An intelligent network connection platform control method and device based on vehicle-mounted communication, and a medium

Abstract

The application discloses a kind of intelligent network connection platform control method, equipment and medium based on vehicle communication, it is related to vehicle control technical field, including, based on safety margin parameter, communication quality index and control demand level, between the dynamic selection of network connection platform centralized main control mode, car-cloud collaborative control mode and vehicle controller full control mode, the vehicle control algorithm parameter corresponding to the selected control mode is adjusted in real time, and vehicle control problem is modeled as a rolling horizon optimization problem, while adding safety margin parameter as the boundary condition of optimization problem, by solving optimization problem, obtain final control instruction, based on final control instruction, select corresponding communication path and issue to vehicle controller for execution;The application generates final control instruction to improve the running safety and stability of intelligent network connection vehicle in complex traffic environment.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to a control method, device and medium for an intelligent connected platform based on vehicle communication. Background Technology

[0002] With the continuous evolution of intelligent transportation, vehicle-to-everything (V2X) communication technology is profoundly driving the transformation of intelligent connected vehicles from single vehicle control to a complex system paradigm of multi-entity collaboration and cloud-edge-device integration. Against this backdrop, intelligent connected platforms based on V2X communication, as the core hub for achieving efficient information interaction between vehicles and infrastructure, vehicles and vehicles, vehicles and the cloud, and vehicles and pedestrians, have become key enabling technologies supporting advanced applications such as autonomous driving, collaborative perception, platooning, and dynamic path optimization. Currently, high-speed, low-latency, and highly reliable communication technologies, represented by 5G-V2X, C-V2X, and the future 6G vehicle-to-everything (V2X) network, have gradually built a wide-area, responsive vehicle-road cooperative network foundation. Relying on this communication architecture, intelligent connected platforms have achieved real-time perception of the overall traffic status and dynamic modeling of the global situation. This enables vehicles not only to acquire their own local environmental information but also to accurately acquire non-locally available information such as macro-level traffic flow, accident warnings, and traffic light timing optimization by integrating massive amounts of historical and real-time data gathered in the cloud.

[0003] Nevertheless, existing vehicle control methods still have room for improvement. First, most control strategies treat the communication link state as a static input or only use historical averages for characterization, failing to fully exploit the probabilistic distribution characteristics of communication parameters. This leads to insufficient estimation of communication delays and command failure risks in the worst-case scenario. Second, there is a general lack of a closed-loop mechanism that directly transforms communication uncertainties into boundary conditions for adjusting and optimizing control algorithm parameters. This results in the use of performance-prioritized control parameters even under high-risk communication conditions, exacerbating trajectory deviations and potential collision risks caused by command delays or loss. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a control method for an intelligent connected platform based on vehicle communication to solve the problems of insufficient estimation of communication delay and command failure risk in the worst case, as well as trajectory deviation and potential collision risk caused by command delay or loss.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a control method for an intelligent connected platform based on vehicle-mounted communication, comprising,

[0008] Collect vehicle status information, surrounding environment perception information, communication link status information, and control command information from the connected platform;

[0009] Based on the communication link status information, calculate the communication quality index that characterizes the overall quality of the current primary communication link;

[0010] Based on the vehicle's physical status information, surrounding environment perception information, and control command information, determine the level of control requirements currently needed by the vehicle.

[0011] Based on communication link status information, the probability distribution of end-to-end delay and packet loss rate in the rolling time domain is predicted, and based on vehicle status information, control requirement level and surrounding environment perception information, the safety margin parameters that meet the preset safety risk limit are calculated.

[0012] Based on safety margin parameters, communication quality index and control requirement level, dynamic selection is made between the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode and the full control mode of the vehicle controller.

[0013] The vehicle control algorithm parameters corresponding to the selected control mode are adjusted in real time, and the vehicle control problem is modeled as a rolling time domain optimization problem. At the same time, a safety margin parameter is added as the boundary condition of the optimization problem. The final control command is obtained by solving the rolling time domain optimization problem.

[0014] Based on the final control command, the appropriate communication path is selected and sent to the vehicle controller for execution.

[0015] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, the vehicle body state information includes vehicle speed, acceleration, and steering wheel angle;

[0016] The surrounding environment perception information includes obstacle distance and pedestrian target status;

[0017] The communication link status information includes historical end-to-end latency, historical latency jitter data, and historical packet loss rate;

[0018] The control command information from the connected platform includes global path planning commands and cooperative driving suggestions.

[0019] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, the calculation of the communication quality index, which characterizes the overall quality of the current primary communication link, refers to normalizing each parameter in the communication link status information, and then weighting and summing all the normalized communication link status information parameters to obtain the communication quality index.

[0020] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, wherein: determining the current control requirement level of the vehicle specifically involves:

[0021] Based on vehicle status information, surrounding environment perception information, and control command information, the driving scenario and task objectives of the vehicle are comprehensively analyzed.

[0022] For each driving scenario and task objective, a control requirement level is assigned.

