A vehicle cruise control method and system

Abstract

The application provides a vehicle cruise control method and system, which is applied to a cloud platform server of a vehicle cruise control system. The method comprises the following steps: acquiring multi-source slope data of a future driving path of a target vehicle to obtain a fused road slope sequence; generating an energy-saving reference trajectory envelope centered on an energy-saving vehicle speed trajectory based on the fused road slope sequence; searching for a target lane-changing sequence and a vehicle speed adjustment sequence based on the energy-saving reference trajectory envelope as a constraint, in combination with multi-lane traffic flow information and real-time vehicle state information uploaded by a terminal platform device of the target vehicle; and issuing a control instruction containing the energy-saving reference trajectory envelope, the target lane-changing sequence and the vehicle speed adjustment sequence to the terminal platform device of the target vehicle, so that the terminal platform device performs vehicle cruise control based on the received control instruction. The vehicle cruise control method can effectively balance energy-saving effect and terminal control flexibility.

Description

Technical Field

[0001] This application relates to the field of new energy vehicle technology, and in particular to a vehicle cruise control method and system. Background Technology

[0002] With the continuous development of vehicle-road-cloud integrated systems, cloud-controlled energy-saving technology for intelligent connected vehicles has become a key direction for current industrial application.

[0003] However, existing energy-saving control technologies for commercial vehicles still have significant shortcomings. Existing solutions are largely limited by the vehicle's local computing power and localized sensing capabilities, making it difficult to balance the energy efficiency of cloud-based global planning with the flexibility of vehicle-side execution under complex road conditions. Furthermore, existing speed optimization schemes fail to fully unleash the vehicle's energy-saving potential, resulting in actual energy-saving effects falling short of expectations. Summary of the Invention

[0004] In view of this, this application provides a vehicle cruise control method and system that balances energy saving and vehicle-side control flexibility.

[0005] In a first aspect, this application provides a vehicle cruise control method, applied to a cloud platform server of a vehicle cruise control system, the method comprising:

[0006] Obtain multi-source slope data of the future driving path of the target vehicle, and perform fusion processing on the multi-source slope data to obtain the fused road slope sequence;

[0007] Based on the fused road slope sequence, an energy-saving reference trajectory envelope interval centered on the energy-saving vehicle speed trajectory is generated;

[0008] Using the energy-saving reference trajectory envelope as a constraint, and combining multi-lane traffic flow information with real-time vehicle status information uploaded by the target vehicle's on-board platform device, the target lane-changing sequence and speed adjustment sequence are searched.

[0009] Control commands, including the energy-saving reference trajectory envelope, the target lane-changing sequence, and the vehicle speed adjustment sequence, are sent to the vehicle-side platform device of the target vehicle, so that the vehicle-side platform device can perform vehicle cruise control based on the received control commands.

[0010] Optionally, the step of generating an energy-saving reference trajectory envelope interval centered on the energy-saving vehicle speed trajectory based on the fused road slope sequence includes:

[0011] Based on the fused road slope sequence, the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated;

[0012] Using a preset allowable energy consumption deviation threshold as a limiting condition, and combining it with the vehicle power system working boundary corresponding to the target gear sequence, the maximum allowable acceleration deviation and the maximum allowable deceleration deviation relative to the energy-saving vehicle speed trajectory are calculated.

[0013] Based on the energy-saving vehicle speed trajectory, the maximum permissible acceleration deviation, and the maximum permissible deceleration deviation, a permissible driving speed range is constructed, and the permissible driving speed range is used as the envelope range of the energy-saving reference trajectory.

[0014] Optionally, the step of calculating the energy-saving vehicle speed trajectory and the matching target gear sequence based on the fused road slope sequence includes:

[0015] Based on the fused road slope sequence, a vehicle dynamics model including slope resistance is established within the driving distance domain;

[0016] Using the weighted sum of the vehicle's overall energy consumption and driving time as the optimization objective, the vehicle dynamics model is used to simulate and extrapolate the driving conditions of the road segment, and the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated.

[0017] Optionally, establishing a vehicle dynamics model including gradient resistance within the driving distance domain includes:

[0018] Within the driving distance domain, a discrete state-space model of vehicle longitudinal dynamics, including gradient resistance, is constructed. This discrete state-space model integrates braking energy recovery characteristics and coasting characteristics, and includes a comprehensive driving resistance model composed of rolling resistance and air resistance.

[0019] Optionally, the step of using the weighted sum of the vehicle's overall energy consumption and driving time as the optimization objective, and simulating the road segment driving conditions based on the vehicle dynamics model to calculate the energy-saving vehicle speed trajectory and the matching target gear sequence, includes:

[0020] Set adjustable energy consumption weighting factors and time penalty weighting factors, and construct an optimization function with the goal of minimizing the weighted sum of comprehensive energy consumption and driving time during vehicle operation.

[0021] Configure driving constraints, which include at least upper and lower limits for vehicle speed, upper and lower limits for acceleration, and feasible gear constraints.

[0022] By calling the intelligent simulation model of the vehicle power system, within the boundaries of the driving constraints, a dynamic programming algorithm is used to find the optimal energy-saving speed trajectory that matches the optimization function and the target gear sequence that matches the energy-saving speed trajectory.

[0023] Optionally, the step of searching for the target lane-changing sequence and speed adjustment sequence by using the energy-saving reference trajectory envelope interval as a constraint, combined with multi-lane traffic flow information and real-time vehicle status information uploaded by the target vehicle's on-board platform device, includes:

[0024] The system acquires multi-lane traffic flow information and real-time vehicle status information uploaded by the vehicle-side platform device of the target vehicle. The real-time vehicle status information includes the vehicle's own driving status data and the side and rear vehicle target information collected by the vehicle-mounted multi-source perception system.

[0025] A forward behavior tree search algorithm is used, with the energy-saving reference trajectory envelope interval as a constraint, combined with the multi-lane traffic flow information and the real-time vehicle status information, to perform lane-changing and speed-coordinated planning;

[0026] During the behavior tree unfolding and deduction process, if a candidate lane-changing behavior or speed adjustment behavior causes the vehicle speed to exceed the energy-saving reference trajectory envelope range, then the branch corresponding to that behavior will be penalized or pruned.

[0027] The collaborative planning is completed with the goal of minimizing the overall cost. The target lane-changing sequence and speed adjustment sequence are obtained by searching. The overall cost includes the cost of the vehicle speed deviating from the energy-saving speed trajectory, the cost of the change in travel time, and the cost of lane-changing safety risks.

[0028] Optionally, the step of acquiring multi-source slope data of the future driving path of the target vehicle and fusing the multi-source slope data to obtain a fused road slope sequence includes:

[0029] The multi-source slope data corresponding to the future driving path of the target vehicle is obtained by classification. The multi-source slope data includes at least high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data.

[0030] Preprocessing is performed on the high-precision map slope data, the real-time vehicle sensor inverted slope data, and the historical vehicle trajectory statistical slope data, respectively.

[0031] Multi-source fusion calculations are performed on preprocessed high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data to output a fused road slope sequence.

[0032] Optionally, the classification process for obtaining multi-source slope data corresponding to the future driving path of the target vehicle includes:

[0033] Access a high-precision map database to obtain discrete elevation data on the future driving path of the target vehicle, and calculate the slope data of the corresponding road segment based on the elevation difference and distance difference between adjacent sampling points, which is used as the slope data of the high-precision map.

