A passive safety vehicle distance control system based on monocular perception

Abstract

The application provides a passive safety distance control system based on monocular perception, and relates to the technical field of vehicle safety distance control, comprising a monocular perception module, the monocular perception module is configured to obtain a real-time relative distance and a real-time braking force coefficient; the real-time relative distance is the distance between the front of the ego vehicle and the tail of the front vehicle, and the real-time braking force coefficient is used to represent the running state of the ego vehicle and the relative running state between the ego vehicle and the front vehicle; a braking calculation module, the braking calculation module is configured to generate a passive braking control instruction according to the real-time relative distance, the real-time braking force coefficient, time-varying distance data and a passive braking model; the passive braking control instruction comprises a time-varying target deceleration; a braking control module, the braking control module is configured to passively brake the ego vehicle according to the passive braking control instruction. The application solves the problems of low application universality and high cost of the existing vehicle passive braking technology through the above system.

Description

Technical Field

[0001] This invention relates to the field of vehicle safety distance control technology, and in particular to a passive safety distance control system based on monocular perception. Background Technology

[0002] As a core component of intelligent driving systems, the evolution of vehicle braking control technology and its braking force calculation schemes reflects the industry's continuous pursuit of driving comfort and safety. Early braking control schemes employed a simple control strategy known as "one-foot braking." While this approach was straightforward in implementation, it revealed numerous technical limitations in practical applications.

[0003] Traditional one-foot braking systems rely on a binary control approach based on a preset safe distance threshold. When the relative distance is detected to be less than the threshold, the system immediately applies a fixed braking force until the distance between vehicles returns to a safe range. This control method has two significant technical drawbacks: First, there are noticeable abrupt changes in braking force at both the start and end of the braking process, with the braking force suddenly increasing from zero to the preset value or suddenly decreasing from the preset value to zero. This discontinuous force change causes passengers to experience a noticeable jolt. Second, a fixed braking force is difficult to adapt to complex and changing real-world driving scenarios. It may result in unnecessary emergency braking due to excessive braking force, or excessive braking distance due to insufficient braking force.

[0004] To address the aforementioned issues, the industry has proposed an improved slope control scheme. This scheme introduces a linear increase in braking force during the initial braking phase and a linear decrease in braking force during the final braking phase, improving occupant comfort through this gradual change. However, this improved scheme introduces new technical challenges: the selection of the slope directly affects the quality of braking control. A fixed slope is difficult to adapt to braking demands under different vehicle speeds, loads, and road conditions, while the design of a dynamic slope lacks theoretical guidance and requires empirical adjustments based on extensive experimental data, resulting in low development efficiency and difficulty in ensuring consistent control performance.

[0005] Regarding adaptive braking force, existing technologies primarily employ empirical fitting methods based on experimental data. This involves collecting braking performance data under specific operating conditions to establish an empirical relationship between braking force and operating parameters, followed by interpolation calculations during actual control. While this method improves control adaptability to some extent, it is essentially a black-box modeling approach based on local data, exhibiting the following inherent drawbacks: First, the braking process involves complex coupling of multiple physical quantities, making it difficult for empirical models to comprehensively cover all influencing factors; second, the model has poor interpretability, making effective root cause analysis difficult when control anomalies occur; and finally, the model's generalization ability is limited, potentially leading to performance degradation under operating conditions exceeding the training data range. Summary of the Invention

[0006] This application provides a passive safety distance control system based on monocular perception to solve the problems of low application coverage and high implementation cost of existing vehicle passive braking technology.

[0007] The system includes:

[0008] A monocular perception module is configured to acquire real-time relative distance and real-time braking force coefficient; the real-time relative distance is the distance between the front of the vehicle and the rear of the vehicle in front, and the real-time braking force coefficient is used to characterize the driving state of the vehicle and the relative driving state between the vehicle and the vehicle in front.

[0009] A braking calculation module is configured to generate passive braking control commands based on the real-time relative vehicle distance, the real-time braking force coefficient, time-varying vehicle distance data, and a passive braking model; the passive braking control commands include a time-varying target deceleration; the time-varying vehicle distance data is acquired during the passive braking process of the vehicle.

[0010] A braking control module, configured to passively brake the vehicle according to the passive braking control command.

