeled vehicles. In particular, for scenarios with a large number of VRUs and complex traffic conditions, vehicle movement involves even more variables.

[0040] The vehicle emergency braking method provided in this invention combines the predicted trajectory of the VRU with the degree of chaos in the traffic flow where the vehicle is located, which can more accurately assess the scene or environment in which the vehicle is located, thereby formulating a braking strategy that is more matched to the environment or scene in which the vehicle is located, and improving the accuracy and reliability of vehicle intelligent control.

[0041] Specifically, such as Figure 1 As shown, an emergency braking method for a vehicle provided by an embodiment of the present invention may include the following steps: Step S101: Collect vehicle driving data and environmental information of the current traffic flow in which the vehicle is located. The environmental information includes characteristic data of multiple vulnerable road users (VRUs).

[0042] The vehicle's driving data is mainly collected in real time during the vehicle's operation by various sensors (such as cameras, laser sensors, radar sensors, speed sensors, throttle opening sensors, etc.) deployed on the vehicle. This data can include the vehicle's speed, current gear, and throttle opening.

[0043] The environmental information of the traffic flow in which the vehicle is currently located can include environmental information about the surrounding environment collected by sensors on the vehicle (such as external cameras, laser sensors, radar sensors, etc.). In addition, this environmental information can also come from roadside perception systems.

[0044] In addition, the aforementioned driving data and environmental data can also come from the vehicle's perception system.

[0045] Step S102: Input the feature data of multiple VRUs into the pre-trained model to predict the predicted trajectory of each VRU.

[0046] Step S103: Calculate the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the characteristic data of each VRU.

[0047] In complex traffic scenarios, the various VRUs (Vehicle Utility Units) that make up the traffic flow exhibit complex nonlinear relationships, leading to chaos. This chaos arises because the randomness and uncertainty of the VRU's (such as pedestrians and electric two-wheelers) trajectories mean that considering only the potential risk of a single object is insufficient to address the cascading accident risks arising from various random disturbances and interactions within the traffic flow, such as potential emergency avoidance by two-wheelers of pedestrians. To further ensure safety, it is necessary to assess and determine the degree of chaos in the traffic flow composed of various VRUs to evaluate the potential threat posed by each VRU in the current environment.

[0048] The degree of chaos in traffic flow is calculated to quantify the traffic flow in which a vehicle is located, so that the complexity of the traffic flow or the scene in which the vehicle is located can be more accurately assessed through this degree of chaos.

[0049] Step S104: Based on the trajectory of each VRU and the degree of chaos in the traffic flow where the vehicle is located, dynamically generate a braking strategy for the vehicle to brake the vehicle using the braking strategy.

[0050] According to the technical solution provided in the embodiments of the present invention, by combining the predicted trajectory of the VRU and the degree of chaos of the traffic flow in which the vehicle is located, a braking strategy is dynamically generated for the vehicle, so that the braking strategy can match the actual scene of the traffic flow in which the vehicle is located. By introducing the predicted trajectory of the VRU, the braking strategy can better ensure the safety of the VRU and can dynamically adapt to the vehicle driving scene, thereby improving the driving comfort and safety of the vehicle.

[0051] To improve prediction accuracy and reduce the computational resource consumption of the model during the prediction process in step S102, which uses a model to predict the trajectory of the VRU, this embodiment of the invention also provides a training process for the model used in step S102. The model training process can be completed on the vehicle side or the server side. For models trained on the server side, they can be deployed on the vehicle side. For models trained on the vehicle side, the training data mainly consists of data generated and collected during the vehicle's movement. Regardless of whether it's on the vehicle side or the server side, the model training process can include: using historical location information from consecutive frame images of historical vulnerable road users (h-VRUs) included in the collected historical traffic flow environmental historical data to construct historical location vector sequences for each h-VRU; iteratively training an adversarial network algorithm incorporating an attention mechanism based on the historical location vector sequences of each h-VRU; and using diverse loss functions to constrain the training during the iterative training process to obtain a pre-trained model.

[0052] This can be achieved using existing image recognition technology to extract historically vulnerable road users (h-VRUs), such as pedestrians and two-wheeled electric vehicles, from each frame of the image. The historical location vector sequence for each h-VRU refers to the fact that each h-VRU corresponds to a specific historical location vector sequence. ,in, = , Indicates the first The position coordinates of an h-VRU in a frame image, i.e., the historical position vector sequence of a certain h-VRU, are the position coordinates of that h-VRU in each frame image in a continuous series of frames. The position coordinates of the h-VRU in each frame image are generally established based on the same coordinate system.

[0053] In addition, adversarial network algorithms that incorporate attention mechanisms can achieve better prediction results in complex scenes. Furthermore, by learning the spatial correlation of each h-VRU through the introduction of attention mechanisms, high-value information can be filtered out from a large amount of h-VRU information, thus improving computational efficiency.

[0054] Furthermore, during the model training process, the technical solution provided in this embodiment of the invention may further include: inputting the historical position vector sequences of each h-VRU into the encoding module (i.e., encoder) of the variational autoencoder, calculating the mean and variance of each historical position vector sequence (i.e., calculating the mean and variance for each historical position vector sequence); inputting the mean and variance of each historical position vector sequence into the decoding module (i.e., decoder) of the variational autoencoder, generating the input latent variables of each h-VRU (exemplarily, for the aforementioned historical position vector sequence). After passing through a variational autoencoder, the resulting latent input variables are: During the decoding process, new elements, namely input latent variables, are constructed by learning the latent properties of the probability distribution in the latent variable space of the historical position vector sequence. The historical position vector sequence is then processed by a variational autoencoder, first encoding and then decoding. This process compresses (removes irrelevant or less influential data) and then amplifies (increases the correlation between features in the processed data), achieving noise reduction of the historical position vector sequence. This makes the training data used in subsequent training processes more accurate, improving the accuracy and reliability of the trained model's predictions. Based on this, a specific implementation plan for an adversarial network algorithm incorporating an attention mechanism through iterative training can include: training the adversarial network algorithm incorporating an attention mechanism using the input latent variables of each h-VRU.

[0055] Among them, the adversarial network algorithm that introduces the attention mechanism includes a multi-layer LSTM (Long Short-Term Memory Recurrent Neural Network). In this algorithm, a pooling layer is added to each adjacent LSTM space. The pooling layer is used to collect the hidden information of the neighboring h-VRU, thereby realizing information sharing between different VRUs.