[0023] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, the calculation of the safety margin parameter that satisfies the preset safety risk upper limit specifically includes:

[0024] Based on communication link status information, time windows are divided according to the rolling time domain. Kernel density estimation is used to calculate the probability distribution of historical end-to-end delay within each time window. The Beta distribution is used to fit parameters to the historical packet loss rate to obtain the probability distribution of packet loss rate.

[0025] Based on the probability distribution of end-to-end delay and the probability distribution of packet loss rate, calculate the upper bound of the deviation caused by the uncertainty of end-to-end delay and the uncertainty of packet loss rate on the effective time of control command information;

[0026] Based on the upper bound of the deviation and the vehicle speed in the vehicle's physical state information, the corresponding range of vehicle position uncertainty is calculated, and a safety risk upper limit threshold is preset.

[0027] Based on the range of location uncertainty and the distance to obstacles in the surrounding environment perception information, calculate the safety margin parameter that meets the upper limit threshold of safety risk.

[0028] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, the dynamic selection among the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller specifically involves:

[0029] Calculate the communication uncertainty risk value, insufficient control capability risk value, and insufficient safety margin risk value for the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller under the current communication quality index and control requirement level.

[0030] The communication uncertainty risk value, control capability insufficiency risk value, and safety margin insufficiency risk value of each control mode are weighted and fused to obtain the comprehensive control risk value corresponding to each control mode.

[0031] The control mode with the lowest overall control risk value is selected as the control mode for the current moment.

[0032] In a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, obtaining the final control command specifically involves:

[0033] The vehicle control algorithm parameters include the lengths of the prediction and control time domains of the model predictive controller, the feedforward and feedback gains of the trajectory tracking controller, and the noise covariance matrix of the state estimator.

[0034] Based on the selected control mode and safety margin parameters, adjust the lengths of the prediction time domain and control time domain of the model predictive controller, adjust the feedforward gain and feedback gain of the trajectory tracking controller, and adjust the noise covariance matrix of the state estimator.

[0035] Based on the adjusted vehicle control algorithm parameters, the vehicle control problem is modeled as a rolling time-domain optimization problem. The objective function of the rolling time-domain optimization problem is designed to minimize the weighted sum of tracking error, control energy consumption, and ride discomfort.

[0036] In the constraint set of the rolling time-domain optimization problem, boundary conditions defined by the safety margin parameter are added;

[0037] The solver solves the rolling time-domain optimization problem with boundary conditions to obtain the optimal control input sequence in the next time domain. The first control variable in the optimal control input sequence is then extracted as the final control command.

[0038] As a preferred embodiment of the intelligent connected platform control method based on vehicle communication described in this invention, the step of selecting the corresponding communication path to send the final control command to the vehicle controller for execution means encapsulating the final control command into a data packet and selecting a communication path to send it to the vehicle controller. After receiving the data packet, the vehicle controller instructs the steering, driving and braking actuators to complete the action.

[0039] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the intelligent connected platform control method based on vehicle communication as described in the first aspect of the present invention.

[0040] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent connected platform control method based on vehicle communication as described in the first aspect of the present invention.

[0041] The beneficial effects of this invention are as follows: Based on the joint modeling method of kernel density estimation and Bayesian update, communication uncertainty is transformed into a quantifiable high-confidence safety margin boundary, breaking through the limitations of the previous static safety distance. A risk assessment mechanism integrating three elements is introduced to realize intelligent dynamic switching of control modes between centralized, distributed and local full authority, ensuring that the optimal control strategy is always matched under different communication conditions. Finally, the safety margin parameter is used as a rigid constraint condition for the rolling time domain optimization problem, so that all control commands are limited to the safe and feasible domain from the source, truly achieving the synergistic goal of "safety first, performance self-adaptation", and improving the operational safety and stability of intelligent connected vehicles in complex traffic environments. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a control method for an intelligent connected platform based on vehicle communication.

[0044] Figure 2 A flowchart for obtaining safety margin parameters.

[0045] Figure 3 This is a flowchart for calculating the communication quality index.

[0046] Figure 4 A flowchart for obtaining the final control command. Detailed Implementation

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0048] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0049] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0050] Reference Figures 1-4 This is one embodiment of the present invention, which provides a control method for an intelligent connected platform based on vehicle communication, comprising the following steps:

[0051] S1. Collect vehicle body status information, surrounding environment perception information, communication link status information, and control command information from the connected platform.

[0052] It should be noted that, through the vehicle controller's local area network bus interface, vehicle speed, longitudinal acceleration, and steering wheel angle are periodically (e.g., every 10ms to 20ms) read from the vehicle powertrain controller, electronic stability controller, and steering controller to constitute the vehicle's status information. The relative distance between the vehicle and the nearest obstacle is measured using forward-facing millimeter-wave radar as the obstacle distance; the pedestrian target status is constructed by outputting a pedestrian bounding box and velocity vector through the vehicle's vision recognition unit; the obstacle distance and pedestrian target status together constitute the surrounding environment perception information. In the cellular communication module's protocol stack, a reception timestamp is recorded for each received application layer acknowledgment data packet. By comparing the sending and receiving timestamps, the historical end-to-end latency is obtained. The standard deviation of consecutive historical end-to-end latency is calculated as historical latency jitter data, and the proportion of unacknowledged data packets is calculated as the historical packet loss rate. The historical end-to-end latency, historical latency jitter data, and historical packet loss rate together constitute the communication link status information. The system receives a data stream encoded in the Protobuf protocol format from the network platform server via a TCP socket connection. By decoding the data stream, it obtains global path planning instructions with path point coordinates, as well as cooperative driving suggestions containing desired vehicle spacing and cooperative speed, which are control instruction information from the network platform.