[0011] Preferably, the monocular sensing module is configured as follows:

[0012] Acquire a real-time vehicle distance image; the real-time vehicle distance image is an image of the area between the vehicle and the vehicle in front.

[0013] The real-time relative vehicle distance and the real-time braking force coefficient are obtained by calculating the vehicle distance based on the real-time vehicle distance image.

[0014] Preferably, the time-varying vehicle distance data includes time-varying inter-vehicle distance and braking time-varying inter-vehicle distance;

[0015] The time-varying inter-vehicle time distance is the change in distance between the vehicle and the vehicle in front per unit time while the vehicle maintains its original driving state;

[0016] The time-varying distance during braking refers to the change in distance between the vehicle and the vehicle in front per unit time during the passive braking process of the vehicle itself.

[0017] Preferably, the monocular sensing module includes:

[0018] An image acquisition unit, configured to acquire the real-time vehicle distance image;

[0019] The vehicle distance calculation unit is configured to perform vehicle distance calculation based on the real-time vehicle distance image to obtain the real-time relative vehicle distance and the real-time braking force coefficient.

[0020] Preferably, the vehicle distance calculation unit is further configured to:

[0021] Import the real-time vehicle distance image into preset coordinates;

[0022] The distance between the vehicle and the vehicle in front is determined in a preset coordinate system to obtain the real-time relative vehicle distance;

[0023] The real-time braking force coefficient is obtained by calculating the distance change based on the position changes of several real-time vehicle distance images in preset coordinates and the real-time vehicle speed.

[0024] Preferably, the system further includes:

[0025] The vehicle speed acquisition module is configured to acquire the real-time vehicle speed of the vehicle.

[0026] The monocular perception module is configured to calculate the vehicle distance based on the real-time vehicle speed and the real-time vehicle distance image, thereby obtaining the real-time relative vehicle distance and the real-time braking force coefficient.

[0027] Preferably, the braking calculation module is configured as follows:

[0028] Based on the real-time relative distance and the real-time braking force coefficient, determine whether the current distance meets the safe distance requirements;

[0029] If so, maintain the vehicle's original movement.

[0030] If not, then a passive braking control command is generated based on the real-time relative distance, the real-time braking force coefficient, the time-varying distance data, and the passive braking model.

[0031] Preferably, the braking calculation module is further configured to:

[0032] The time-varying inter-vehicle time distance is obtained by calculating based on the real-time relative vehicle distance;

[0033] If the real-time relative distance does not meet the safe distance requirement, the time-varying distance during braking is calculated based on the real-time relative distance; the safe distance requirement is a dynamic distance that varies according to the vehicle's speed.

[0034] Passive braking control commands are generated based on the real-time relative vehicle distance, the real-time braking force coefficient, the time-varying vehicle distance, the braking time-varying vehicle distance, and the passive braking model.

[0035] Preferably, the braking calculation module includes:

[0036] A safe distance judgment unit is configured to judge the safe distance based on the real-time relative vehicle distance and the real-time vehicle speed, and generate safe distance information or unsafe distance information.

[0037] A time-distance calculation unit is configured to calculate the time-varying inter-vehicle time distance based on the real-time relative vehicle distance.

[0038] The instruction generation unit is configured to generate an initial passive braking control instruction based on the unsafe distance information, the real-time relative distance, the time-varying inter-vehicle distance, the real-time braking force coefficient, the time-varying inter-vehicle distance, and the passive braking model; the initial passive braking control instruction is used to control the braking control module to perform initial passive braking on the vehicle.

[0039] Preferably, the time-distance calculation unit is further configured to calculate the time-distance between vehicles during braking based on the unsafe distance information and the real-time relative distance.

[0040] The instruction generation unit is further configured to generate a target passive braking control instruction based on the unsafe distance information, the real-time relative distance, the time-varying distance between vehicles, the time-varying distance during braking, the real-time braking force coefficient, the time-varying distance between vehicles, and the passive braking model; the target passive braking control instruction is used to control the braking control module to perform target passive braking on the vehicle until the real-time relative distance meets the safe distance requirement.