[0034] The longitudinal acceleration data and vehicle speed change data collected by the vehicle inertial measurement unit are acquired. Based on the longitudinal dynamics of the vehicle, the longitudinal acceleration data and the vehicle speed change data are inverted and calculated to obtain a real-time slope estimate, which is used as the real-time vehicle sensor inverted slope data.

[0035] Historical vehicle trajectory data collected when vehicles travel on the target road segment is obtained. The trajectory data is statistically analyzed, and slope feature information is extracted to generate historical vehicle trajectory statistical slope data.

[0036] Optionally, the step of performing multi-source fusion calculation on the preprocessed high-precision map slope data, real-time vehicle sensor-derived slope data, and historical vehicle trajectory statistical slope data, and outputting the fused road slope sequence, includes:

[0037] Initial weight coefficients are assigned to the slope data of high-precision map, the slope data inverted by real-time vehicle sensors, and the slope data statistically obtained from historical vehicle trajectories, and the sum of the initial weight coefficients of the three types of slope data is a set value.

[0038] For the sampling interval along the future driving path of the target vehicle, the confidence level of the three types of slope data in the corresponding sampling interval is calculated in real time;

[0039] The weight coefficients of the three types of slope data are updated based on the confidence scores obtained in real time. The weighted Kalman filter algorithm is used to complete the fusion calculation of multi-source slope data and output the fused road slope sequence.

[0040] Secondly, this application provides a vehicle cruise control system, including a cloud platform server and a vehicle-end platform device connected to the cloud platform server via vehicle network communication, wherein the cloud platform server is used to execute the steps in the vehicle cruise control method provided in the embodiments of this application.

[0041] The vehicle cruise control method provided in this application acquires multi-source slope data of the target vehicle's future driving path from a cloud platform server, fuses and processes this data to obtain a high-precision road slope sequence, and generates an energy-saving reference trajectory envelope interval centered on the energy-saving speed trajectory based on the road slope sequence, replacing the single fixed speed trajectory of the traditional scheme, thus balancing energy-saving effect and vehicle-side control flexibility. Simultaneously, using the energy-saving reference trajectory envelope interval as the core constraint, and combining global traffic flow information and real-time vehicle status, the method collaboratively plans the target lane-changing sequence and speed adjustment sequence, achieving deep collaboration between lane-changing decisions and energy-saving targets. Finally, control commands are sent to the vehicle for execution, completing intelligent energy-saving cruise with vehicle-cloud collaboration. This vehicle cruise control method can solve the problems of poor anti-disturbance capability of the single optimal speed trajectory, disconnect between lane-changing decisions and energy-saving planning, and unsatisfactory global energy-saving effect in traditional vehicle cruise schemes, effectively reducing driving energy consumption during long-distance cruise while ensuring driving safety and scenario adaptability. Attached Figure Description

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

[0043] Figure 1 This is a schematic diagram of the steps of the vehicle cruise control method provided in the embodiments of this application;

[0044] Figure 2 This is a schematic diagram illustrating the calculation steps of the energy-saving reference trajectory envelope interval provided in the embodiments of this application;

[0045] Figure 3 This is a schematic diagram illustrating the calculation steps of the energy-saving vehicle speed trajectory provided in the embodiments of this application;

[0046] Figure 4 This is a schematic diagram of the calculation steps for the energy-saving vehicle speed trajectory provided in another embodiment of this application;

[0047] Figure 5 This is a schematic diagram of the search steps for the target lane change sequence and vehicle speed adjustment sequence provided in the embodiments of this application;

[0048] Figure 6 This is a schematic diagram of the structure of the vehicle cruise control system provided in the embodiments of this application;

[0049] Figure 7 This is a structural schematic diagram of the target vehicle and vehicle-side platform equipment provided in the embodiments of this application.

[0050] Explanation of reference numerals in the attached figures:

[0051] 1-Target vehicle, 100-Vehicle cruise control system, 101-Cloud platform server, 102-Vehicle platform equipment. Detailed Implementation

[0052] 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 a part of the embodiments of the present invention, and not all of them. 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.

[0053] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0054] In this document, references to "embodiment" or "implementation" mean that a particular feature, structure, or characteristic described in connection with an embodiment or implementation may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0055] Before introducing the technical solution of this application, let's go over the technical issues in related technologies in detail.

[0056] With the continuous development of vehicle-road-cloud integrated systems, cloud-controlled energy-saving technology for intelligent connected vehicles has become a key area for industrial application. However, existing energy-saving control technologies for commercial vehicles still have significant technical bottlenecks and shortcomings.

[0057] Among these shortcomings, existing technical solutions lack a cloud-controlled layered architecture design for energy-saving driving applications. Existing solutions largely rely on vehicle-side computing power and the local perception capabilities of onboard sensors, making it difficult to handle large-scale, multi-target collaborative optimization problems. They also severely underutilize the abundant computing resources and global traffic information available in the cloud, limiting the system's performance ceiling. Furthermore, traditional cloud-controlled cruise solutions often employ a mode of directly issuing a single optimal vehicle speed or acceleration sequence from the cloud. The optimal energy-saving solution is highly susceptible to failure in scenarios such as lane changes and traffic disturbances, and the vehicle-side control freedom is severely limited.

[0058] Meanwhile, existing studies on energy-saving vehicle speed optimization based on slope information do not fully consider the characteristics of the power system, such as regenerative braking and coasting in neutral gear, in electric heavy-duty trucks, resulting in limited energy-saving effects. Furthermore, existing technologies suffer from a disconnect between static and dynamic traffic information, often relying solely on static road slope information or localized dynamic traffic information, failing to achieve deep integration and collaborative optimization of long-distance static geographic information and real-time dynamic traffic flow.

[0059] Furthermore, the decision-making level and scope of existing technologies are limited. Lane-changing decisions are usually based on short-sighted rules or local game theory, and are not integrated with long-term energy-saving speed planning based on global road information. This can easily lead to short-sighted decision-making, frequent lane changes, or missing the best passage opportunity, which will increase vehicle energy consumption and make it impossible to achieve globally optimal energy-saving control.

[0060] In view of this, to solve the above problems, this application provides a vehicle cruise control method, applied to the cloud platform server 101 of the vehicle cruise control system 100. Please refer to... Figure 1 The method includes the following steps:

[0061] S10. Obtain multi-source slope data of the future driving path of the target vehicle, and perform fusion processing on the multi-source slope data to obtain the fused road slope sequence.

[0062] Optionally, the aforementioned vehicle cruise control system 100 can be understood as an intelligent driving system with global path planning, energy-saving trajectory calculation, and vehicle-to-cloud data interaction functions. It can be applied to scenarios such as high-speed trunk line cruise for electric commercial vehicles and intercity long-distance cruise for passenger vehicles. This application does not impose any special limitations on the applicable vehicle models. The executing entity of the aforementioned vehicle cruise control method is the cloud platform server 101. The cloud platform server 101 and the vehicle-side platform device 102 of the target vehicle 1 are connected bidirectionally via a vehicle network communication link, which can realize real-time data interaction and the issuance of control commands.

[0063] Optionally, the aforementioned target vehicle 1 can be understood as a vehicle equipped with the vehicle cruise control system 100 of this application, having completed identity binding with the cloud platform server 101, and currently performing a cruise driving task.