[0056] Specifically, the pooling layer collects hidden information of neighboring h-VRUs based on the following calculation formula (1); (1) in, Indicates the first The h-VRU is in the grid at time t. The association between the internal and neighboring h-VRUs; , ) indicates the first The position coordinates of each h-VRU at time t; , ) indicates the relationship with the first The h-VRU is the nearest one The position coordinates of each h-VRU at time t; express Are the sensors collecting environmental information in use? The range of grid data collected at any time Inside, Indicates the first The total number of neighboring h-VRUs of each h-VRU. This indicates that the LSTM output is related to the first... The h-VRU is the nearest one One h-VRU in The hidden state of the LSTM at time step [time]. It's worth noting that this is related to the [time step]. The h-VRU is the nearest one In h-VRU exist Within the range of values. Additionally... The value of is the total number of all VRUs detected from the traffic flow.

[0057] Among them, LSTM consists of a forget gate Input gate and output gate Composed of, among which, the Gate of Oblivion Input gate and output gate The following set of calculation formulas (2) must be satisfied: (2) in, This indicates a deviation from the forget gate; Indicates the deviation of the input gate; Indicates the deviation of the output gate; Indicates the weight of the forget gate; Indicates the weights of the input gates; Indicates the weight of the output gate; This represents the input value at the previous time step; This represents the input value at the current moment. This represents the activation function used for the output of hidden layer neurons. Understandably, and As input to the LSTM, it comes from the output of the previous layer. For the first LSTM layer, its input comes from the output of the variational autoencoder. In other words, the input to each LSTM layer... and They are different values.

[0058] In addition, the diversification loss function is calculated using the following formula (3): (3) in, This represents the model loss value; Indicates hyperparameters, Indicates the first The actual trajectory points of each h-VRU; Indicates the first The predicted trajectory points of h-VRU.

[0059] That is, the corresponding model architecture for the training process, such as Figure 2 As shown, the architecture involved in the model mainly includes two parts. One part is the encoding and decoding of the variational autoencoder (VA). That is, after the feature data of h-VRU (i.e., the historical position vector sequence) is input into the encoder EN of the variational autoencoder (VA), it is encoded by the encoder EN. The output of the encoder EN is directly input into the decoder DE of the variational autoencoder (VA). The decoder DE decodes the output of the encoder EN to obtain the input latent variables. The input latent variables are sent to the Generative Adversarial Network (GAN) algorithm that introduces the attention mechanism. The GAN contains multiple layers of LSTM, and a pooling layer is added to each adjacent LSTM. The pooling layer collects the hidden information of the neighboring h-VRU to achieve the training of the GAN architecture, thereby obtaining the pre-trained model mentioned above.

[0060] The model trained above is particularly effective for low-speed, random, and complex VRU traffic flow scenarios, improving both prediction accuracy and computational efficiency.

[0061] In this embodiment of the invention, a specific implementation plan for calculating the degree of chaos in the traffic flow where a vehicle is located may include: for each VRU, determining the relative distance, relative speed, and azimuth angle of the VRU relative to the vehicle's direction of travel; and using the relative distance, relative speed, and azimuth angle of the VRU relative to the vehicle's direction of travel, calculating the degree of chaos in the traffic flow where the vehicle is located. Here, the relative distance between the VRU and the vehicle refers to the absolute value of the distance between the VRU and the vehicle along the vehicle's direction of travel. The relative speed between the VRU and the vehicle refers to the speed difference between the vehicle and the VRU along the vehicle's direction of travel; this speed difference can be used to monitor and characterize potential speed abrupt changes in VRUs (especially for children, electric two-wheelers, etc.). The azimuth angle of the VRU relative to the vehicle's direction of travel is the angle between the VRU's direction of motion and the vehicle's direction of travel; this azimuth angle can be used to monitor the diagonal and S-shaped weaving behaviors of VRUs (especially electric two-wheelers).

[0062] In the technical solution provided in the embodiments of the present invention, by combining the relative distance and relative speed between the VRU and the vehicle, as well as the directional angle of the VRU relative to the vehicle's direction of travel, during the quantification of the degree of chaos in traffic flow, the calculated degree of chaos can more realistically reflect the traffic flow situation in which the vehicle is located.

[0063] Furthermore, calculating the degree of chaos in the traffic flow where a vehicle is located may also include: assuming the motion of each VRU as an ellipsoidal system, such as... Figure 3 As shown, where, Figure 3 This illustrates the trajectory change of an ellipsoidal system of a VRU from time t0 to time t. For the ... Each VRU, and the principal axis length of its corresponding ellipsoidal system. The three states (such as) Figure 3 shown , and ) respectively correspond to the first The relative distance between the VRU and the vehicle, the first The relative speed between the VRU and the vehicle and the first The direction angle of each VRU relative to the vehicle's direction of travel; input the principal axis length of the ellipsoidal system corresponding to the motion of each VRU into the following calculation formula (4), the degree of chaos of the traffic flow where the vehicle is located; (4) in, Indicates the degree of chaos in the traffic flow in which the vehicle is located; This indicates the total number of VRUs included in the traffic flow in which the vehicle is located; , Indicates the first The type weight coefficients for each VRU type. These weight coefficients can be dynamically adjusted for different VRU types. For example, for bicycles with low trajectory predictability, their... For adult pedestrians whose reactions are sudden and unexpected, For child pedestrians whose movements are sudden and whose trajectories are difficult to predict, For VRUs with high speeds and frequent traffic violations (such as driving against traffic and cutting in front), such as electric bicycles and tricycles, their... .

[0064] Through the above calculations, a specific value of MLE∈(-∞, +∞) can be obtained. Among them, MLE<0 indicates that the traffic flow with VRUs has a low degree of chaos and is in a stable state; MLE>0 indicates that the traffic flow is in a chaotic state, and the higher the value, the higher the degree of chaos, that is, the more chaotic the VRU traffic flow behavior in the surrounding environment.

[0065] Furthermore, in this embodiment of the invention, in order to further reduce the amount of computation and improve the efficiency of braking strategy generation, the above-mentioned vehicle emergency braking method may further include: calculating the trajectory overlap rate between the predicted trajectory of each VRU and the vehicle's driving trajectory; selecting target VRUs with a trajectory overlap rate not lower than a preset trajectory overlap threshold; preferably, the preset trajectory overlap threshold is 20%, that is, selecting VRUs with an overlap rate of 20% or higher between the predicted trajectory and the vehicle's driving trajectory.