[0053] Furthermore, the sampling period of 10ms to 20ms keeps the state update frequency consistent with the vehicle chassis control cycle, ensuring real-time control while avoiding excessive computational and communication load, thereby improving the stability and safety of rolling time-domain optimization control.

[0054] S2. Based on the communication link status information, calculate the communication quality index, which characterizes the overall quality of the current primary communication link.

[0055] It should be noted that the min-max normalization method is used to normalize the historical end-to-end delay, historical delay jitter data, and historical packet loss rate in the communication link status information. A preset weight coefficient is assigned to each of the normalized historical end-to-end delay, historical delay jitter data, and historical packet loss rate, and they are multiplied. The three product results are summed to obtain a comprehensive value, namely the communication quality index.

[0056] Furthermore, the weighting coefficients are set to differentiate the degree of impact of each parameter in the communication link status information on vehicle control safety. The historical packet loss rate directly determines whether the control command information can successfully reach the vehicle controller. Once packet loss occurs, the control command information will not take effect or the control behavior will be delayed, which has the greatest impact on vehicle control safety. Therefore, the historical packet loss rate is assigned the highest weight (e.g., 0.4–0.6). The historical end-to-end delay determines the degree of lag in the control command and has a significant impact on vehicle trajectory deviation. Therefore, it is assigned the second highest weight (e.g., 0.25–0.4). The historical delay jitter data mainly affects the stability of the control cycle and has a relatively small direct impact on safety. Therefore, it is assigned the lowest weight (e.g., 0.1–0.3). The weighting coefficient of the historical packet loss rate is greater than that of the historical end-to-end delay, which is greater than that of the historical delay jitter data. The sum of all weighting coefficients is 1.

[0057] S3. Based on the vehicle's status information, surrounding environment perception information, and control command information, determine the current control requirement level of the vehicle.

[0058] S3.1 It should be noted that, based on the vehicle's speed, acceleration, and steering wheel angle in the vehicle's own state information, the system determines whether the vehicle is currently cruising, accelerating, braking, or turning. Combining obstacle distances and pedestrian target states from the surrounding environment perception information, the system identifies the local traffic environment in which the vehicle is located, including whether there is a close-range vehicle ahead, whether there is a risk of a vehicle cutting in from the side, whether pedestrians are in a conflict zone, and whether the vehicle is deviating from the center of the lane. Simultaneously, the system analyzes the global path planning instructions from the control command information from the connected vehicle platform to clarify the macro-level driving intention for the next stage, such as "exit the highway at the next exit" or "turn left at the intersection ahead"; for cooperative driving suggestions, it clarifies whether there are specific collaborative requirements such as platooning or cooperative passage through intersections. Based on the results of the above judgments, identifications, and analyses, the vehicle's current driving scenario and task objective are classified into one of the following situations: highway cruising, following another vehicle in congested traffic, autonomous lane changing, intersection passage, roundabout passage, emergency obstacle avoidance, remote shuttle, or automatic parking.

[0059] S3.2 It should be noted that, for each identified driving scenario and task objective, a control requirement level is assigned according to preset allocation rules (the control requirement level is defined as highest level, high level, medium level, and low level). The allocation rules are pre-set based on a comprehensive analysis of the functional safety risks of the driving scenario and task objective, the latency sensitivity of control commands, the decision complexity of control actions, and the degree of dependence on the globality of collaborative information. For example, the driving scenario "emergency obstacle avoidance" is assigned the highest level due to the urgent risk of collision and extremely stringent requirements for latency and reliability. The driving scenarios "pedestrian avoidance," "intersection crossing," and "roundabout crossing" are assigned the high level due to the dynamic complexity and strong interactivity of the environment and high requirements for real-time control and decision accuracy. The driving scenarios "congested road following" and "autonomous lane changing" occur in a structured environment, allowing for moderate reaction time and control margin, and are assigned the medium level. The driving scenarios "highway cruise", "maintaining a specific distance in convoy according to fleet instructions", "remote shuttle" and "automatic parking" have the lowest requirements for real-time control due to their highly structured nature, slow state changes or offline planning.

[0060] S4. Based on communication link status information, predict the probability distribution of end-to-end delay and packet loss rate in the rolling time domain, and calculate the safety margin parameters that meet the preset safety risk upper limit according to vehicle status information, control requirement level and surrounding environment perception information.