[0041] As described above, this application provides a passive safety distance control system based on monocular perception. The system includes a monocular perception module configured to acquire real-time relative distance and real-time braking force coefficient. The real-time relative distance is the distance between the front of the vehicle and the rear of the vehicle in front. The real-time braking force coefficient characterizes the vehicle's driving state and its relative driving state with respect to the vehicle in front. A braking calculation module is configured to generate passive braking control commands based on the real-time relative distance, the real-time braking force coefficient, time-varying distance data, and a passive braking model. The passive braking control commands include a time-varying target deceleration. The time-varying distance data is acquired during the passive braking process. A braking control module is configured to passively brake the vehicle according to the passive braking control commands. This application solves the problems of low application coverage and high implementation cost of existing vehicle passive braking technologies through the above system. Attached Figure Description

[0042] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of a passive safe distance control system based on monocular perception according to this application;

[0044] Figure 2 This is a schematic diagram of a monocular perception module in a passive safe distance control system based on monocular perception according to this application;

[0045] Figure 3 This is a schematic diagram of the braking calculation module in a passive safe distance control system based on monocular perception according to this application;

[0046] Figure 4 This is a graph showing the change in braking force when the first type of existing vehicle is passively braked.

[0047] Figure 5 This is a graph showing the change in braking force when the second type of existing vehicle is passively braked.

[0048] Figure 6 This is a graph showing the change in braking force when a vehicle using the solution described in this application undergoes passive braking.

[0049] Legend:

[0050] 100 - Monocular perception module; 200 - Braking calculation module; 300 - Braking control module; 400 - Vehicle speed acquisition module; 110 - Image acquisition unit; 120 - Vehicle distance calculation unit; 210 - Safe distance judgment unit; 220 - Time distance calculation unit; 230 - Command generation unit. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0053] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0054] When a vehicle is traveling on the road and the distance between it and the vehicle in front is less than the safe distance due to reasons such as inattention, dangerous driving, the vehicle in front slowing down, being overtaken, or being cut off, the vehicle is in a high-risk state, and the probability of a rear-end collision will increase significantly. This high-risk driving state not only poses a great risk to the driver of the vehicle, but also to other road users, and has a significant negative impact on overall traffic safety.

[0055] To reduce the traffic risks caused by insufficient safe following distance, with the current industry technology, it is entirely possible to actively control the brakes to decelerate through ADAS controllers, thereby increasing the relative distance between vehicles and reducing the traffic risks caused by the relative distance being less than the safe following distance, thus improving the safety of the traffic system and the safety of the vehicle itself.

[0056] If we want to control safe following distance through technical methods, then the following basic requirements must be met:

[0057] ADAS systems can detect vehicles ahead (they can know that there is a vehicle in front).

[0058] ADAS systems can identify lane lines and other road users (to determine operating conditions).

[0059] ADAS systems can detect the real-time distance between the vehicle and the vehicle in front;

[0060] ADAS systems can detect the relative speed between vehicles;

[0061] ADAS systems can acquire information such as the vehicle's speed, acceleration, angular acceleration, and vehicle parameters;

[0062] There are many technical solutions that can meet the above requirements, but each technical solution has its own advantages and disadvantages, as shown in Table 1.

[0063] Table 1

[0064]

[0065] From the perspective of technical solution selection, there are many solutions that can meet the requirements of safe vehicle distance. The reason we chose monocular perception is that its cost is low enough and the hardware product form is simple. The problem is that the perception results are not particularly accurate. For example, the ranging error is large (meter level), the relative speed error is large, and the lateral position and speed accuracy is insufficient.

[0066] As can be seen from ADAS products on the market, almost all ADAS products with braking functions rely on at least monocular + millimeter-wave radar for perception. This is limited by the industry level of monocular perception capabilities and the ability to develop ADAS braking functions based on monocular perception.

[0067] ADAS products are defined as driver assistance systems, which mainly intervene passively at critical moments rather than taking full control of the vehicle from the driver. Therefore, the passive safety distance keeping function has completely different requirements from active driver assistance technology, and requires demand analysis and technical solution design from different perspectives.

[0068] The methods for calculating vehicle braking force have undergone significant changes in the industry. Initially, it was a one-foot braking method: once the relative distance was less than the safe distance, a fixed braking force was applied until the relative distance met the requirement. The drawback of this method is that regardless of the amount of force applied, it always starts from 0, directly applying a value greater than zero; and at the end, it instantly drops from a fixed value greater than zero to 0, resulting in a very poor user experience. Figure 4 As shown.