[0064] Optionally, the aforementioned future driving path can be understood as the complete driving route from the current location to the destination planned by the target vehicle 1 through the vehicle navigation system. The cloud platform server 101 can obtain the complete information of the future driving path from the vehicle-side platform device 102 in real time and lock the road segment range where the target vehicle 1 is about to travel.

[0065] Optionally, the aforementioned multi-source slope data can be understood as at least two or more data sources that can characterize the slope information of the target road segment. Different data sources can complement each other and solve the problems of insufficient accuracy and poor adaptability of a single slope data source.

[0066] Optionally, the above fusion process can be understood as the process of standardizing, calibrating, and aligning multi-source slope data, and then merging the data through a preset fusion algorithm to output high-precision and high-reliability slope data.

[0067] Optionally, the above road slope sequence can be understood as a continuous slope dataset with equal intervals that matches the future driving path. Each sampling point in the sequence corresponds to a unique location within the driving distance domain, as well as the road slope value corresponding to that location.

[0068] Optionally, the data source type for multi-source slope data can be configured according to the actual application scenario, such as high-precision map slope data, real-time vehicle sensor-derived slope data, and historical vehicle trajectory statistical slope data.

[0069] Optionally, the cloud platform server 101 can update the coverage of the future driving path according to the current driving position of the target vehicle 1 at a preset cycle to ensure that complete slope data within the preset distance ahead is always obtained, thus ensuring the continuity of the planning during the cruise.

[0070] S20. Based on the fused road slope sequence, generate an energy-saving reference trajectory envelope interval centered on the energy-saving vehicle speed trajectory.

[0071] Optionally, the aforementioned energy-saving vehicle speed trajectory can be understood as a global speed planning sequence that achieves an optimal balance between overall vehicle energy consumption and travel time along the future travel path of the target vehicle 1. The aforementioned energy-saving reference trajectory envelope can be understood as the allowable driving speed range that fluctuates up and down, with the energy-saving vehicle speed trajectory as the core. As long as the vehicle's driving speed is within this range, the deviation between its driving energy consumption and the globally optimal energy consumption can be controlled within a preset threshold, thus taking into account the core objective of global energy saving while reserving sufficient speed adjustment space for the vehicle to cope with sudden traffic scenarios.

[0072] Optionally, the cloud platform server 101 uses the fused road slope sequence as the core input to complete vehicle dynamics modeling within the travel distance domain. Compared to traditional time-domain modeling, travel distance-domain modeling can accurately match the distribution characteristics of road slope changes with travel distance, improving the accuracy of energy consumption simulation and trajectory planning. Subsequently, using the weighted sum of the comprehensive energy consumption and travel time throughout the vehicle's journey as the optimization objective, the global energy-saving speed trajectory of the target vehicle 1 on its future travel path is obtained. Then, based on a preset allowable energy consumption deviation threshold, the maximum allowable acceleration deviation and the maximum allowable deceleration deviation relative to the energy-saving speed trajectory are calculated, ultimately constructing an energy-saving reference trajectory envelope interval centered on the energy-saving speed trajectory.

[0073] It should be noted that the preset allowable energy consumption deviation thresholds can be flexibly configured according to user needs, vehicle type, and driving scenario. For example, a smaller deviation threshold can be set for long-haul freight scenarios to prioritize energy-saving performance; a larger deviation threshold can be set for cruising around cities to prioritize driving flexibility. Optionally, the energy-saving reference trajectory envelope can be dynamically adjusted according to changes in road gradient. On road sections with drastic gradient changes, such as long downhill slopes and mountain curves, the envelope range is narrowed to strengthen energy-saving constraints; on long straight sections and highways with gentle gradients, the envelope range is widened to improve vehicle-side control flexibility.

[0074] S30. Using the energy-saving reference trajectory envelope interval as a constraint, and combining multi-lane traffic flow information with real-time vehicle status information uploaded by the target vehicle's vehicle-side platform equipment, the target lane-changing sequence and speed adjustment sequence are searched and obtained.

[0075] Optionally, the aforementioned multi-lane traffic flow information can be understood as the traffic flow status information of each lane on the future travel path of the target vehicle 1. This information can be obtained through vehicle-road cooperative systems, real-time data uploaded by connected vehicles on the same road segment, and traffic big data platforms, including but not limited to information such as the average vehicle speed, traffic density, following distance, and location of slow-moving road sections ahead for each lane.

[0076] Optionally, the aforementioned real-time vehicle status information can be understood as information collected and uploaded in real time by the vehicle-side platform device 102, representing the vehicle's current driving status and surrounding environment.

[0077] Optionally, the above lane-changing sequence can be understood as a planning sequence of lane-changing timing, target lane, and number of lane changes for target vehicle 1 on its future travel path. The above speed adjustment sequence can be understood as a segment-specific speed adjustment planning sequence that matches the lane-changing sequence.

[0078] Optionally, the cloud platform server 101 first acquires the predicted multi-lane traffic flow information within a preset future distance domain, and simultaneously receives the vehicle status information uploaded in real time by the vehicle-side platform device 102 of the target vehicle 1. Then, using the generated energy-saving reference trajectory envelope as the core constraint, a global path search algorithm is employed to collaboratively plan lane-changing and speed adjustment behaviors. During the planning process, if a candidate lane-changing or speed adjustment behavior causes the vehicle speed to exceed the energy-saving reference trajectory envelope, the candidate behavior is penalized or directly eliminated. Ultimately, with the goal of minimizing the overall cost, the globally optimal lane-changing sequence and speed adjustment sequence are searched, thereby achieving deep collaboration between lane-changing decisions and energy-saving goals. This effectively solves the problem of disconnect between lane-changing decisions and energy-saving planning, and the additional energy consumption caused by lane-changing behavior in traditional cruise control schemes.

[0079] It should be noted that the aforementioned global path search algorithm can be selected based on the cloud computing power configuration, including but not limited to forward behavior tree search algorithm or model predictive control algorithm, etc. This application does not impose any special restrictions on the algorithm type. Optionally, the cloud platform server 101 can update multi-lane traffic flow information and vehicle status information in real time according to a preset rolling time domain window, and continuously update the target lane change sequence and vehicle speed adjustment sequence to ensure the matching degree between the planning results and real-time traffic conditions.

[0080] S40. A control command containing the energy-saving reference trajectory envelope, the target lane-changing sequence, and the vehicle speed adjustment sequence is sent to the vehicle-side platform device of the target vehicle, so that the vehicle-side platform device performs vehicle cruise control based on the received control command.

[0081] Optionally, the aforementioned vehicle-side platform device 102 can be understood as an on-board device installed on the target vehicle 1, possessing data acquisition, communication interaction, and vehicle control functions, including but not limited to components such as an autonomous driving domain controller, vehicle controller, on-board gateway, and on-board communication unit. It can receive control commands issued by the cloud platform server 101 and convert them into lateral and longitudinal control actions for the vehicle. The aforementioned vehicle-to-everything (V2X) communication link can adopt wireless communication methods such as 5G, C-V2X, and 4G, and can adaptively switch according to the actual application scenario to ensure the real-time performance and reliability of control command issuance.