[0061] S4.1 It should be noted that, based on the communication link status information, a rolling time domain of a preset duration (e.g., 1 second) is defined backward from the current time as a time window. Within the time window, the historical end-to-end delays of all records are extracted to form a delay sample sequence. The probability density function value of the delay sample sequence over the continuous value domain is calculated using a Gaussian kernel function. The probability density function value is the probability distribution of the end-to-end delay within the rolling time domain, and the formula is:

[0062] ;

[0063] in, The probability density function value represents the end-to-end delay. This indicates the number of samples in the time-delayed sample sequence. Indicates bandwidth parameter, Represents the Gaussian kernel function. The time delay independent variable is the end-to-end time delay of the probability density to be evaluated. Represents the first time delay sample in the sequence. The time delay observation value of each observation sample.

[0064] The bandwidth parameter can be adaptively determined according to Scott's rule, expressed by the formula:

[0065] ;

[0066] in, Indicates bandwidth parameter, The standard deviation of the time-delayed sample sequence is represented by the sample standard deviation. Indicates the number of samples in the time-delayed sample sequence;

[0067] Furthermore, constants This is a theoretically derived constant in Scott's rule, based on rigorous mathematical optimization. Its purpose is to minimize the asymptotic integral mean square error of the Gaussian kernel density estimate under the assumption that the data follows a normal distribution. The negative fifth power of the number of samples in the time-delayed sample sequence represents the scaling relationship between the theoretically optimal bandwidth and the number of samples, describing how the bandwidth should decrease as the sample size increases.

[0068] S4.2 It should be noted that within the same time window (1 second), the total number of packets sent by the network platform and the number of data packets successfully acknowledged by the vehicle terminal are counted to calculate the historical packet loss rate. The Beta distribution is considered as a prior distribution describing the unknown probability of packet loss rate, and its shape is determined by two parameters α and β. Initially, no-information prior parameters can be set for the Beta distribution, for example, α=1 and β=1. Within the time window, the difference between the observed total number of packets sent and the number of successfully acknowledged data packets is calculated to obtain the number of lost packets. Utilizing the property that the Beta distribution is the conjugate prior of the Bernoulli distribution, the prior parameters are updated. The update rule is: sum the prior parameter α with the number of successfully acknowledged data packets to obtain the posterior parameter α'; sum the prior parameter β with the observed number of lost packets to obtain the posterior parameter β'. The updated posterior parameters α' and β' are the new Beta distribution. The new Beta distribution is the probability distribution of the packet loss rate in the rolling time domain.

[0069] Furthermore, parameters α and β are shape parameters of the distribution, controlling the shape of the Beta distribution and reflecting the distribution of the packet loss rate. Specifically, α represents the number of successful events (such as successfully acknowledged data packets), and β represents the number of failed events (such as packet loss).

[0070] S4.3. It should be noted that a high confidence level, such as 99%, is set. On the probability distribution of end-to-end delay, the right-tailed quantile corresponding to the confidence level is calculated. The right-tailed quantile represents the upper bound of the pessimistic estimate of the control command information transmission delay when considering delay uncertainty, and is denoted as the upper bound component of the delay deviation. On the probability distribution of packet loss rate, the right-tailed quantile corresponding to the confidence level is calculated again, representing the pessimistic estimate of the packet loss rate. Based on the pessimistic estimate of the packet loss rate, combined with the retransmission timeout and maximum retransmission count of the transport layer protocol, the maximum possible time required for a control command message to be successfully received from its initial transmission is calculated under the worst-case packet loss scenario. This maximum possible time is marked as the upper bound component of the delay caused by packet loss. The formula for calculating the maximum possible time required for the control command information to be successfully received from its initial transmission is:

[0071] ;

[0072] in, This indicates the maximum possible time required for a control command to be successfully received from its initial transmission. This indicates a retransmission timeout (as defined by the transport layer protocol). This represents a pessimistic estimate of the packet loss rate. Indicates the maximum number of retransmissions (as specified by the transport layer protocol).

[0073] Transport layer protocols are the fundamental communication protocols responsible for providing end-to-end reliability assurance for the transmission of control command information between the connected platform and the vehicle controller.

[0074] The sum of the upper bound component of the delay deviation and the upper bound component of the delay caused by packet loss is the upper bound of the deviation caused by the combined effect of end-to-end delay uncertainty and packet loss rate uncertainty on the effective time of control command information.

[0075] Furthermore, setting the high confidence level to 99% is to transform the probability distribution of communication uncertainty into a deterministic worst-case time boundary that can be used for vehicle safety control design. At the same time, it can provide a statistically significant and quantifiable worst-case time boundary for the safety design of vehicle control, thereby enabling pre-compensation for communication uncertainty at the algorithm level and improving the functional safety of intelligent connected vehicles in unreliable communication environments.

[0076] S4.4. It should be noted that the current vehicle speed is read from the vehicle's physical state information. Assume that the vehicle maintains its current speed at a constant velocity throughout the entire time window corresponding to the upper bound of the deviation. Multiplying the maximum time deviation represented by the upper bound of the deviation by the current vehicle speed yields the maximum additional displacement the vehicle can generate relative to its ideal, unbiased expected position within the maximum time deviation. This calculated maximum additional displacement is quantified as the range of vehicle position uncertainty caused by communication uncertainties. The ideal, unbiased expected position refers to the expected trajectory point that the vehicle should reach within the current control cycle under ideal communication conditions, without considering communication latency and packet loss, based on the expected trajectory corresponding to the global path planning instructions from the connected platform.