[0069] It has poor adaptability; in some scenarios, the force feels too strong, and the braking is too abrupt; in other scenarios, the force feels too weak, and the brakes cannot be stopped. A fixed value is difficult to meet the needs of all scenarios.

[0070] Based on the above issues, subsequent revision plans emerged, with the basic idea being:

[0071] Regarding body sensation:

[0072] Add a diagonal line at the beginning and end of the braking process, such as... Figure 5 As shown.

[0073] This approach offers a straightforward optimization, but it also presents significant challenges. The key issue is determining the appropriate slope for braking force during uphill and downhill driving. Using a fixed slope will result in inconsistent performance under different conditions: the braking force may be applied too slowly, too quickly, or decrease too rapidly. Conversely, deviating from a fixed value requires designers to create dynamic climbing rates based on actual test results. This process is tedious, and its effectiveness is limited by the inherent limitations of the proposed solution.

[0074] Regarding force adaptability:

[0075] A common approach to optimizing force adaptation capability is to conduct on-road tests and, based on the actual braking effect, explore a set of ideal force values ​​under various working conditions. Then, fit these empirical values ​​to one-dimensional or multi-dimensional data and interpolate them according to the actual scenario to achieve the goal of force adaptation.

[0076] This optimization approach is very intuitive, and the technical route is simple and straightforward. However, the problem with this optimization scheme is also obvious: the relationship between braking force and body sensation is multivariate and nonlinear, making it difficult to traverse all variables. Data collection can only be carried out by fixing most variables in some scenarios. Therefore, its adaptive capability is also localized. Once the actual working conditions are different from the working conditions when the test data is collected, it is difficult to guarantee whether the effect of its adaptive capability is positive or negative.

[0077] In summary, the fundamental problem with this black-box algorithm is the lack of an interpretable mathematical model. Its generalization ability is insufficient, and the path to analyzing the problem is basically blocked. Ultimately, it can achieve limited adaptability in certain characteristic scenarios, despite the high cost of manpower and time.

[0078] Based on the above problems, this application provides the following embodiments.

[0079] Figure 1 This is a schematic diagram of a passive safe distance control system based on monocular perception according to this application.

[0080] See Figure 1 As can be seen, this embodiment provides a passive safe distance control system based on monocular perception, the system comprising:

[0081] A monocular perception module 100 is configured to acquire real-time relative distance and real-time braking force coefficient; the real-time relative distance is the distance between the front of the vehicle and the rear of the vehicle in front, and the real-time braking force coefficient is used to characterize the driving state of the vehicle and the relative driving state between the vehicle and the vehicle in front.

[0082] Specifically, conventional technologies for acquiring relative vehicle distance and relative vehicle speed often employ binocular perception technology, a combination of monocular perception and lidar, or a combination of binocular perception technology and lidar, in order to improve the accuracy and stability of relative vehicle speed and relative vehicle distance acquisition.

[0083] In this embodiment, only the monocular perception module 100 is used to acquire images. The purpose of this is to reduce the cost of image acquisition equipment. The corresponding acquisition accuracy is compensated by subsequent calculation algorithms, thereby reducing costs and improving the generalizability of the technology application while maintaining calculation accuracy and stability.

[0084] It should be noted that the monocular perception module 100 used in this embodiment is only one implementation method. Other combinations of technologies, such as binocular perception, binocular perception and lidar combination, and monocular perception and lidar combination, can also be used to achieve the same result.

[0085] The system also includes:

[0086] The braking calculation module 200 is configured to generate passive braking control commands based on the real-time relative distance, the real-time braking force coefficient, time-varying distance data, and the passive braking model; the passive braking control commands include time-varying target deceleration; the time-varying distance data is acquired during the passive braking process of the vehicle.

[0087] Specifically, in this embodiment, the vehicle distance is calculated by the braking calculation module 200. The main variable in the algorithm relied upon by the braking calculation module 200 is the relative vehicle distance, while the relative vehicle speed and acceleration are secondary variables. It can be understood that the calculation weight of the vehicle distance in the braking calculation module 200 is much greater than that of the speed variable, thereby offsetting the problem that the monocular perception module 100 has a low speed variable acquisition capability.