[0066] Based on the selected target VRUs, a specific implementation scheme for calculating the degree of chaos in the traffic flow where the vehicle is located may include: calculating the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each target VRU; a specific implementation scheme for dynamically generating braking strategies for the vehicle may include: dynamically generating braking strategies corresponding to the target VRUs for the vehicle. In other words, an optimization scheme provided by this embodiment of the invention focuses only on target VRUs with a trajectory overlap rate not lower than a preset trajectory overlap threshold, and generates braking strategies for these target VRUs, ensuring vehicle driving safety while effectively improving computational efficiency and reducing the consumption of computational resources.

[0067] It is worth noting that for VRUs other than the target VRU, the vehicle's existing braking system braking strategy, such as fixed safe distance braking, can be used directly to control the vehicle.

[0068] Furthermore, to improve the braking effect of the above braking strategy, before dynamically generating the braking strategy for the vehicle, the emergency braking method may further include: determining whether the vehicle meets the following conditions; if the vehicle meets these conditions, dynamically generating the braking strategy for the vehicle; conditions: the vehicle is in an outdoor parking scenario, the vehicle is in a forward gear, and the vehicle speed is between 0 km / h and 45 km / h. That is, the braking scheme provided in this embodiment of the invention is used to control the vehicle only when the vehicle is in an outdoor parking scenario, the vehicle is in a forward gear, and the vehicle speed is between 0 km / h and 45 km / h. If the vehicle is in an indoor environment or the vehicle speed exceeds 45 km / h, the vehicle's built-in conventional braking strategy is used to brake the vehicle, thereby improving the safety and reliability of vehicle braking.

[0069] More specifically, a specific implementation of dynamically generating a braking strategy for a vehicle corresponding to a target VRU may include: in terms of the degree of chaos Under these conditions, the maximum braking deceleration is set to 6 m / s². 2 .

[0070] Another specific implementation of dynamically generating a braking strategy for the vehicle corresponding to the target VRU may include setting the maximum braking deceleration to 3 m / s² when the chaos level is greater than or equal to 0.8 and less than 1.3. 2 .

[0071] Another specific implementation of dynamically generating a braking strategy for a vehicle corresponding to a target VRU may include: when the chaos level is greater than or equal to 1.3, not taking active braking when the distance to the target VRU is more than 0.15 m, and giving a prompt message indicating chaotic traffic flow.

[0072] In addition to the aforementioned braking strategies, differentiated information prompts can be used to alert the driver for different braking strategies. For example, when the chaos level is less than 0.8, yellow flashing is triggered on both sides of the central control screen. When the chaos level is greater than or equal to 0.8 but less than 1.3, green flashing is triggered on both sides of the central control screen. When the chaos level is greater than or equal to 1.3, and the distance to the target VRU exceeds 0.15 m, red flashing is triggered on both sides of the central control screen, and a message can be displayed on the central control screen stating, "Current traffic flow is chaotic; vehicles will not intervene excessively. Please be aware of oncoming two-wheeled vehicles and pedestrians." It is worth noting that the above prompts and warnings are only examples; in actual use, other differentiated prompts and warnings can also be employed.

[0073] Furthermore, in practical applications, the thresholds at each level of the chaos level classification strategy can be tested and calibrated according to actual conditions. The final control strategy is executed collaboratively by the braking system and the information prompting system.

[0074] Furthermore, the specific implementation plan for dynamically generating a braking strategy corresponding to the target VRU for the vehicle may also include: when the degree of chaos is less than 1.3, using the following safety distance correction model (5), dynamically generating a braking safety distance and a warning safety distance for the vehicle; (5) in, This indicates the safe braking distance of the vehicle relative to the target VRU; This indicates the safe warning distance between the vehicle and the target VRU. Indicates the vehicle's speed; This indicates the relative speed of the target VRU with respect to the vehicle; Indicates the vehicle's maximum braking deceleration; This indicates the maximum braking deceleration of the target VRU; Indicates system delay time; Indicates the driver's reaction time; This represents the minimum distance between the vehicle and the target VRU after the vehicle reaches the safe braking distance relative to the target VRU, from the start of braking. Indicates the shortest warning time; This represents the correction factor for low-speed road conditions. The recommended default value is 0.7; Represents the traffic flow chaos correction coefficient, where, , This represents the set scaling factor. ; This indicates the degree of chaos in the traffic flow in which the vehicle is located.

[0075] Furthermore, the aforementioned vehicle emergency braking method may also include: sequentially transforming multiple future trajectory points included in the predicted trajectory of each VRU through coordinate mapping, curve fitting, and smoothing to obtain trajectory curves for each VRU, and visualizing the trajectory curves of each VRU. This allows the driver to clearly understand the actual traffic flow situation in which the vehicle is located.

[0076] The aforementioned vehicle emergency braking method is applied to a vehicle emergency braking system. Accordingly, the vehicle emergency braking method may further include: when the dynamic braking mode of the vehicle emergency braking system is activated, executing a step of dynamically generating a braking strategy for the vehicle. That is, by having both the dynamic braking mode and the conventional braking mode inherent in the vehicle emergency braking system coexist, the user can choose to activate or deactivate the dynamic braking mode, thereby further enhancing the user experience of the vehicle emergency braking system.

[0077] When the dynamic braking mode of the vehicle's emergency braking system is turned off, other braking strategies of the vehicle's emergency braking system are used to brake the vehicle.

[0078] In addition, dynamic braking mode can be activated when the environmental information of the current traffic flow meets the activation conditions. This means that dynamic braking mode is automatically activated without user intervention, improving the driving experience and vehicle safety in complex traffic conditions.