[0077] The system reads the obstacle distance between the current vehicle and the nearest obstacle ahead from surrounding environment perception information. A preset safety risk upper limit threshold is established, representing the ratio of the minimum safe distance stipulated by traffic safety standards to the obstacle distance. The difference between the obstacle distance and the range of positional uncertainty is calculated to obtain the effective safe distance under the influence of communication uncertainty. The safety risk upper limit threshold and the obstacle distance are multiplied to obtain the theoretical minimum required safe distance. The ratio of the effective safe distance to the minimum required safe distance is calculated to obtain the safety margin parameter that satisfies the safety risk upper limit threshold. The safety margin parameter is a dimensionless safety margin coefficient, theoretically ranging from 0 to positive infinity, but typically falling between 0.5 and 2.0 in actual vehicle driving scenarios. A safety margin parameter greater than 1 indicates sufficient safety distance, equal to 1 indicates just meeting safety requirements, and less than 1 indicates insufficient safety distance.

[0078] The upper limit threshold for safety risks is set based on the degree of collision risk that may be caused by vehicle position deviation due to communication uncertainty. It is used to ensure that the vehicle can still maintain an operating state of not less than the minimum safe distance under the worst communication conditions. The ratio of the minimum safe distance (which varies with vehicle speed) specified by traffic safety standards to the distance to the obstacle is calculated, and the result of the ratio calculation is used as the upper limit threshold for safety risks.

[0079] S5. Based on safety margin parameters, communication quality index, and control requirement level, it dynamically selects between the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller.

[0080] S5.1 It should be noted that the preset mode dependency weighting coefficients associated with each control mode will (1 The communication quality index is multiplied by the mode dependence weight coefficient of the corresponding control mode to obtain the communication uncertainty risk value of each control mode.

[0081] Furthermore, the mode dependency weight coefficients for each control mode can be allocated according to the degree of dependence of the control decision location on communication. The principle is: the more a control mode relies on the control command information issued by the network platform, the greater the control risk caused by communication anomalies, and the larger the mode dependency weight coefficient, such as 0.9; the more a control mode relies on the local closed loop of the vehicle controller, the smaller the mode dependency weight coefficient, such as 0.2; the vehicle-cloud collaborative control mode involves the joint participation of the network platform and the vehicle controller in control decision-making, and the mode dependency weight coefficient is set to medium, such as 0.6.

[0082] S5.2 It should be noted that, based on the theoretical control accuracy required by the control demand level and the typical control accuracy that each control mode can provide, the risk value of insufficient control capability for each control mode is calculated. For example, the theoretical lateral control accuracy required for the "high" control demand level is 0.1 meters, the typical control accuracy of the centralized master control mode of the connected platform is 0.05 meters, the vehicle-cloud collaborative control mode is 0.08 meters, and the on-board controller full-authority control mode is 0.15 meters. Therefore, the risk value of insufficient control capability for each control mode can be calculated as follows:

[0083] ;

[0084] in, This indicates that the control mode's control capability is insufficient, which poses a risk value. This indicates the required control precision at the control demand level. This indicates the typical control accuracy of the control mode.

[0085] S5.3 It should be noted that a safety margin expectation threshold is preset for each of the network platform centralized master control mode, vehicle-cloud collaborative control mode, and vehicle controller full-authority control mode; the difference between the safety margin expectation threshold and the safety margin parameter is calculated to obtain the safety margin insufficient risk value (0 is taken when the difference is less than 0). For the network platform centralized master control mode, vehicle-cloud collaborative control mode, and vehicle controller full-authority control mode, the calculation process for communication uncertainty risk value, control capability insufficient risk value, and safety margin insufficient risk value is executed independently.

[0086] Furthermore, the expected safety margin thresholds for each control mode are set based on the degree of dependence of each control mode on the communication link and the sensitivity of communication uncertainty to the effective time of control command information. For example, the centralized master control mode of the connected platform mainly relies on the remote connected platform to generate control command information, and is most sensitive to communication latency and packet loss. Therefore, a larger expected safety margin threshold is required to offset the position deviation caused by communication uncertainty (e.g., 1.50). The vehicle-cloud collaborative control mode involves the joint participation of the vehicle controller and the connected platform in control decision-making, and has a secondary degree of dependence on communication. Therefore, the corresponding expected safety margin threshold is set at a medium level (e.g., 1.25). The full-authority control mode of the vehicle controller mainly relies on local perception and local control, and has the lowest degree of dependence on the communication link. Therefore, the corresponding expected safety margin threshold is set at the lowest level (e.g., 1.10).

[0087] S5.4 It should be noted that for the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller, the corresponding communication uncertainty risk value, control capability insufficiency risk value, and safety margin insufficiency risk value are obtained respectively. Preset risk fusion weight coefficients are assigned to each of these risk values. The communication uncertainty risk value, control capability insufficiency risk value, and safety margin insufficiency risk value for each control mode are multiplied by their corresponding risk fusion weight coefficients and then summed to obtain the comprehensive control risk value for each control mode. The control mode with the lowest comprehensive control risk value is selected as the control mode at the current moment.