[0088] The passive braking calculation performed by the braking calculation module 200 can be understood as "anthropomorphism," which can be understood as:

[0089] In everyday motor vehicle driving, drivers cannot accurately obtain the distance between their vehicle and the vehicle in front, as well as their vehicle's precise speed, during braking. Furthermore, as the vehicle's speed increases, the safe braking distance must also continuously increase to ensure that the vehicle does not collide with the vehicle in front. During this braking process, the driver must rely on the time required to safely stop the vehicle within a certain distance from the vehicle in front to determine the required distance. This embodiment, based on this braking method, converts the distance from the spatial domain to the time domain and uses the time domain to measure the safe distance between the vehicle and the vehicle in front, thereby achieving passive braking.

[0090] For example, assuming the safe braking time of the vehicle using the system provided in this embodiment is 1 second, this safe braking time is the minimum braking time required for the vehicle to suddenly stop when the preceding vehicle is moving. When the safe braking time, represented by the relative distance between vehicles, is less than 1 second, passive braking is required to reduce the vehicle speed to a certain extent, and passive braking completely stops when the speed returns to 1 second. During passive braking, the braking force gradually increases from the beginning to a constant level and then gradually decreases, thereby achieving smooth passive braking. The change in braking force is described in [reference needed]. Figure 6 .

[0091] The time-varying vehicle distance data includes time-varying inter-vehicle distance and braking time-varying inter-vehicle distance; the time-varying inter-vehicle distance is the change in distance between the vehicle and the vehicle in front per unit time while the vehicle is maintaining its original driving state; the braking time-varying inter-vehicle distance is the change in distance between the vehicle and the vehicle in front per unit time during the passive braking process of the vehicle. Both the time-varying inter-vehicle distance and the braking time-varying inter-vehicle distance are calculated by the braking calculation module 200 based on the real-time relative vehicle distance and time.

[0092] The system also includes:

[0093] Braking control module 300, configured to passively brake the vehicle according to the passive braking control command.

[0094] Specifically, in this embodiment, the braking control module 300 is used to directly apply passive braking to the vehicle. The specific form of the module will not be described in detail in this embodiment.

[0095] The system also includes:

[0096] The vehicle speed acquisition module 400 is configured to acquire the real-time vehicle speed of the vehicle.

[0097] Specifically, in this embodiment, since the aforementioned "safe braking time" changes continuously with the vehicle speed, the vehicle speed is also an influencing factor in the calculation of the "safe braking time". Therefore, the vehicle speed acquisition module 400 is set to acquire the real-time vehicle speed.

[0098] Figure 2 This is a schematic diagram of a monocular perception module in a passive safe distance control system based on monocular perception according to this application.

[0099] See Figure 2 It can be seen that, further, in some embodiments, the monocular sensing module 100 includes:

[0100] Image acquisition unit 110, the image acquisition unit 110 being configured to acquire the real-time vehicle distance image;

[0101] The vehicle distance calculation unit 120 is configured to perform vehicle distance calculation based on the real-time vehicle distance image to obtain the real-time relative vehicle distance and the real-time braking force coefficient.

[0102] Specifically, in this embodiment, in order to improve generalization, the monocular perception module 100 only uses a conventional monocular perception device. In a conventional monocular perception device, there must be components for image acquisition and image processing. Therefore, the monocular perception module 100 is provided with the image acquisition unit 110 and the vehicle distance calculation unit 120 respectively. The image acquisition unit 110 acquires the actual image, and the vehicle distance calculation unit 120 calculates the vehicle distance and vehicle speed based on the image.

[0103] See Figure 3 It is understood that, further, in some embodiments, the braking calculation module 200 includes:

[0104] The safe distance judgment unit 210 is configured to judge the safe distance based on the real-time relative distance and the real-time vehicle speed, and generate safe distance information or unsafe distance information.

[0105] Specifically, in this embodiment, a safe distance needs to be determined before passive braking is performed. If the relative distance does not meet the safe distance requirement, then subsequent passive braking is performed. The safe distance determination unit 210 does not simply measure whether the safe distance is met by the distance between vehicles, but rather by combining the real-time relative distance and the real-time speed of the vehicle. This can be understood as the "safe braking time" mentioned above, which is used to measure whether the vehicle is currently at a safe distance.

[0106] The braking calculation module 200 also includes:

[0107] The time-distance calculation unit 220 is configured to calculate the time-varying inter-vehicle time distance based on the real-time relative vehicle distance.