[0079] In conclusion, as follows: Figure 4As shown, the technical solution provided by this invention is mainly divided into several parts: the first part (Ly1) VRU detection and tracking, the second part (Ly2) VRU trajectory prediction, the third part (Ly3) trigger condition judgment, the fourth part (Ly4) chaotic discrimination of traffic flow where the vehicle is located, and the fifth part (Ly5) vehicle dynamic braking. Specifically, for the first part (Ly1) VRU detection and tracking: it can track and detect the VRUs around the vehicle by using environmental data (such as images) detected by sensors such as radar Ra and camera Ca, and extract the feature data Data1 of the VRU from the environmental data, and provide Data1 to the second part (Ly2). The second part (Ly2) predicts the trajectory of the VRU based on Data1. Specifically, in the second part (Ly2) VRU trajectory prediction, it is necessary to train the model used to predict the VRU trajectory, and obtain the predicted trajectory Data2 for each VRU by inputting Data1 into the trained model, and provide the predicted trajectory Data2 to the fifth part (Ly5). The third part (Ly3) of the trigger condition judgment mainly analyzes the data collected in the first part (Ly1) to detect whether the vehicle's environment meets the conditions for initiating vehicle dynamic braking (i.e., the fifth part, Ly5). If the third part (Ly3) determines that the vehicle meets the conditions for initiating vehicle dynamic braking, it can trigger the first part (Ly1) to perform VRU detection and tracking. Alternatively, the third part (Ly3) can also directly activate the fifth part (Ly5), and the VRU detection and tracking performed by the first part (Ly1) can be unaffected by the third part (Ly3). The fourth part (Ly4) of the traffic flow chaos judgment requires information obtained from the first part (Ly1), from which the relative distance between the VRU and the vehicle, the relative speed between the VRU and the vehicle, and the azimuth angle of the VRU relative to the vehicle's direction of travel are extracted. Then, by abstracting the VRU as an ellipsoidal system, the chaos level Data3 of the VRU is calculated based on this. The predicted trajectory Data2 of the VRU obtained from the second part (Ly2) and the chaos level Data3 obtained from the fourth part (Ly4) are provided to the fifth part (Ly5). Part 5 (Ly5) Vehicle Dynamic Braking: Step 1 (St1): Select the target VRU based on the overlap rate between the VRU's predicted trajectory (Data2) and the vehicle's driving trajectory. Step 2 (St2): Construct a modified safety distance correction model based on the basic safety distance model. Step 3 (St3): When the chaos level (Data3) is below 1.3, dynamically generate braking safety distance and warning safety distance for the vehicle using the safety distance correction model. Step 4 (St4): Perform graded dynamic braking for the vehicle according to the chaos level. In the graded dynamic braking process, the braking safety distance and warning safety distance are provided by Step 3 (St3).

[0080] Figure 4The process steps achieved by the cooperation between the different parts shown are as follows: Figure 5 As shown. Figure 5 As shown, achieving dynamic braking of a vehicle mainly includes the following steps: Step S501: Trigger condition judgment. Based on the data detected by Ly1 in the first part, determine whether the vehicle meets the conditions for dynamic braking (the vehicle is in an outdoor parking scene, the vehicle is in forward gear, and the vehicle speed is between 0 km / h and 45 km / h).

[0081] Step 502, VRU detection and tracking: Detect and track VRUs from the data detected in the first part Ly1.

[0082] Step 503, trajectory prediction: Using the detected and tracked VRU location information Data1, the trajectory of the VRU is predicted, and the predicted trajectory Data2 of the VRU is input into the fifth part Ly5.

[0083] Step 504, Traffic flow chaos determination: Based on the relative distance between the VRU and the vehicle, the relative speed between the VRU and the vehicle, and the directional angle of the VRU relative to the vehicle's direction of travel, calculate the degree of chaos in the traffic flow where the vehicle is located (Data3).

[0084] Step S505, target VRU screening: Based on the overlap rate between the predicted trajectory Data2 of the VRU and the vehicle's driving trajectory, target VRUs are screened.

[0085] Step S506: Dynamically generate braking safety distance and warning safety distance for the vehicle using the safety distance correction model.

[0086] Specifically, the safety distance correction model in step S506 is obtained by correcting the basic safety distance model in step S500. Step S500 can be constructed before the method of dynamically braking the vehicle.

[0087] Step S507: Generate a braking strategy for the vehicle in a hierarchical and dynamic manner. This step takes into account the degree of chaos, braking safety distance and warning safety distance to dynamically generate a braking strategy for the vehicle.

[0088] Step S508: Perform vehicle braking, coordinate the various systems of the vehicle to brake the vehicle according to the braking strategy, and provide collision warning and alert.

[0089] Compared to existing vehicle braking strategies that only consider the static characteristics of the vehicle and the environment, resulting in frequent braking triggers in low-speed traffic flows dominated by VRUs, the technical solution provided by this invention balances driving safety and ease of handling. By assessing the degree of potential danger in the current environment through the degree of chaos, it enables AEB to dynamically adjust and selectively intervene or not intervene based on the traffic environment, thereby reducing intervention frequency and trigger distance while ensuring active safety performance. It also resolves the conflict over driving rights caused by excessive program intervention, thus improving traffic efficiency. In particular, the technical solution provided by this invention is especially effective in low-speed congestion scenarios where various VRUs converge, such as auxiliary roads and intersections. It achieves multi-objective optimization in complex, chaotic, and random human-vehicle-environment multi-coupling and interaction systems, improving driving safety, ease of handling, and traffic efficiency, and is expected to resolve the conflict over driving rights caused by excessive program intervention. Furthermore, this research can also provide a reference for decision-making regarding vehicle collision avoidance with pedestrians, two-wheeled vehicles, and other vulnerable road users, and for the development of active safety technologies.

[0090] In summary, the technical solution provided by the embodiments of the present invention, through reasonable selection and improvement of algorithms and optimization of braking strategies in combination with the degree of chaos, improves active driving safety while taking into account the ease of vehicle handling and the scalability of intelligent driving systems, and has significant technological progress and industrial application prospects.

[0091] Furthermore, embodiments of the present invention provide a vehicle emergency braking device. For example... Figure 6 As shown, the vehicle emergency braking device 600 may include: a data acquisition module 601, a processing module 602, a dynamic strategy generation module 603, and a braking module 604, wherein... The data acquisition module 601 is used to collect vehicle driving data and environmental information of the current traffic flow in which the vehicle is located. The environmental information includes the characteristic data of multiple vulnerable road users (VRUs). The processing module 602 is used to input the feature data of multiple VRUs into a pre-trained model, predict the predicted trajectory of each VRU, and calculate the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each VRU. The dynamic strategy generation module 603 is used to dynamically generate braking strategies for vehicles based on the trajectories of each VRU and the degree of chaos in the traffic flow where the vehicle is located. Braking module 604 is used to brake a vehicle using a braking strategy.