[0082] In this embodiment, the cloud platform server 101 sends cruise control commands, including the energy-saving reference trajectory envelope, the target lane-changing sequence, and the vehicle speed adjustment sequence, to the vehicle-side platform device 102 of the target vehicle 1 via the vehicle network communication link. After receiving the control commands, the vehicle-side platform device 102, through the vehicle controller, the autonomous driving domain controller, and other execution units, completes the coordinated execution of longitudinal speed control and lateral lane-changing control based on the commands, ultimately realizing intelligent energy-saving cruise control with vehicle-cloud collaboration.

[0083] In summary, the vehicle cruise control method provided in this embodiment acquires multi-source slope data of the future driving path of the target vehicle 1 through a cloud platform server 101, and obtains a high-precision road slope sequence through fusion processing. Based on the road slope sequence, an energy-saving reference trajectory envelope interval centered on the energy-saving speed trajectory is generated, replacing the single fixed speed trajectory of the traditional scheme, thus balancing energy-saving effect and vehicle-side control flexibility. Simultaneously, using the energy-saving reference trajectory envelope interval as the core constraint, combined with global traffic flow information and real-time vehicle status, the target lane-changing sequence and speed adjustment sequence are collaboratively planned to achieve deep collaboration between lane-changing decisions and energy-saving targets. Finally, control commands are sent to the vehicle for execution, completing intelligent energy-saving cruise with vehicle-cloud collaboration. This vehicle cruise control method can solve the problems of poor anti-disturbance capability of the single optimal speed trajectory, disconnect between lane-changing decisions and energy-saving planning, and unsatisfactory global energy-saving effect in traditional vehicle cruise schemes. It effectively reduces driving energy consumption during long-distance cruise while ensuring driving safety and scenario adaptability.

[0084] Optionally, the step of acquiring multi-source slope data of the future driving path of target vehicle 1 and fusing the multi-source slope data to obtain a fused road slope sequence includes the following steps:

[0085] S11. Classify and obtain multi-source slope data corresponding to the future driving path of the target vehicle. The multi-source slope data includes at least high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data.

[0086] Optionally, the aforementioned high-precision map slope data can be understood as forward-looking global slope data generated based on elevation information from a high-precision map database, possessing the core advantages of wide coverage, high path matching accuracy, and no real-time acquisition delay. The aforementioned real-time vehicle sensor-derived slope data can be understood as current road segment slope data derived from real-time driving data collected by the on-board sensors of the target vehicle, possessing the core advantages of high vehicle-to-everything (V2X) matching accuracy, strong real-time performance, and the ability to reflect the true slope resistance under the current environment. The aforementioned historical vehicle trajectory statistical slope data can be understood as crowdsourced slope data generated based on the statistical analysis of trajectory data from a large number of historical vehicles traveling on the same road segment, possessing the core advantages of large sample size, high statistical accuracy, and the ability to reflect the long-term true slope characteristics of the road segment.

[0087] Optionally, the three types of slope data acquired by the cloud platform server 101 have a spatial coverage area that matches the future driving path of the target vehicle 1, and also include data on the already traveled road segments within a preset distance behind the vehicle's current position, used to complete data benchmark calibration and error correction. Optionally, in addition to the three core data sources, the cloud platform server 101 can also supplement the data by acquiring road segment slope data collected by vehicle-road cooperative roadside equipment and slope inversion data uploaded in real time by other connected vehicles on the same road segment, further enriching the data source dimensions and improving data reliability. For road segments with drastic slope changes, such as mountain slopes and long-distance uphill and downhill sections, the cloud platform server 101 can automatically increase the sample size of the acquired data and encrypt the data sampling points to ensure the accuracy of the data source for complex road segments.

[0088] S12. Perform preprocessing on the high-precision map slope data, the real-time vehicle sensor inverted slope data, and the historical vehicle trajectory statistical slope data, respectively.

[0089] In one optional embodiment of this application, the cloud platform server 101 performs general standardized preprocessing and data source-specific preprocessing sequentially for the three types of slope data. Specifically, this includes the following operations: outlier filtering, removing outliers and jump values ​​from the three types of data, filtering invalid slope data caused by sensor malfunctions, missing map data, or abnormal historical data conditions, and retaining valid data samples; and spatial coordinate matching, uniformly mapping the spatial coordinates of the three types of slope data to a driving distance coordinate system based on the future driving path of the target vehicle 1, so that the sampling points of all data are based on the cumulative driving distance. To establish a location benchmark, spatial coordinate misalignment issues between different data sources are eliminated. Sampling intervals are uniformly processed using linear interpolation, spline interpolation, and other methods to adjust the sampling intervals of the three types of data to a preset fixed interval. This interval is consistent with the discrete sampling interval of the subsequent vehicle dynamics model, ensuring a one-to-one correspondence between all data sampling points and providing a foundation for point-by-point fusion calculations. Benchmark calibration is performed using the actual vehicle-derived slope data of the road segment already traveled by target vehicle 1 as a benchmark. Zero-point deviations and system errors in the high-precision map slope data and historical vehicle trajectory statistical slope data are calibrated to eliminate benchmark offsets between different data sources.

[0090] Optionally, depending on the characteristics of different data sources, the cloud platform server 101 performs dedicated preprocessing operations: for real-time vehicle sensor inversion slope data, it performs moving average filtering and adaptive noise reduction to eliminate data fluctuations caused by vehicle driving bumps and sensor noise; for historical vehicle trajectory statistical slope data, it performs data weighting and classification processing, classifying historical data according to data collection time, vehicle type, and sample quality, and removing low-quality samples; for high-precision map slope data, it performs data integrity verification and completes road sections with missing elevation data.

[0091] Optionally, all preprocessing operations are performed within the driving distance coordinate system to ensure that the three types of preprocessed data have corresponding valid slope values ​​at each driving distance sampling point, with no missing data or spatial misalignment. Optionally, the cloud platform server 101 can adaptively adjust the preprocessing parameters according to the type and quality of the data source. For example, for real-time inversion data with large fluctuations, it can automatically adjust the size of the filtering window to ensure the smoothness and accuracy of the data.

[0092] S13. Perform multi-source fusion calculation on the preprocessed high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data, and output the fused road slope sequence.

[0093] Optionally, the above-mentioned fused road slope sequence can be understood as a continuous slope dataset with equal intervals that corresponds one-to-one with the driving distance domain of the future driving path of the target vehicle 1. Each driving distance sampling point in the sequence corresponds to a unique fused slope value, which can accurately characterize the road slope characteristics at the corresponding location.

[0094] In one optional embodiment of this application, the cloud platform server 101 takes the preprocessed three types of slope data as input and uses a weighted Kalman filter algorithm to complete the multi-source fusion calculation. The fusion algorithm can be flexibly selected based on the cloud computing power configuration and driving scenario. In addition to the weighted Kalman filter algorithm, federated average fusion algorithm, Bayesian estimation fusion algorithm, etc., can also be used. This application does not impose any special limitation on the specific type of fusion algorithm, as long as it can integrate the effective information from multiple source data and output a high-precision fused slope sequence.

[0095] Optionally, the cloud platform server 101 can adaptively adjust the initial allocation rules of the fusion weights according to the road segment type. For example, in long-distance, flat highway sections, the initial weight of historical vehicle trajectory statistical slope data is increased; in mountainous sections with drastic slope changes, the initial weight of high-precision map slope data is increased; and in continuous road sections where vehicles have already traveled, the weight of real-time vehicle sensor inverted slope data is increased to further improve the fusion accuracy.