[0108] Specifically, in this embodiment, the time distance calculation unit 220 performs the relevant calculations on vehicle distance and vehicle speed, thereby generating the changes in vehicle distance and vehicle speed within a unit of time.

[0109] The braking calculation module 200 also includes:

[0110] Instruction generation unit 230, the instruction generation unit 230 is configured to:

[0111] Based on the unsafe distance information, the real-time relative distance, the time-varying inter-vehicle distance, the real-time braking force coefficient, the time-varying inter-vehicle distance, and the passive braking model, an initial passive braking control command is generated; the initial passive braking control command is used to control the braking control module 300 to perform initial passive braking on the vehicle.

[0112] The unsafe distance information, the real-time relative distance, the time-varying distance between vehicles, the time-varying distance during braking, the real-time braking force coefficient, the time-varying distance between vehicles, and the passive braking model generate a target passive braking control command; the target passive braking control command is used to control the braking control module 300 to perform target passive braking on the vehicle until the real-time relative distance meets the safe distance requirement.

[0113] Specifically, in this embodiment, the instruction generation unit 230 generates instructions based on vehicle distance-related data. The vehicle distance-related data includes the time-varying vehicle distance during braking. The time-varying vehicle distance during braking is generated when passive braking has begun. Therefore, the time-varying vehicle distance during braking does not exist at the beginning of passive braking, or it can be defined as 0 and its value changes continuously as passive braking progresses.

[0114] This can be understood as the passive braking control command changing continuously during the passive braking process.

[0115] The specific implementation of this embodiment is as follows:

[0116] During the actual research and development process, the problems of the black-box solution were thoroughly analyzed, and a mathematical model was derived and verified to calculate the braking force under different working conditions. The specific mathematical model is as follows:

[0117] ;

[0118] Where A(t) is the time-varying target deceleration, with dimensions of m / s². 2 D(t): time-varying relative distance between vehicles, dimension: m; HT(t): time-varying inter-vehicle time distance, dimension: s; HT thr : The time distance between the vehicle and the engine at the start of braking, dimension: s (Note: Depending on the actual scenario, this variable can also be a time variable); a: The force control coefficient, with independent variables including relative velocity Δv(t) and vehicle speed V. ego (t), vehicle acceleration a ego (t), relative acceleration Δa(t), etc. This is just a typical model, and it can also include other variables / signals / information related to the specified force.

[0119] Using HT to represent the inter-vehicle distance transforms the distance from the spatial domain to the time domain. A key feature of this model is that when braking begins, the time-varying HT == HT. thrAt this point, the calculated value of A is almost zero. As time changes, the distance decreases, and HT naturally decreases. As a result, A gradually increases. The rate of change of A is mainly affected by the change in HT, that is, mainly by the change in relative distance. The faster the relative distance changes, the faster A increases, and vice versa. Thus, the initial braking sensation is guaranteed, and it is an adaptive sensation that does not need to rely on experience for setting.

[0120] Another characteristic of this model is that the braking termination process is the reverse of the braking initiation process. That is, due to braking, the vehicle speed decreases, the relative distance to other vehicles gradually approaches the safe distance, and A decreases until HT == HT. thr At this point, the braking force A = 0, and the braking process ends. This ensures that the braking process is also a time-varying, continuous process that returns to zero, without any sudden changes in braking force. Meanwhile, the change in relative distance is directly reflected in the rate of change of A, which is referred to as the speed of brake release in this embodiment. Thus, for the black-box model, the adaptive characteristics of this mathematical model are obvious, and the principle is clear.

[0121] At this point, the adaptive motion control of this model, compared to the black-box control model, has been completely solved. It can achieve adaptive force climbing and force recovery without the need for manual design of experience values ​​or the establishment of different numerical networks for interpolation. Instead, it is perfectly solved through an interpretable mathematical model.

[0122] The next issue is solving the problem of adaptive force control throughout the braking process. Compared to the black-box model, the third characteristic of this mathematical model is that its output force A is mainly affected by distance. During the entire braking process, when the relative distance is less than the safe distance, the smaller the relative distance, the larger A becomes. Here, the characteristics of a time-varying system must be considered, that is, the faster the relative distance changes, the faster A increases. This continues until the relative distance is less than or equal to the safe distance. Thus, this embodiment shows that, except for the initial and final force changes which are guaranteed to start from 0 and return to zero, the force change throughout the braking process is adaptive and mainly controlled by distance changes. At this point, the vehicle spacing becomes a closed-loop feedback variable of the model during the entire braking process, ultimately forming a distance-based closed-loop control force adaptive calculation model.