[0092] In embodiments of the present invention, such as Figure 6As shown, the vehicle emergency braking device 600 may further include: a model training module 605, which is used to construct a historical position vector sequence for each h-VRU by utilizing the historical position information in the continuous frame images of each historical vulnerable road user (h-VRU) included in the historical traffic flow environmental historical data; based on the historical position vector sequence of each h-VRU, iteratively training an adversarial network algorithm with an attention mechanism, and using diverse loss functions to constrain the training during the iterative training process to obtain a pre-trained model.

[0093] In this embodiment of the invention, the model training module 605 is further configured to input the historical position vector sequences of each h-VRU into the encoding module of the variational autoencoder to calculate the mean and variance of each historical position vector sequence; input the mean and variance of each historical position vector sequence into the decoding module of the variational autoencoder to generate the input latent variables of each h-VRU; and use the input latent variables of each h-VRU to train the adversarial network algorithm that incorporates the attention mechanism.

[0094] In this embodiment of the invention, the adversarial network algorithm that introduces an attention mechanism includes a multi-layer LSTM, wherein a pooling layer is added to each adjacent LSTM in the space, and the pooling layer is used to collect hidden information of the neighboring h-VRU.

[0095] In this embodiment of the invention, the pooling layer collects hidden information of neighboring h-VRUs based on the following calculation formula (1); (1) in, express Are the sensors collecting environmental information in use? The range of grid data collected at any time Inside, Indicates the first The neighboring h-VRUs of each h-VRU, Indicates the LSTM output of the first... One h-VRU in The hidden state of the LSTM at time step.

[0096] In this embodiment of the invention, LSTM is constructed using a forget gate. Input gate and output gate Composed of, among which, the Gate of Oblivion Input gate and output gate The following set of calculation formulas (2) must be satisfied: (2) in, This indicates a deviation from the forget gate; Indicates the deviation of the input gate; Indicates the deviation of the output gate; Indicates the weight of the forget gate; Indicates the weights of the input gates; Indicates the weight of the output gate; This represents the input value at the previous time step; This represents the input value at the current moment. This represents the activation function used for the output of hidden layer neurons.

[0097] In this embodiment of the invention, the diversification loss function is calculated using the following formula (3): (3) in, Indicates hyperparameters, This represents the actual trajectory points of the h-VRU; This represents the predicted trajectory point of h-VRU.

[0098] In this embodiment of the invention, the processing module is further configured to determine, for each VRU, the relative distance and relative speed between the VRU and the vehicle, and the directional angle of the VRU relative to the vehicle's direction of travel; and to calculate the degree of chaos in the traffic flow where the vehicle is located using the relative distance and relative speed between the VRU and the vehicle, and the directional angle of the VRU relative to the vehicle's direction of travel.

[0099] In this embodiment of the invention, the processing module 602 is further configured to assume the motion of each VRU as an ellipsoidal system, for the first VRU... Each VRU, and the principal axis length of its corresponding ellipsoidal system. The three states correspond to the first The relative distance between the VRU and the vehicle, the first The relative speed between the VRU and the vehicle and the first The direction angle of each VRU relative to the vehicle's driving direction; input the principal axis length of the ellipsoidal system corresponding to the motion of each VRU into the following calculation formula (4), the degree of chaos of the traffic flow where the vehicle is located; (4) in, Indicates the degree of chaos in the traffic flow in which the vehicle is located; This indicates the total number of VRUs included in the traffic flow in which the vehicle is located; , Indicates the first The type weight coefficient of the type to which each VRU belongs.

[0100] In this embodiment of the invention, the processing module 602 is further configured to calculate the trajectory overlap rate between the predicted trajectory of each VRU and the driving trajectory of the vehicle; filter out target VRUs with a trajectory overlap rate not lower than a preset trajectory overlap threshold; calculate the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each target VRU; and dynamically generate a braking strategy corresponding to the target VRU for the vehicle.

[0101] In this embodiment of the invention, the dynamic strategy generation module 603 is further configured to adjust the degree of chaos. Under these conditions, the maximum braking deceleration is set to 6 m / s². 2 .

[0102] In this embodiment of the invention, the dynamic strategy generation module 603 is further configured to set the maximum braking deceleration to 3 m / s² when the chaos level is greater than or equal to 0.8 and less than 1.3. 2 .

[0103] In this embodiment of the invention, the dynamic strategy generation module is further configured to not take active braking when the distance to the target VRU exceeds 0.15 m when the chaos level is greater than or equal to 1.3, and to provide a prompt message indicating chaotic traffic flow.

[0104] In this embodiment of the invention, the dynamic strategy generation module 603 is further used to dynamically generate a braking safety distance and a warning safety distance for the vehicle using the following safety distance correction model (5) when the degree of chaos is less than 1.3. (5) in, This indicates the safe braking distance of the vehicle relative to the target VRU; This indicates the safe warning distance between the vehicle and the target VRU. Indicates the vehicle's speed; This indicates the relative speed of the target VRU with respect to the vehicle; Indicates the vehicle's maximum braking deceleration; This indicates the maximum braking deceleration of the target VRU; Indicates system delay time; Indicates the driver's reaction time; This represents the minimum distance between the vehicle and the target VRU after the vehicle reaches the safe braking distance relative to the target VRU, from the start of braking. Indicates the shortest warning time; This represents the correction factor for low-speed road conditions. The recommended default value is 0.7; Represents the traffic flow chaos correction coefficient, where, , Indicates the scaling factor. ; Indicates the degree of chaos.

[0105] In this embodiment of the invention, the dynamic strategy generation module 603 is further used to display warning information corresponding to the degree of chaos through the central control screen.

[0106] In this embodiment of the invention, the dynamic strategy generation module 603 is further used to determine whether the vehicle meets the following conditions. If the vehicle meets the conditions, a braking strategy is dynamically generated for the vehicle. The conditions are: the vehicle is in an outdoor parking scene, the vehicle is in forward gear, and the vehicle speed is between 0 km / h and 45 km / h.

[0107] In this embodiment of the invention, the processing module 602 is further used to sequentially transform, curve fit and smooth the multiple future trajectory points included in the predicted trajectory of each VRU to obtain the trajectory curve for the VRU, and visualize the trajectory curve of each VRU.