[0096] In this embodiment, the vehicle cruise control method acquires three core data sources: high-precision map slope data, real-time vehicle sensor-derived slope data, and historical vehicle trajectory statistical slope data. High-precision map data addresses the problem of limited vehicle-end perception range and inability to obtain long-distance forward-looking slope information; real-time inversion data addresses the issue of outdated static map data and disconnection from the current operating conditions of the actual vehicle; and historical statistical data addresses the problem of insufficient real-time data sample size and large random errors for a single vehicle. This vehicle cruise control method breaks through the limitations of traditional solutions relying on a single data source, providing comprehensive and highly reliable raw data support for the generation of high-precision slope sequences from the data source end, thus effectively adapting to the long-distance, all-scenario forward-looking planning needs of commercial vehicles for long-distance trunk line cruises. Furthermore, this solution filters out invalid samples caused by sensor malfunctions and missing data through outlier filtering, maps all data uniformly to a driving distance domain coordinate system based on the vehicle's driving path through spatial coordinate matching, and eliminates zero-point deviations and system errors between different data sources through real-vehicle inversion data benchmark calibration. Furthermore, this solution integrates effective information from three types of data sources through a multi-source fusion algorithm, which can effectively eliminate random and systematic errors from a single data source. The final output of the fused road slope sequence has higher accuracy than that from a single data source, effectively controlling slope data errors.

[0097] Optionally, the classification to obtain multi-source slope data corresponding to the future driving path of target vehicle 1 includes the following steps:

[0098] S111. Access the high-precision map database, obtain discrete elevation data on the future driving path of the target vehicle, and calculate the slope data of the corresponding road segment based on the elevation difference and distance difference between adjacent sampling points, as the slope data of the high-precision map.

[0099] Optionally, the aforementioned high-precision map database can be understood as a professional map database with decimeter-level spatial accuracy and containing spatial information of all road elements. It can be accessed through commercial high-precision map service platforms or high-precision map platforms built by car manufacturers. The database stores discrete elevation point data of the road network. Each elevation point corresponds to a unique geographic coordinate and altitude value. The data accuracy can meet the needs of long-distance cruise energy-saving planning.

[0100] Optionally, the aforementioned discrete elevation data can be understood as a dataset of discrete points containing geographic coordinates and elevation values ​​sampled at preset equal intervals along the future driving path of the target vehicle 1, which is the basic raw data for calculating road slope.

[0101] In one optional embodiment of this application, the cloud platform server 101 first locks the complete geographical range and mileage interval of the future driving path based on the navigation planning path uploaded by the target vehicle 1. Then, it accesses the high-precision map database through a standardized interface to retrieve the full amount of discrete elevation data within the coverage area of ​​the driving path, with the data sampling interval consistent with the discrete sampling interval of the subsequent vehicle dynamics model. Next, it maps the geographical coordinates of the discrete elevation data to the driving distance domain coordinate system based on the driving path of the target vehicle 1 to obtain the cumulative driving distance and elevation value corresponding to each sampling point. Finally, it calculates the slope value of the corresponding road segment point by point through the elevation difference and distance difference between adjacent sampling points, generating high-precision map slope data that corresponds one-to-one with the future driving path.

[0102] Optionally, when acquiring high-precision map elevation data, the cloud platform server 101 simultaneously performs data version verification and integrity verification, prioritizing the use of the latest version of high-precision map data. For road sections with missing elevation data, interpolation of adjacent valid sampling points is used to complete the data, ensuring that there is no missing data along the entire path. For road sections with drastic slope changes, such as mountain slopes and continuous uphill / downhill sections, the cloud platform server 101 can automatically increase the sampling interval of elevation data, improve the resolution of slope calculation, and ensure the accuracy of slope data for complex road sections.

[0103] In one specific embodiment of this application, the cloud platform server 101 obtains road elevation information at discrete sampling locations along the vehicle's future driving path by accessing a commercial or self-built high-precision map database. Let the road distance coordinates be... and The corresponding elevation values ​​are respectively and Then the road slope angle within this interval It is calculated using the following formula:

[0104] .

[0105] S112. Acquire longitudinal acceleration data and vehicle speed change data collected by the vehicle inertial measurement unit. Perform inversion calculation on the longitudinal acceleration data and vehicle speed change data based on the vehicle longitudinal dynamics relationship to obtain a real-time slope estimate, which is used as the real-time vehicle sensor inversion slope data.

[0106] Optionally, the aforementioned vehicle inertial measurement unit can be understood as a high-precision on-board sensor installed at the end of the target vehicle 1 to measure the vehicle's three-axis acceleration and angular velocity, which can collect longitudinal acceleration data during the vehicle's driving process in real time.

[0107] Optionally, the aforementioned vehicle speed change data can be understood as the real-time driving speed and speed change rate of the target vehicle 1, which is collected in real time by the vehicle controller and speed sensor, and can accurately characterize the speed change status during vehicle driving.

[0108] Optionally, the core principle of the above-mentioned inversion calculation based on the longitudinal dynamics of the vehicle is as follows: the measured longitudinal acceleration during the vehicle's movement is composed of the acceleration generated by the vehicle's own acceleration and deceleration and the gravitational acceleration component corresponding to the road slope; given the known rate of change of the vehicle's speed, the interference component of the vehicle's own acceleration and deceleration can be eliminated from the measured longitudinal acceleration, and the gravitational acceleration component corresponding to the road slope can be derived in reverse, finally calculating an accurate real-time slope estimate, thus completely solving the industry problem of inaccurate slope measurement during vehicle acceleration and deceleration.

[0109] In one optional embodiment of this application, the cloud platform server 101 acquires real-time vehicle operation data of the target vehicle 1's driven road segment uploaded by the vehicle-side platform device 102 via a vehicle-to-everything (V2X) communication link. This data includes longitudinal acceleration data collected by the vehicle inertial measurement unit and vehicle speed change data collected by the vehicle speed sensor. Simultaneously, it acquires auxiliary parameters such as the vehicle's real-time gear position, driving torque, and braking torque. Subsequently, the acquired raw data undergoes preprocessing such as moving average filtering and outlier removal to eliminate measurement errors caused by vehicle driving bumps and sensor noise. Based on the vehicle's longitudinal dynamic equilibrium equation, the preprocessed longitudinal acceleration data and vehicle speed change data are inverted to eliminate acceleration components caused by the vehicle's own acceleration and deceleration, extract acceleration components generated by road slope, and finally calculate the real-time slope estimate of the corresponding driven road segment, generating real-time vehicle sensor inverted slope data.

[0110] Optionally, during the inversion calculation process, the cloud platform server 101 automatically filters valid operating condition data, using only steady-state operating condition data of uniform vehicle speed and smooth acceleration / deceleration for inversion, and eliminating non-steady-state operating condition data of rapid acceleration, sudden braking, and turning, to ensure the accuracy of the inverted slope estimate. Furthermore, the cloud platform server 101 can adaptively calibrate the parameters of the inversion calculation model based on the historical driving data of the target vehicle 1, further improving the accuracy of the slope inversion.

[0111] In one specific embodiment of this application, to enhance the accuracy of slope information under complex road or map error conditions, the cloud platform server 101 uses inertial measurement unit data and speed change information collected during historical vehicle travel to perform inverse estimation of the actual road slope. In the longitudinal dynamics of the vehicle, the gravity component along the slope direction and the longitudinal acceleration measured by the vehicle satisfy the following relationship:

[0112] .