[0123] Next, this embodiment reviews the time-varying characteristics of braking force from a temporal perspective: Since the initial force starts from 0, the vehicle speed does not decrease significantly during the initial braking phase. Therefore, the relative distance decreases rapidly, and the difference between the relative distance and the safe distance becomes increasingly larger and less safe. At this time, A increases rapidly. As A increases, the vehicle speed begins to decrease, and the relative distance gets closer and closer to the safe distance (the safe distance is related to the vehicle speed; generally, the lower the vehicle speed, the smaller the safe distance, and vice versa). Then A begins to decrease again. Before A decreases, the braking force continues to increase until the relative distance between the two vehicles gets closer and closer to the safe distance, after which the braking process begins. The braking force change curve during the entire process is shown in [reference needed]. Figure 6 .

[0124] Currently, this embodiment has not discussed the role of the force coefficient 'a', but it has analyzed how the model simultaneously solves the basic logic of braking sensation and adaptive braking force in real time. The following embodiment will explain why the force coefficient 'a' is added to the model.

[0125] This embodiment reveals a problem with the mathematical model without the factor 'a': how to ensure that the safe distance is maintained within the safe vehicle distance range before a collision occurs? This is achieved through the factor 'a'. The force coefficient 'a' controls the magnitude of the braking force throughout the entire braking process. Its main function is to adjust the base force based on real-time operating conditions, ensuring that the relative distance remains within the safe range while avoiding a collision. Specifically:

[0126] Based on relative velocity and relative acceleration, information such as collision time and collision distance can be calculated. This information is then converted into force control coefficients. When the expected collision time is small, the force coefficient 'a' increases, thus increasing the force. When the expected collision time is large, the force coefficient 'a' decreases, ensuring that the force is not excessive.

[0127] The strength coefficient 'a' is an open sub-mathematical model, and its specific definition can be defined according to the actual application scenario, ultimately serving as a supplement to the main model (as shown in the following formula).

[0128] ;

[0129] Thus, the force coefficient 'a' ensures the effectiveness of the entire braking process and increases the flexibility of the model. However, the main model must play a primary role in order to guarantee the three features described above.

[0130] The above mathematical model has been simulated and tested, and it is consistent with the design goals. It can even be compatible with the characteristics of braking system delay and control deviation. At the same time, it does not have high requirements for the accuracy of relative speed and relative distance. Therefore, it can achieve safe distance maintenance based on monocular perception. When applying other perception models, it is still applicable when higher accuracy perception results can be obtained. It can even further expand and optimize the force control coefficient 'a'.

[0131] This embodiment has the following advantages:

[0132] The system displays the expressions, thus solving the problem of insufficient interpretability compared to the black-box model;

[0133] The interpretability of the system is the foundation of its scalability. Based on this, force calculation methods for other scenarios can be derived, and scenario-customized development can be supported.

[0134] The system's design is based on human drivers' reflections and analyses of their own driving behavior. The algorithm was developed through extensive analysis of the designer's own driving experience, followed by mathematical methods to describe and model braking decisions during driving. Therefore, the model possesses anthropomorphic characteristics.

[0135] The system employs a model with minimal mathematical computation and requires almost no external computing power, making it suitable for use in most embedded systems.

[0136] The system primarily relies on sensing and ranging results. As long as the single-mode sensing and ranging results are continuous and smooth, errors within a certain range are tolerable because the initial force is relatively small. Whether or not high-precision relative velocity and relative acceleration are required depends mainly on the configuration of external sensors; they are used if available, and not used if not. Therefore, this model also possesses good scalability.

[0137] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the discussion in some embodiments is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The selection and description of the above embodiments are for the purpose of better explaining the contents of this disclosure, thereby enabling those skilled in the art to better utilize the embodiments.