[0108] The aforementioned vehicle emergency braking device is applied to the vehicle emergency braking system. It can be used as a functional device of the vehicle emergency braking system. When the dynamic braking mode of the vehicle emergency braking system is activated, the dynamic strategy generation module of this functional device performs the step of dynamically generating a braking strategy for the vehicle.

[0109] Furthermore, in the vehicle emergency braking system, a dynamic braking mode is configured for this function. When the environmental information of the traffic flow where the vehicle is currently located meets the activation conditions of the dynamic braking mode, the dynamic braking mode is activated.

[0110] Furthermore, in the vehicle emergency braking system, when the dynamic braking mode of the vehicle emergency braking system is turned off, other braking strategies of the vehicle emergency braking system are used to brake the vehicle.

[0111] Furthermore, embodiments of the present invention provide an electronic device for emergency braking of a vehicle, which is applied to a vehicle. Specifically, the electronic device for emergency braking of a vehicle may include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle emergency braking method provided in any of the above embodiments.

[0112] Furthermore, embodiments of the present invention provide a vehicle. This vehicle may include the vehicle emergency braking device provided in the above embodiments or the electronic device for vehicle emergency braking provided in the above embodiments.

[0113] In addition, the aforementioned vehicle emergency braking device can also be applied to servers.

[0114] Figure 7 An exemplary vehicle system architecture 700 is shown, to which the vehicle emergency braking method or vehicle emergency braking device of the present invention can be applied.

[0115] like Figure 7 As shown, the vehicle system architecture 700 may include various systems, such as a driving control system 701, a power system 702, a sensor system 703, a control system 704, a vehicle emergency braking system 705, one or more peripheral devices 706, a power supply 707, a computer system 708, and a user interface 709. The vehicle emergency braking method provided in this embodiment can be implemented through interaction with the aforementioned systems. Optionally, the vehicle system architecture 700 may include more or fewer systems, and each system may include multiple components. Furthermore, each system and component of the vehicle system architecture 700 may be interconnected via wired or wireless means.

[0116] The vehicle system architecture 700 includes a driving control system 701, which can be in a fully or partially automated driving mode. For example, the driving control system 701 can brake the vehicle under the control of the vehicle emergency braking system 705 without human interaction, and the driving control system 701 can also drive the vehicle.

[0117] The powertrain 702 may include components that provide power to the vehicle. For example, the powertrain 702 may include an engine, an energy source, a transmission, wheels, tires, etc. The engine may be an internal combustion engine, an electric motor, an air-compressed engine, or other combinations of engines, such as a hybrid engine consisting of a gasoline engine and an electric motor, or a hybrid engine consisting of an internal combustion engine and an air-compressed engine. The engine converts the energy source into mechanical energy to supply the transmission. Examples of energy sources may include gasoline, diesel, other petroleum-based fuels, propane, other compressed gas-based fuels, ethanol, solar panels, batteries, and other electrical sources. The energy source may also provide energy to other systems in the vehicle. Furthermore, the transmission may include a gearbox, a differential, a drive shaft, and a clutch, etc.

[0118] Sensor system 703 may include sensors for sensing the vehicle's surrounding environment and pressure sensors for sensing whether there are passengers in the seats. Examples include a positioning system (which may be a Global Positioning System (GPS), BeiDou Navigation Satellite System, or other positioning systems), radar, a laser rangefinder, an inertial measurement unit (IMU), a camera, and metal detection devices such as capacitive sensors. The positioning system can be used to determine the vehicle's geographical location. The IMU is used to sense changes in the vehicle's position and orientation based on inertial acceleration. In one embodiment, the IMU may be a combination of an accelerometer and a gyroscope. Radar can use radio signals to sense objects in the vehicle's surrounding environment. In some embodiments, in addition to sensing objects, radar can also be used to sense the speed and / or direction of travel of objects.

[0119] To detect environmental information and objects outside the vehicle, cameras can be configured at appropriate locations on the vehicle's exterior. For example, to acquire environmental images of the vehicle's sides, a camera can be mounted on the side mirror. The camera can be a still or video camera.

[0120] The control system 704 may include software systems for implementing vehicle driving control, such as systems for analyzing the vehicle's surrounding environment, pretensioning seat belts, route planning, obstacle avoidance, and image analysis. The control system 1604 may also include hardware systems such as an accelerator, steering wheel system, seat belt system, and airbag system. Furthermore, the control system 704 may add or replace components other than those shown and described. Alternatively, some of the components shown above may be omitted.

[0121] Furthermore, as described above, the control system 704 can also serve as an auxiliary system to the aforementioned vehicle emergency braking system 705.

[0122] Additionally, the control system 704 can interact with external sensors, other autonomous driving devices, other computer systems, or users via peripheral devices 706. Peripheral devices 706 may include wireless communication systems, on-board computers, microphones, and / or speakers.

[0123] In some embodiments, peripheral device 706 provides a means for user interaction with the control system 704 via a user interface. For example, an onboard computer may provide information to a user of the vehicle. The user interface may also operate the onboard computer to receive user input. The onboard computer may be operated via a touchscreen. In other cases, peripheral device may provide a means for communicating with other devices located within the vehicle. For example, a microphone may receive audio (e.g., voice commands or other audio input) from a user of the control system 704. Similarly, a speaker may output audio to a user of the control system 704.

[0124] Wireless communication systems can communicate wirelessly with one or more devices, either directly or via a communication network. For example, wireless communication systems can use networks such as cellular networks, WiFi, and wireless local area networks (WLANs), or they can use infrared links, Bluetooth, or ZigBee to communicate directly with devices. Other wireless protocols include those used in various autonomous driving communication systems.

[0125] The power source 707 can provide power to various components of the vehicle. The power source 707 can be a rechargeable lithium-ion or lead-acid battery.

[0126] Some or all of the functions enabling emergency braking of the vehicle are controlled by a computer system 708. The computer system 708 may include at least one processor that executes instructions stored in a non-transitory computer-readable medium such as memory. The computer system 708 provides execution code for the aforementioned control system to implement vehicle safety protection.