[0113] in, The road slope angle corresponding to the vehicle's current position. The acceleration value is obtained from the vehicle's longitudinal acceleration sensor. The rate of change of longitudinal speed of the vehicle can be calculated from GPS or wheel speed signals. This refers to gravitational acceleration. In this embodiment, by statistically filtering the slope inversion results of multiple historical vehicles on the same road segment, slope information that more closely reflects the actual road characteristics can be obtained.

[0114] S113. Obtain trajectory data collected when historical vehicles travel on the target road section, perform statistical analysis on the trajectory data, extract slope feature information, and generate historical vehicle trajectory statistical slope data.

[0115] Optionally, the aforementioned historical vehicle trajectory data can be understood as the real-time driving trajectory data, sensor data, and vehicle operation data uploaded by all connected vehicles within a preset time period on the target road segment currently being traveled by target vehicle 1. This data is stored in a cloud-based crowdsourced driving database, characterized by a large sample size, comprehensive scenario coverage, and long time span, effectively eliminating random errors from single samples. The aforementioned statistical analysis can be understood as a process of data cleaning, effective sample selection, and large-sample statistical fitting of massive historical trajectory data. This process can extract the true slope distribution characteristics of the target road segment from massive discrete data, eliminating data deviations caused by sensor errors of individual vehicles and abnormal operating conditions.

[0116] In one optional embodiment of this application, the cloud platform server 101 first locks the target road segment range corresponding to the future driving path of the target vehicle 1, and retrieves all historical vehicle trajectory data within the target road segment and the preset time period from the cloud crowdsourced driving database; then, it performs data cleaning and effective sample screening on the massive historical data, removing low-quality samples with missing data, abnormal operating conditions, and sensor failures, and retaining effective trajectory data and driving data under steady-state driving conditions; then, it performs large-sample statistical analysis on all effective samples, using methods such as Gaussian fitting and weighted average statistics to extract the slope feature information corresponding to each driving distance sampling point, including the slope mean, variance, and confidence interval; finally, based on the statistically obtained slope feature information, it generates historical vehicle trajectory statistical slope data that corresponds one-to-one with the driving distance domain of the target road segment.

[0117] Optionally, during statistical analysis, the cloud platform server 101 can assign weights to historical samples, allocating different weight coefficients based on data collection time, vehicle type, and sample quality. For example, samples collected more recently have higher weights, and samples of the same vehicle type as the target vehicle 1 have higher weights, further improving the compatibility of statistical data with the target vehicle 1. For road sections where road conditions have changed due to construction diversions, road resurfacing, or other reasons, the cloud platform server 101 can automatically filter historical data before the road conditions changed, using only the new samples after the changes for statistical analysis, ensuring that the statistical slope data is consistent with the actual road conditions.

[0118] In this embodiment, the vehicle cruise control method acquires data through three categories of differentiated data sources, covering three dimensions: long-distance forward-looking global information, real-time vehicle calibration information, and massive crowdsourced statistical information. This effectively solves the problems of insufficient accuracy, poor scenario adaptability, and disconnect between static data and real-time operating conditions caused by traditional single slope data sources. It provides high-quality, multi-dimensional raw data support for subsequent multi-source data fusion processing, ensuring the accuracy and reliability of subsequent global energy-saving planning from the data source end.

[0119] Optionally, the step of performing multi-source fusion calculation on the preprocessed high-precision map slope data, real-time vehicle sensor-derived slope data, and historical vehicle trajectory statistical slope data, and outputting the fused road slope sequence, includes the following steps:

[0120] S131. Assign initial weight coefficients to the high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data respectively, and the sum of the initial weight coefficients of the three types of slope data is the set value.

[0121] Optionally, the aforementioned initial weighting coefficients can be understood as quantitative coefficients assigned to each type of slope data at the start of the fusion calculation, representing the reliability of its initial data. The coefficient value ranges from 0 to 1; the larger the coefficient value, the higher the contribution of that type of data in the initial fusion calculation. The above setting is a fixed value of 1, meaning that the sum of the initial weighting coefficients of the three types of slope data is always equal to 1, ensuring the normalization of weight allocation during the fusion calculation process and avoiding deviations in the fusion results caused by weight imbalance.

[0122] In one optional embodiment of this application, the cloud platform server 101 assigns differentiated initial weight coefficients to the three types of slope data based on the driving scenario, road segment type, and sample coverage of the three types of data sources of the target vehicle 1. The core principle of the initial weight allocation is to prioritize assigning higher initial weights to the data sources with the most complete coverage and the highest data stability, while assigning appropriate basic weights to data sources with real vehicle calibration characteristics and crowdsourcing statistical characteristics.

[0123] Optionally, the initial weighting coefficients can be adaptively adjusted according to the driving scenario, without needing to be fixed to the default rules. Optionally, for road sections with drastic slope changes in mountainous areas, the spatial coverage integrity of high-precision map data is higher, and its initial weighting coefficient can be increased to 0.6-0.7; for long-distance, flat highway sections, the statistical accuracy of historical crowdsourced data is higher, and its initial weighting coefficient can be increased to 0.4-0.5; for scenarios where target vehicle 1 has traveled a long distance continuously and the sample size of real-vehicle inversion data is sufficient, the initial weighting coefficient of real-time inversion data can be increased to 0.3-0.4, thereby adapting to the data source characteristics of different scenarios.

[0124] S132. For the sampling intervals along the future driving path of the target vehicle, calculate the confidence values ​​of the three types of slope data in the corresponding sampling intervals in real time.

[0125] Optionally, the aforementioned sampling interval can be understood as a discrete interval within the driving distance domain of the target vehicle 1's future driving path, corresponding one-to-one with the sampling interval of the preprocessed slope data. Each sampling interval corresponds to a unique cumulative driving distance range and a valid sample of the three types of slope data, ensuring that the spatial position of the confidence calculation and subsequent fusion calculation corresponds one-to-one, without spatial misalignment. The aforementioned confidence value can be understood as a normalized value that quantifies the reliability and accuracy of a single type of slope data within the corresponding sampling interval, with a value range of 0-1. The closer the value is to 1, the higher the reliability and the better the data quality of that type of data within the corresponding interval.

[0126] In one optional embodiment of this application, the cloud platform server 101 calculates confidence scores for three types of slope data for each sampling interval within the driving distance domain. The specific calculation rules are as follows: For the confidence scores of high-precision map slope data, the calculation dimensions include at least the version freshness of the high-precision map data, the elevation sampling density of the corresponding sampling interval, the matching accuracy between historical data and real-vehicle collected data, and data integrity. The newer the map version, the higher the sampling density, and the higher the historical matching accuracy, the closer the confidence score is to 1. For the confidence scores of real-time vehicle sensor-derived slope data, the calculation dimensions include at least the [missing information - likely related to vehicle inertial measurement unit]. The confidence level is as follows: calibration accuracy, sample size of effective steady-state conditions within the corresponding interval, stability of vehicle driving state during data acquisition, and signal-to-noise ratio of sensor data. The higher the sensor calibration accuracy, the larger the effective sample size, and the more stable the vehicle driving state, the closer the confidence level is to 1. For the confidence level of historical vehicle trajectory slope data, the calculation dimensions should include at least the total number of historical effective samples within the corresponding sampling interval, the time span of data acquisition, the dispersion of historical slope data within the same interval, and the matching degree between sample vehicle type and target vehicle 1. The larger the historical sample size, the closer the acquisition time, the smaller the data dispersion, and the higher the vehicle type matching degree, the closer the confidence level is to 1.