Claims

1. A passive safe distance control system based on monocular perception, characterized in that, The system includes: A monocular perception module (100) is configured to acquire real-time relative distance and real-time braking force coefficient; the real-time relative distance is the distance between the front of the vehicle and the rear of the vehicle in front, and the real-time braking force coefficient is used to characterize the driving state of the vehicle and the relative driving state between the vehicle and the vehicle in front. A braking calculation module (200) is configured to generate a passive braking control command based on the real-time relative distance, the real-time braking force coefficient, time-varying distance data, and a passive braking model; the passive braking control command includes a time-varying target deceleration; the time-varying distance data is acquired during the passive braking process of the vehicle. A braking control module (300) is configured to passively brake the vehicle according to the passive braking control command; The monocular sensing module (100) is configured as follows: Acquire a real-time vehicle distance image; the real-time vehicle distance image is an image of the area between the vehicle and the vehicle in front. The real-time relative vehicle distance and the real-time braking force coefficient are obtained by calculating the vehicle distance based on the real-time vehicle distance image. The time-varying vehicle distance data includes time-varying inter-vehicle distance and braking time-varying inter-vehicle distance; The time-varying inter-vehicle time distance is the change in distance between the vehicle and the vehicle in front per unit time while the vehicle maintains its original driving state; The time-varying distance during braking refers to the change in distance between the vehicle and the vehicle in front per unit time during the passive braking process of the vehicle itself. The monocular sensing module (100) includes: An image acquisition unit (110) is configured to acquire the real-time vehicle distance image; Vehicle distance calculation unit (120) is configured to perform vehicle distance calculation based on the real-time vehicle distance image to obtain the real-time relative vehicle distance and the real-time braking force coefficient; The vehicle distance calculation unit (120) is also configured to: Import the real-time vehicle distance image into preset coordinates; The distance between the vehicle and the vehicle in front is determined in a preset coordinate system to obtain the real-time relative vehicle distance; The real-time braking force coefficient is obtained by calculating the distance change based on the position changes of several real-time vehicle distance images in preset coordinates and the real-time vehicle speed.

2. The passive safe distance control system based on monocular perception according to claim 1, characterized in that, The system also includes: Vehicle speed acquisition module (400), the vehicle speed acquisition module (400) is configured to acquire the real-time vehicle speed of the vehicle.

3. A passive safe distance control system based on monocular perception according to claim 2, characterized in that, The braking calculation module (200) is configured as follows: Based on the real-time relative distance and the real-time braking force coefficient, determine whether the current distance meets the safe distance requirements; If so, maintain the vehicle's original movement. If not, then a passive braking control command is generated based on the real-time relative distance, the real-time braking force coefficient, the time-varying distance data, and the passive braking model.

4. A passive safe distance control system based on monocular perception according to claim 3, characterized in that, The braking calculation module (200) is also configured to: The time-varying inter-vehicle time distance is obtained by calculating based on the real-time relative vehicle distance; If the real-time relative distance does not meet the safe distance requirement, the time-varying distance during braking is calculated based on the real-time relative distance. The safe following distance requirement is a dynamic distance that changes according to the speed of the vehicle. Passive braking control commands are generated based on the real-time relative vehicle distance, the real-time braking force coefficient, the time-varying vehicle distance, the braking time-varying vehicle distance, and the passive braking model.

5. A passive safe distance control system based on monocular perception according to claim 4, characterized in that, The braking calculation module (200) includes: A safe distance judgment unit (210) is configured to judge the safe distance based on the real-time relative distance and the real-time vehicle speed, and generate safe distance information or unsafe distance information. A time-distance calculation unit (220) is configured to calculate the time-varying inter-vehicle time distance based on the real-time relative vehicle distance. The instruction generation unit (230) is configured to generate an initial passive braking control instruction based on the unsafe distance information, the real-time relative distance, the time-varying distance between vehicles during braking, the real-time braking force coefficient, the time-varying distance between vehicles, and the passive braking model; the initial passive braking control instruction is used to control the braking control module (300) to perform initial passive braking on the vehicle.

6. A passive safe distance control system based on monocular perception according to claim 5, characterized in that, The time-distance calculation unit (220) is also configured to calculate the braking time-distance based on the unsafe distance information and the real-time relative distance. The instruction generation unit (230) is further configured to generate a target passive braking control instruction based on the unsafe distance information, the real-time relative distance, the time-varying vehicle distance, the braking time-varying vehicle distance, the real-time braking force coefficient, and the passive braking model. The target passive braking control command is used to control the braking control module (300) to perform target passive braking on the vehicle until the real-time relative distance meets the safe distance requirement.

Images