[0127] The processor can be any conventional processor, such as a commercially available central processing unit (CPU). Alternatively, the processor can be a special-purpose device such as an application-specific integrated circuit (ASIC) or other hardware-based processor. Those skilled in the art will understand that the processor, computer, or memory can actually include multiple processors, computers, or memories that may or may not be stored in the same physical housing. For example, memory can be a hard disk drive or other storage media located in a housing different from that of a computer. Therefore, references to processors or computers will be understood to include references to a collection of processors or computers or memories that may or may not operate in parallel. Unlike using a single processor to perform the steps described herein, some components, such as steering and deceleration components, may each have their own processor that performs calculations only related to the component's specific function.

[0128] User interface 709 is used to provide information to or receive information from a user of the vehicle. Optionally, user interface 709 may include one or more input / output devices within a set of peripheral devices 706, such as wireless communication systems, on-board computers, microphones, and speakers.

[0129] It should be understood that the components described above are merely an example. In actual applications, components in the various modules or systems mentioned above may be added or removed as needed. Figure 7 This should not be construed as a limitation on the embodiments of this application.

[0130] The following is for reference. Figure 8 It shows a schematic diagram of the structure of a computer system 800 suitable for implementing an emergency braking method for a vehicle according to an embodiment of the present invention. Figure 8 The computer system shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0131] like Figure 8 As shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 802 or programs loaded from storage section 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the system 800. The CPU 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0132] The following components are connected to I / O interface 805: an input section 806; an output section 807 including devices such as cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; a storage section 808 including devices such as hard disks; and a communication section 809 including network interface cards such as LAN cards and modems. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 810 as needed so that computer programs read from it can be installed into storage section 808 as needed.

[0133] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention 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 communication section 809, and / or installed from removable medium 811. When the computer program is executed by central processing unit (CPU) 801, it performs the functions defined above in the system of this invention.

[0134] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, 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 device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0135] 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 the present invention. 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 a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may 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.

[0136] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor; for example, a processor can be described as including the aforementioned data acquisition module, processing module, dynamic strategy generation module, and braking module. The names of these modules or units do not necessarily limit the module or unit itself; for example, a data acquisition module can also be described as "a module or unit that collects vehicle driving data and environmental information about the current traffic flow of the vehicle."

[0137] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs that, when executed by the device, cause the device to include: collecting vehicle driving data and environmental information of the traffic flow in which the vehicle is currently located, wherein the environmental information includes feature data of multiple vulnerable road users (VRUs); inputting the feature data of the multiple VRUs into a pre-trained model to predict the predicted trajectory of each VRU; calculating the degree of chaos in the traffic flow in which the vehicle is located based on the vehicle driving data and the feature data of each VRU; and dynamically generating a braking strategy for the vehicle based on the trajectories of each VRU and the degree of chaos in the traffic flow in which the vehicle is located, so as to brake the vehicle using the braking strategy.

[0138] According to the technical solution of the present invention, by combining the predicted trajectory of the VRU and the degree of chaos of the traffic flow in which the vehicle is located, a braking strategy is dynamically generated for the vehicle, so that the braking strategy can match the actual scene of the traffic flow in which the vehicle is located. By introducing the predicted trajectory of the VRU, the braking strategy can better ensure the safety of the VRU and can dynamically adapt to the vehicle driving scene, thereby improving the driving comfort of the vehicle.

[0139] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A vehicle emergency braking method, characterized in that, include: Collect vehicle driving data and environmental information of the traffic flow in which the vehicle is currently located, wherein the environmental information includes characteristic data of multiple vulnerable road users; The feature data of multiple vulnerable road users are input into a pre-trained model to predict the predicted trajectory of each vulnerable road user. The degree of chaos in the traffic flow where the vehicle is located is calculated based on the vehicle's driving data and the characteristic data of each vulnerable road user. Based on the trajectories of each vulnerable road user and the degree of chaos in the traffic flow in which the vehicle is located, a braking strategy is dynamically generated for the vehicle to brake using the braking strategy.

2. The vehicle emergency braking method according to claim 1, characterized in that, Also includes: By utilizing the historical location information in consecutive frame images of each historically disadvantaged road user from the collected historical traffic flow environmental historical data, a historical location vector sequence is constructed for each of the historically disadvantaged road users. Based on the historical location vector sequences of each historically disadvantaged road user, an adversarial network algorithm with an attention mechanism is iteratively trained, and during the iterative training process, a variety of loss functions are used to constrain the training to obtain a pre-trained model.

3. The vehicle emergency braking method according to claim 2, characterized in that, Also includes: The historical location vector sequences of each historically disadvantaged road user are input into the encoding module of the variational autoencoder to calculate the mean and variance of each historical location vector sequence. The mean and variance of each of the historical location vector sequences are input into the decoding module of the variational autoencoder to generate the input latent variables of each of the historical vulnerable road users. The iterative training of the adversarial network algorithm with an attention mechanism includes: training the adversarial network algorithm with an attention mechanism using the input latent variables of each of the historically disadvantaged road users.

4. The vehicle emergency braking method according to claim 2 or 3, characterized in that, Adversarial network algorithms that incorporate attention mechanisms include multi-layer LSTM, wherein a pooling layer is added to each adjacent LSTM in the space, and the pooling layer is used to collect hidden information of nearby historically disadvantaged road users.

5. The vehicle emergency braking method according to claim 4, characterized in that, The pooling layer collects hidden information of nearby historically disadvantaged road users based on the following first calculation formula; First calculation formula: in, express Are the sensors collecting environmental information in use? The range of grid data collected at any time Inside, Indicates the first The neighboring h-VRUs of each h-VRU, Indicates the LSTM output of the first... One h-VRU in The hidden state of the LSTM at time step.

6. The vehicle emergency braking method according to claim 4, characterized in that, LSTM by the forget gate Input gate and output gate Composed of, among which, the Gate of Oblivion Input gate and output gate The following set of calculation formulas must be satisfied: Calculation formula set: in, This indicates a deviation from the forget gate; Indicates the deviation of the input gate; Indicates the deviation of the output gate; Indicates the weight of the forget gate; Indicates the weights of the input gates; Indicates the weight of the output gate; This represents the input value at the previous time step; This represents the input value at the current moment. This represents the activation function used for the output of hidden layer neurons.

7. The vehicle emergency braking method according to any one of claims 2, 3, 5, and 6, characterized in that, The diversification loss function is calculated using the following second formula: in, Indicates hyperparameters, This represents the actual trajectory points of the h-VRU; This represents the predicted trajectory point of h-VRU.