[0088] Furthermore, the insufficient safety margin risk value directly reflects whether the safe distance between the vehicle and the obstacle meets the requirements, and has the most direct impact on collision risk. Therefore, it has the largest risk fusion weight coefficient, such as 0.6. The communication uncertainty risk value indirectly affects the safe distance by influencing the timeliness and reliability of control command information, so its risk fusion weight coefficient is the second largest, such as 0.25. The insufficient control capability risk value mainly affects the trajectory tracking accuracy, and its effect on collision risk is a secondary indirect impact. Therefore, its risk fusion weight coefficient is the smallest, such as 0.15. The sum of the risk fusion weight coefficients is 1.

[0089] S6. Adjust the vehicle control algorithm parameters corresponding to the selected control mode in real time, and model the vehicle control problem as a rolling time-domain optimization problem. At the same time, add a safety margin parameter as the boundary condition of the optimization problem, and obtain the final control command by solving the optimization problem.

[0090] S6.1 It should be noted that, based on the selected control mode and safety margin parameter, when the safety margin parameter is greater than 1.2, it indicates that there is sufficient safety margin, which can improve control performance. For example, the prediction time domain and control time domain length of the model predictive controller can be extended to 120% of the original length for better look-ahead planning; the feedforward gain and feedback gain of the trajectory tracking controller can be increased (e.g., 10%-20%) to improve the accuracy and response speed of trajectory tracking; the process noise covariance matrix and observation noise covariance matrix of the state estimator can be reduced (e.g., the diagonal elements of the process noise covariance matrix and observation noise covariance matrix can be reduced by 10%) to improve observation confidence.

[0091] When the safety margin parameter is between 0.8 and 1.2, it indicates that the safety status is at a critical or normal level. The prediction time domain length and control time domain length of the model predictive controller are kept at their original lengths. The feedforward gain of the trajectory tracking controller is reduced (e.g., by 5%), the feedback gain of the trajectory tracking controller is kept at its original value, and the noise covariance matrix of the state estimator is maintained at a preset baseline value.

[0092] When the safety margin parameter is less than 0.8, it indicates a tight safety boundary and potential risk. In this case, the primary control objective is to ensure stability and safety. Shortening the prediction and control time domains of the model predictive controller (e.g., to 80% of their original length) reduces reliance on uncertain futures and focuses on real-time control. Reducing the feedforward and feedback gains of the trajectory tracking controller (e.g., by 10%-20%) lowers system bandwidth and prevents oscillations caused by errors or noise. Increasing the process noise and observation noise covariance of the state estimator reflects higher environmental uncertainties and perception errors, making the filter smoother (e.g., increasing the diagonal elements of the process noise covariance matrix by 20% and the diagonal elements of the state estimator's observation noise covariance matrix by 10%).

[0093] Furthermore, the safety margin parameter range division and corresponding controller parameter adjustment rules are set based on the characterization relationship between the safety margin parameter and the vehicle's safe distance redundancy. This ensures that scenarios with smaller safety margin parameters correspond to more conservative controller parameter configurations, while scenarios with larger safety margin parameters correspond to more performance-oriented controller parameter configurations. The numerical adjustment of controller parameters (such as gain, time domain length, noise covariance matrix, etc.) is not static, but is flexibly adjusted according to specific communication conditions, vehicle status, and environmental information. The aim is to achieve the optimal balance between performance and safety without affecting vehicle stability.

[0094] Each control mode adjusts its parameters based on the same safety margin parameter. However, since the reference controller parameters for different control modes are different, the controller parameter adjustment results obtained by each control mode are not the same under the same safety margin parameter conditions.

[0095] S6.2 It should be noted that the adjusted prediction time domain length of the model predictive controller is used as the prediction time domain of the optimization problem, and the adjusted control time domain length of the model predictive controller is used as the control time domain of the optimization problem. The objective function of the rolling time domain optimization problem is constructed as a weighted quadratic function, which contains the sum of three core costs: the first term is the sum of squares of the deviations between the vehicle's predicted trajectory and the expected trajectory provided by the connected platform, used to minimize the tracking error; the second term is the sum of squares of the control input vectors (including steering wheel angle and longitudinal acceleration), used to penalize excessive control energy consumption; the third term is the quadratic cost of the rate of change of control commands to characterize ride discomfort; the three costs are multiplied by their corresponding weight coefficients and then summed to obtain the objective function of the rolling time domain optimization problem.

[0096] Furthermore, the tracking error cost directly determines whether the vehicle can stay within the safe driving trajectory, and has the most direct impact on safety. Therefore, its corresponding objective function weight coefficient is the largest (e.g., 0.6). The energy consumption control cost mainly constrains the actuator output amplitude, affecting the actuator load and energy consumption level. Its impact on safety is secondary, so its corresponding objective function weight coefficient is moderate (e.g., 0.3). The ride discomfort cost only affects the ride experience and has the least direct impact on safety. Therefore, its corresponding objective function weight coefficient is the smallest (e.g., 0.1).