[0127] Furthermore, after the cloud platform server 101 completes the single-dimensional score calculation, it obtains the final confidence value of each type of data in the corresponding sampling interval by weighted summation, and completes the normalization processing of the 0-1 interval to ensure that the confidence values ​​of the three types of data have a unified comparison benchmark.

[0128] It should be noted that the confidence score calculation is performed independently for each interval. The confidence score values ​​of the same type of data within different sampling intervals can change dynamically to match the actual data quality within the corresponding interval. Optionally, for sampling intervals with missing data or insufficient sample size, the confidence score value of the corresponding data source can be directly set to 0 to eliminate interference from invalid data sources during the fusion calculation.

[0129] S133. Based on the confidence scores obtained in real time, update the weight coefficients of the three types of slope data. Use the weighted Kalman filter algorithm to complete the fusion calculation of multi-source slope data and output the fused road slope sequence.

[0130] Optionally, the core rule for updating the weight coefficients is as follows: the higher the confidence value of a single data source, the larger the weight coefficient assigned to it, and the sum of the weight coefficients of the three types of slope data after updating is always kept at 1; if the confidence value of a single data source is 0, its weight coefficient is updated to 0 synchronously and does not participate in the fusion calculation of the corresponding sampling interval.

[0131] The weighted Kalman filter algorithm described above can be understood as an iterative optimization algorithm applicable to multi-source sensor data fusion. Its core advantage lies in its ability to combine the weight coefficients and noise variance of the data, and through the iterative process of prediction and updating, to obtain the minimum variance optimal estimate of the slope value of the corresponding sampling interval. Compared with simple weighted average fusion, it can further eliminate random noise in the data and effectively improve the fusion accuracy.

[0132] In one optional embodiment of this application, the cloud platform server 101 first adaptively updates the weight coefficients of the three types of data according to the weight update rules based on the calculated confidence values ​​of the three types of data in the corresponding sampling interval; then, based on the updated weight coefficients, it sets the observation noise matrix and weight matrix of the weighted Kalman filter algorithm, and uses the three types of slope data as the observation input of the algorithm; through the prediction step and update step of the Kalman filter, it calculates the globally optimal estimate of the slope value in the sampling interval; after traversing all sampling intervals of the future driving path of the target vehicle 1 and completing the point-by-point fusion calculation of the entire road segment, it arranges the optimal slope estimates of all sampling intervals in the order of the driving distance domain, and finally outputs a fused road slope sequence that corresponds one-to-one with the future driving path and is continuously equidistant.

[0133] It should be noted that the iterative process of the weighted Kalman filter algorithm can combine the fusion results of the previous sampling interval to achieve a smooth transition of the fusion results across the entire road segment, avoiding abrupt changes in slope values. Optionally, the cloud platform server 101 can perform moving average smoothing on the final output fused road slope sequence to further eliminate minor fluctuations in the data and ensure that the slope sequence matches the continuous slope change characteristics of the actual road. For road segments with generally low confidence levels, it can automatically mark them and send prompts to the vehicle, while simultaneously increasing the safety redundancy of subsequent planning and ensuring the stability of cruise control.

[0134] In one specific embodiment of this application, the cloud platform server 101 employs a weighted Kalman filter or an equivalent fusion method to fuse slope estimates from different sources to obtain the final slope estimate used for planning. Specifically, a weighted Kalman filter is used:

[0135] .

[0136] in, The fused road slope angle generated in the cloud at the prediction step k. It is the acceleration due to gravity. The slope estimate is obtained by inverting real-time or recent vehicle sensor data. This is a slope estimate obtained based on statistical analysis of historical vehicle trajectory data. These are weighting coefficients corresponding to various types of slope information, used to characterize the confidence level of different data sources.

[0137] Furthermore, the weighting coefficients can be adaptively adjusted based on data freshness, sensor accuracy, map reliability, or historical statistical stability, and satisfy the following:

[0138] .

[0139] In this embodiment, the vehicle cruise control method first sets a normalized initial weight benchmark, then quantizes and calculates the confidence level of the data source interval by interval, and finally dynamically updates the weights based on the confidence level and completes the optimal fusion through weighted Kalman filtering. This solves the problems of traditional fixed-weight fusion, such as inability to adapt to dynamic changes in data quality, insufficient fusion accuracy, and susceptibility to interference from low-quality data. It maximizes the advantages of the three complementary data sources and outputs a road slope sequence with high accuracy, high reliability, and full-scene adaptability, providing accurate core basic data for subsequent vehicle dynamics modeling and global energy-saving optimal trajectory planning.

[0140] Please see Figure 2 The process of generating an energy-saving reference trajectory envelope centered on the energy-saving vehicle speed trajectory based on the fused road slope sequence includes the following steps:

[0141] S21. Based on the fused road slope sequence, the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated.

[0142] Optionally, the above-mentioned energy-saving vehicle speed trajectory can be understood as a continuous vehicle speed planning sequence that achieves the optimal balance between the overall energy consumption and driving time of the vehicle within the entire driving distance domain of the future driving path of the target vehicle 1, by comprehensively matching road slope changes, vehicle power characteristics, and driving safety constraints. Each sampling point in the sequence corresponds to a unique position within the driving distance domain, as well as the globally optimal target vehicle speed corresponding to that position.

[0143] Optionally, the above target gear sequence can be understood as the optimal gear planning sequence of the vehicle transmission at each driving distance sampling point, which corresponds one-to-one with the energy-saving vehicle speed trajectory. This sequence is used to match the efficient operating range of the motor under different vehicle speeds and slopes, further reducing vehicle driving energy consumption and maximizing braking energy recovery efficiency.

[0144] In one optional embodiment of this application, the target vehicle 1 is an electric heavy-duty long-haul truck, and the driving scenario is a section of a highway from point A to point B. The cloud platform server 101 plans a speed strategy for uphill sections to accelerate in advance and use the vehicle's inertia to climb the slope, avoiding high-power drive during the climb. For long downhill sections, it plans a speed strategy for coasting in neutral gear and maximizing braking energy recovery. At the same time, it matches the corresponding optimal gear to ensure that the vehicle drive motor always works in the high-efficiency range of 85%-95%. Finally, it solves the continuous energy-saving speed trajectory and the matched target gear sequence for the entire road section.

[0145] S22. Using a preset allowable energy consumption deviation threshold as a limiting condition, and combining it with the working boundary of the vehicle power system corresponding to the target gear sequence, calculate the maximum allowable acceleration deviation and the maximum allowable deceleration deviation relative to the energy-saving vehicle speed trajectory.

[0146] Optionally, the aforementioned preset allowable energy consumption deviation threshold can be understood as the maximum allowable deviation ratio of the vehicle's actual driving energy consumption relative to the global optimal energy consumption, which is preset by the operator or the system. This serves as a hard constraint to ensure that the energy-saving effect during cruising does not fail, and can be set to values ​​such as 3%, 5%, or 8%.