8. The vehicle emergency braking method according to claim 1, characterized in that, The calculation of the degree of chaos in the traffic flow where the vehicle is located includes: For each vulnerable road user, determine the relative distance and relative speed between the vulnerable road user and the vehicle, as well as the azimuth angle of the vulnerable road user relative to the vehicle's direction of travel; The degree of chaos in the traffic flow where the vehicle is located is calculated using the relative distance and relative speed to the vehicle, as well as the directional angle of the vulnerable road user relative to the vehicle's direction of travel.

9. The vehicle emergency braking method according to claim 8, characterized in that, Calculating the degree of chaos in the traffic flow where a vehicle is located also includes: Assuming the motion of each vulnerable road user is a ellipsoidal system, for the first... For a vulnerable road user, the principal axis length of the corresponding ellipsoidal system The three states correspond to the first The relative distance between vulnerable road users and vehicles, the first The relative speed between vulnerable road users and vehicles and the first The directional angle of a vulnerable road user relative to the direction of vehicle travel; Input the principal axis length of the ellipsoidal system corresponding to the motion of each VRU into the third calculation formula below to determine the degree of chaos in the traffic flow where the vehicle is located. Third calculation formula: in, Indicates the degree of chaos in the traffic flow in which the vehicle is located; This indicates the total number of vulnerable road users included in the traffic flow in which the vehicle is located; , Indicates the first The type weight coefficient of the type to which the vulnerable road user belongs.

10. The vehicle emergency braking method according to claim 1 or 8, characterized in that, Also includes: Calculate the trajectory overlap rate between the predicted trajectory of each vulnerable road user and the driving trajectory of the vehicle; Select vulnerable road users whose trajectory overlap rate is not lower than a preset trajectory overlap threshold; The calculation of the degree of chaos in the traffic flow where the vehicle is located includes: calculating the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the characteristic data of each of the target vulnerable road users. The dynamic generation of braking strategies for the vehicle includes: dynamically generating braking strategies for the vehicle corresponding to the target vulnerable road user.

11. The vehicle emergency braking method according to claim 10, characterized in that, Dynamically generating a braking strategy for the vehicle corresponding to the target vulnerable road user, including: The degree of chaos Under these conditions, the maximum braking deceleration is set to 6 m / s². 2 ; And / or, When the degree of chaos is greater than or equal to 0.8 and less than 1.3, the maximum braking deceleration is set to 3 m / s². 2 ; And / or, When the level of chaos is greater than or equal to 1.3, no active braking will be applied when the distance to the target vulnerable road user exceeds 0.15m, and a warning message indicating chaotic traffic flow will be given.

12. The vehicle emergency braking method according to claim 1 or 11, characterized in that, The method of dynamically generating a braking strategy for the vehicle corresponding to the target vulnerable road user further includes: When the degree of chaos is less than 1.3, Using the following safety distance correction model, a braking safety distance and a warning safety distance are dynamically generated for the vehicle; Safe distance correction model: in, This indicates the safe braking distance of a vehicle relative to a vulnerable road user. This indicates the safe warning distance between the vehicle and the vulnerable road user. Indicates the vehicle's speed; Indicates the relative speed of the target vulnerable road user relative to the vehicle; Indicates the vehicle's maximum braking deceleration; Indicates the maximum braking deceleration for the target vulnerable road user; Indicates system delay time; Indicates the driver's reaction time; This represents the minimum distance between the vehicle and the target vulnerable road user after the vehicle has reached a safe braking distance relative to the target vulnerable road user, from the start of braking. Indicates the shortest warning time; This represents the correction factor for low-speed road conditions. The recommended default value is 0.7; Represents the traffic flow chaos correction coefficient, where, , Indicates the scaling factor. ; Indicates the degree of chaos.

13. The vehicle emergency braking method according to claim 1 or 10, characterized in that, Dynamically generating a braking strategy for the vehicle corresponding to the target vulnerable road user, including: Warning information corresponding to the level of chaos is displayed on the central control screen.

14. The vehicle emergency braking method according to any one of claims 1, 8, 9 and 11, characterized in that, Prior to dynamically generating a braking strategy for the vehicle, the method further includes: Determine whether the vehicle meets the following conditions, and if the vehicle meets the conditions, execute a dynamically generated braking strategy for the vehicle; Conditions: The vehicle is in an outdoor parking environment, in forward gear, and its speed is between 0 km / h and 45 km / h.

15. The vehicle emergency braking method according to claim 1, characterized in that, Also includes: The predicted trajectories of each vulnerable road user include multiple future trajectory points that are sequentially transformed by coordinate mapping, curve fitting, and smoothing to obtain trajectory curves for each vulnerable road user, and the trajectory curves of each vulnerable road user are visualized.

16. The vehicle emergency braking method according to claim 1, characterized in that, Used in vehicle emergency braking systems, The vehicle emergency braking method further includes: when the dynamic braking mode of the vehicle emergency braking system is activated, performing the step of dynamically generating a braking strategy for the vehicle.

17. The vehicle emergency braking method according to claim 16, characterized in that, When the dynamic braking mode of the vehicle emergency braking system is turned off, the vehicle is braked using other braking strategies of the vehicle emergency braking system. And / or, When the environmental information of the traffic flow in which the vehicle is currently located meets the activation conditions of the dynamic braking mode, the dynamic braking mode is activated.

18. A vehicle emergency braking device, characterized in that, include: The system comprises a data acquisition module, a processing module, a dynamic strategy generation module, and a braking module. The data acquisition module is used to collect vehicle driving data and environmental information of the traffic flow in which the vehicle is currently located, wherein the environmental information includes characteristic data of multiple vulnerable road users. The processing module is used to input the feature data of multiple vulnerable road users into a pre-trained model, predict the predicted trajectory of each vulnerable road user, and calculate the degree of chaos in the traffic flow where the vehicle is located based on the vehicle's driving data and the feature data of each vulnerable road user. The dynamic strategy generation module is used to dynamically generate a braking strategy for the vehicle based on the trajectory of each vulnerable road user and the degree of chaos in the traffic flow where the vehicle is located. A braking module for braking the vehicle using the braking strategy.

19. An electronic device for emergency braking of a vehicle, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle emergency braking method as described in claims 1-17.

20. A vehicle, characterized in that, This includes the vehicle emergency braking device as described in claim 18 or the electronic device for vehicle emergency braking as described in claim 19.

Images