[0097] S6.3. It should be noted that the real-time distance between the vehicle and surrounding obstacles, and the lateral distance between the vehicle and the lane boundary are extracted from the surrounding environment perception information. These real-time distances and lateral distances are used as the basic safety boundaries. The basic safety boundaries are then compared with the safety margin parameter to obtain a scaled, more stringent conservative safety boundary. In the state constraints of the rolling time-domain optimization problem, a constraint requiring the predicted position of the vehicle throughout the entire prediction time domain to always be greater than or equal to the conservative safety boundary is added. In the control input constraints of the rolling time-domain optimization problem, the physical limits of the steering wheel angle and acceleration are multiplied by the reciprocal of the safety margin parameter to form a more stringent control action boundary, which is then added to the control input constraints. After constraint processing, the safety margin parameter is directly transformed into a hard constraint on the vehicle's future motion state space and control action space, constituting the boundary conditions defined by the safety margin parameter in the rolling time-domain optimization problem.

[0098] Furthermore, the physical limits of vehicle steering wheel angle and acceleration are constants that are directly obtained from vehicle platform characteristic calibration and remain essentially unchanged throughout the vehicle's lifespan.

[0099] S6.4 It should be noted that the rolling time-domain optimization problem, with boundary conditions defined by safety margin parameters, is input to the numerical optimization solver. Under the premise of satisfying all constraints, the numerical optimization solver performs iterative calculations based on minimizing the weighted quadratic objective function. This iterative process is repeated until the change in the objective function value is less than a preset convergence threshold, and the degree of violation of all constraints is less than a preset tolerance. At this point, the iterative calculation is considered converged. After convergence, the last set of control input sequences output by the numerical optimization solver is a series of optimal control input values ​​in the future control time domain that minimize the weighted quadratic objective function under all constraints. This series of optimal control input values ​​constitutes the optimal control input sequence. From the optimal control input sequence, the first control quantity corresponding to the current control cycle is extracted, which is the final control command. The final control command includes the steering wheel angle command and the longitudinal acceleration command.

[0100] Furthermore, the convergence threshold is set to ensure that the numerical optimization solver obtains sufficiently accurate control input results within a finite computation time. It is determined by statistically analyzing the change in the objective function value between two consecutive iterations. For example, 5% to 10% of the standard deviation of the objective function change can be used as the convergence threshold.

[0101] The tolerance is set to ensure that the numerical optimization solution will not cause safety boundary failure in actual execution; the specific value can be determined according to the allowable error ratio of the vehicle's key safety constraints (such as minimum safety distance or maximum lateral deviation), for example, taking 1% to 5% of the corresponding safety constraint limit as the tolerance.

[0102] S7. Based on the final control command, select the appropriate communication path and send it to the vehicle controller for execution.

[0103] It should be noted that the final control commands (including steering wheel angle commands and longitudinal acceleration commands), along with timestamps and checksums, are encapsulated into application-layer data packets according to a preset communication protocol (such as the SOME / IP service protocol in the AUTOSAR standard). These application-layer data packets are then encapsulated into final data frames that can be transmitted over a selected communication link, passing through the transport layer, network layer, and data link layer. An optimal path is dynamically selected based on the current available communication link status; for example, when both cellular networks and direct communication links are available, the direct link with lower latency is preferred. The encapsulated final data frame is sent to the vehicle controller via the selected physical communication path. Upon receiving the data frame, the vehicle controller's communication interface decodes, verifies, and decapsulates the data to extract the final control commands. The steering wheel angle command is sent to the steering actuator controller to drive the steering motor, and the longitudinal acceleration command is converted into a specific drive torque request or braking pressure request, which is then sent to the drive motor controller and brake controller, respectively, thereby driving the drive actuator and brake actuator to complete the action.

[0104] This embodiment also provides a computer device applicable to the intelligent connected platform control method based on vehicle communication, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent connected platform control method based on vehicle communication as proposed in the above embodiment.

[0105] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0106] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the intelligent connected platform control method based on vehicle communication as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0107] In summary, this invention, through a joint modeling method based on kernel density estimation and Bayesian updating, transforms communication uncertainty into a quantifiable, high-confidence safety margin boundary, overcoming the limitations of previous static safety distances. It introduces a risk assessment mechanism that integrates three elements, enabling intelligent dynamic switching between centralized, distributed, and local full-authority control modes. This ensures that the optimal control strategy is always matched under different communication conditions. Finally, the safety margin parameter is used as a rigid constraint in the rolling time-domain optimization problem, limiting all control commands to the safe and feasible domain from the source. This truly achieves the synergistic goal of "safety first, performance adaptive," improving the operational safety and stability of intelligent connected vehicles in complex traffic environments.

[0108] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A control method for an intelligent connected platform based on vehicle communication, characterized in that: include, Collect vehicle status information, surrounding environment perception information, communication link status information, and control command information from the connected platform; Based on the communication link status information, calculate the communication quality index that characterizes the overall quality of the current primary communication link; Based on the vehicle's physical status information, surrounding environment perception information, and control command information, determine the level of control requirements currently needed by the vehicle. Based on communication link status information, the probability distribution of end-to-end delay and packet loss rate in the rolling time domain is predicted, and based on vehicle status information, control requirement level and surrounding environment perception information, the safety margin parameters that meet the preset safety risk limit are calculated. Based on safety margin parameters, communication quality index and control requirement level, dynamic selection is made between the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode and the full control mode of the vehicle controller. The vehicle control algorithm parameters corresponding to the selected control mode are adjusted in real time, and the vehicle control problem is modeled as a rolling time domain optimization problem. At the same time, a safety margin parameter is added as the boundary condition of the optimization problem. The final control command is obtained by solving the rolling time domain optimization problem. Based on the final control command, the appropriate communication path is selected and sent to the vehicle controller for execution.

2. The intelligent connected platform control method based on vehicle communication as described in claim 1, characterized in that: The vehicle's status information includes vehicle speed, acceleration, and steering wheel angle; The surrounding environment perception information includes obstacle distance and pedestrian target status; The communication link status information includes historical end-to-end latency, historical latency jitter data, and historical packet loss rate; The control command information from the connected platform includes global path planning commands and cooperative driving suggestions.

3. The intelligent connected platform control method based on vehicle communication as described in claim 2, characterized in that: The communication quality index, which represents the overall quality of the current primary communication link, is calculated by normalizing each parameter in the communication link status information, and then weighting and summing all the normalized communication link status information parameters to obtain the communication quality index.

4. The intelligent connected platform control method based on vehicle communication as described in claim 3, characterized in that: The determination of the current control requirement level of the vehicle is specifically as follows: Based on vehicle status information, surrounding environment perception information, and control command information, the driving scenario and task objectives of the vehicle are comprehensively analyzed. For each driving scenario and task objective, a control requirement level is assigned.

5. The intelligent connected platform control method based on vehicle communication as described in claim 4, characterized in that: The calculation of the safety margin parameter that satisfies the preset safety risk upper limit is specifically as follows: Based on communication link status information, time windows are divided according to the rolling time domain. Kernel density estimation is used to calculate the probability distribution of historical end-to-end delay within each time window. The Beta distribution is used to fit parameters to the historical packet loss rate to obtain the probability distribution of packet loss rate. Based on the probability distribution of end-to-end delay and the probability distribution of packet loss rate, calculate the upper bound of the deviation caused by the uncertainty of end-to-end delay and the uncertainty of packet loss rate on the effective time of control command information; Based on the upper bound of the deviation and the vehicle speed in the vehicle's physical state information, the corresponding range of vehicle position uncertainty is calculated, and a safety risk upper limit threshold is preset. Based on the range of location uncertainty and the distance to obstacles in the surrounding environment perception information, calculate the safety margin parameter that meets the upper limit threshold of safety risk.

6. The intelligent connected platform control method based on vehicle communication as described in claim 5, characterized in that: The dynamic selection between the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller is specifically as follows: Calculate the communication uncertainty risk value, insufficient control capability risk value, and insufficient safety margin risk value for the centralized master control mode of the connected platform, the vehicle-cloud collaborative control mode, and the full control mode of the vehicle controller under the current communication quality index and control requirement level. The communication uncertainty risk value, control capability insufficiency risk value, and safety margin insufficiency risk value of each control mode are weighted and fused to obtain the comprehensive control risk value corresponding to each control mode. The control mode with the lowest overall control risk value is selected as the control mode for the current moment.

7. The intelligent connected platform control method based on vehicle communication as described in claim 6, characterized in that: Obtain the final control command, specifically: The vehicle control algorithm parameters include the lengths of the prediction and control time domains of the model predictive controller, the feedforward and feedback gains of the trajectory tracking controller, and the noise covariance matrix of the state estimator. Based on the selected control mode and safety margin parameters, adjust the lengths of the prediction time domain and control time domain of the model predictive controller, adjust the feedforward gain and feedback gain of the trajectory tracking controller, and adjust the noise covariance matrix of the state estimator. Based on the adjusted vehicle control algorithm parameters, the vehicle control problem is modeled as a rolling time-domain optimization problem. The objective function of the rolling time-domain optimization problem is designed to minimize the weighted sum of tracking error, control energy consumption, and ride discomfort. In the constraint set of the rolling time-domain optimization problem, boundary conditions defined by the safety margin parameter are added; The solver solves the rolling time-domain optimization problem with boundary conditions to obtain the optimal control input sequence in the next time domain. The first control variable in the optimal control input sequence is then extracted as the final control command.

8. The intelligent connected platform control method based on vehicle communication as described in claim 7, characterized in that: The phrase "based on the final control command, select the corresponding communication path and send it to the vehicle controller for execution" means that the final control command is encapsulated into a data packet and a communication path is selected to send it to the vehicle controller. After receiving the data packet, the vehicle controller instructs the steering, driving and braking actuators to complete the action.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent connected platform control method based on vehicle communication as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent connected platform control method based on vehicle communication as described in any one of claims 1 to 8.

Images