[0147] Optionally, the target gear sequence is used to calibrate the transmission gears that enable the vehicle's powertrain to operate within a preset high-efficiency range at various positions along the vehicle's driving path. In this embodiment, using a preset allowable energy consumption deviation threshold as a constraint, and combining the vehicle's powertrain efficiency characteristics and driving and braking capability boundaries corresponding to the target gear sequence, the maximum allowable acceleration deviation and maximum allowable deceleration deviation relative to the optimal energy-saving speed trajectory are calculated. Furthermore, the target gear sequence generated in this step also serves as a constraint for subsequent forward behavior tree search. In the lane-changing cooperative planning step, candidate behaviors that do not meet the gear constraints are pruned. During the vehicle-side execution phase, the vehicle-side platform device 102 must synchronously match the target gear at the corresponding position within the allowable speed range to ensure the energy-saving effect during vehicle operation.

[0148] Optionally, the aforementioned maximum permissible acceleration deviation can be understood as the maximum permissible acceleration increment of the vehicle relative to the optimal energy-saving speed at the corresponding sampling point, without exceeding a preset permissible energy consumption deviation threshold, representing the upper limit to which the vehicle speed can fluctuate upwards. The aforementioned maximum permissible deceleration deviation can be understood as the maximum permissible deceleration reduction of the vehicle relative to the optimal energy-saving speed at the corresponding sampling point, without exceeding a preset permissible energy consumption deviation threshold, representing the lower limit to which the vehicle speed can fluctuate downwards.

[0149] Optionally, the cloud platform server 101, based on the obtained energy-saving vehicle speed trajectory and target gear sequence, and combined with the established vehicle longitudinal dynamics model, uses a preset allowable energy consumption deviation threshold as a constraint. Through full-condition energy consumption simulation calculation, it iterates through the changes in vehicle energy consumption under different speed deviations, and solves for the maximum allowable acceleration deviation and maximum allowable deceleration deviation corresponding to each sampling point on the future driving path. It should be noted that the maximum allowable acceleration deviation and maximum allowable deceleration deviation are not fixed uniform values ​​for the entire road section, but rather dynamically changing values ​​with road gradient, current optimal vehicle speed, and vehicle dynamic characteristics. For example, on uphill sections, vehicle acceleration leads to a significant increase in driving energy consumption, so the maximum allowable acceleration deviation at the corresponding sampling point will be correspondingly narrowed; on long downhill sections, excessive vehicle deceleration will result in a loss of braking energy recovery benefits, so the maximum allowable deceleration deviation at the corresponding sampling point will be correspondingly narrowed, ensuring that the energy consumption deviation for the entire road section is always controlled within the preset threshold.

[0150] Optionally, the cloud platform server 101 can dynamically adjust the convergence accuracy of the deviation calculation according to the road segment type. On mountainous roads with steep gradients, such as slopes and curves, the sampling density and convergence accuracy of the simulation calculation can be increased, narrowing the speed fluctuation boundary and strengthening energy-saving constraints. On long straight sections and gently sloping highways, the calculation density can be appropriately reduced, the speed fluctuation boundary can be widened, and the flexibility of vehicle-side control can be improved. In one optional embodiment of this application, the operator presets an allowable energy consumption deviation threshold of 5%. For sampling points on flat highways, the optimal energy-saving speed is 65 km / h, and the maximum allowable acceleration deviation is +8 km / h and the maximum allowable deceleration deviation is -8 km / h, obtained through simulation calculation. For sampling points on uphill sections, the optimal energy-saving speed is 55 km / h, and the maximum allowable acceleration deviation is +4 km / h and the maximum allowable deceleration deviation is -6 km / h, ensuring that the energy consumption deviation after speed fluctuation never exceeds 5%.

[0151] S23. Based on the energy-saving vehicle speed trajectory, the maximum permissible acceleration deviation, and the maximum permissible deceleration deviation, construct the permissible vehicle speed driving range, and use the permissible vehicle speed driving range as the envelope range of the energy-saving reference trajectory.

[0152] Optionally, the aforementioned permissible driving speed range can be understood as a continuous strip-shaped range consisting of the upper and lower speed limits of each sampling point within the entire distance domain of the future driving path. The upper speed limit is the sum of the energy-saving optimal speed and the maximum permissible acceleration deviation of the corresponding sampling point, and the lower speed limit is the sum of the energy-saving optimal speed and the maximum permissible deceleration deviation of the corresponding sampling point.

[0153] Optionally, the aforementioned energy-saving reference trajectory envelope is the aforementioned continuous permissible driving speed range. Its core characteristics are: with the energy-saving speed trajectory as the central baseline, the upper and lower boundaries change dynamically with the driving distance. The actual driving speed of the vehicle is within this range, which can ensure that the deviation between the energy consumption of the entire driving process and the global optimal energy consumption does not exceed the preset threshold, thereby ensuring the core goal of global energy saving, and also reserving sufficient speed adjustment space for the vehicle to cope with sudden traffic disturbances.

[0154] Optionally, for each travel distance sampling point along the future travel path, the cloud platform server 101 calculates the upper and lower speed limits of the vehicle at that point, using the optimal energy-saving speed at that point as a benchmark. The upper speed limits of all sampling points along the entire road segment are fitted into a continuous upper boundary curve, and the lower speed limits of all sampling points are fitted into a continuous lower boundary curve. Finally, the upper boundary curve, the lower boundary curve, and the central energy-saving vehicle speed trajectory together constitute a strip-shaped energy-saving reference trajectory envelope.

[0155] Optionally, the cloud platform server 101 can perform smoothing filtering on the fitted upper and lower boundary curves to avoid abrupt changes at the boundaries and ensure the smoothness of vehicle speed control. It can also set a warning threshold within the envelope interval. For example, when the deviation between the actual vehicle speed and the optimal vehicle speed reaches 80% of the maximum allowable deviation, a warning prompt is sent to the vehicle to guide the vehicle to return the speed to near the optimal trajectory, further ensuring energy-saving effect.

[0156] In one optional embodiment of this application, the target vehicle 1 is an electric heavy-duty long-haul truck. The cloud platform server 101 constructs a dynamically changing energy-saving reference trajectory envelope based on the energy-saving vehicle speed trajectory over a 200km road segment, as well as the maximum permissible acceleration deviation and maximum permissible deceleration deviation at each sampling point. When the vehicle encounters a slower vehicle ahead, the vehicle can directly decelerate and follow within the envelope without leaving the energy-saving constraint range or frequently requesting replanning from the cloud. This ensures that the energy-saving effect does not deviate from the preset target and significantly improves the cruise control's anti-disturbance capability and driving smoothness.

[0157] In this embodiment, the vehicle cruise control method first solves the global energy-saving optimal vehicle speed and gear reference, then calculates the vehicle speed floating boundary based on energy consumption hard constraints, and finally constructs the dynamic energy-saving reference trajectory envelope interval through a step-by-step execution logic. This solves the problems of poor anti-disturbance capability of a single fixed vehicle speed trajectory in traditional cruise schemes, easy deviation from the energy-saving target under traffic disturbances, and the need for frequent global replanning. Under the premise of effectively ensuring the core goal of global energy saving, it reserves sufficient control flexibility for the vehicle, thereby greatly improving the scenario adaptability and operational stability of energy-saving cruise control.

[0158] Please see Figure 3 The calculation of the energy-saving vehicle speed trajectory and the matching target gear sequence based on the fused road slope sequence includes the following steps:

[0159] S211. Based on the fused road slope sequence, establish a vehicle dynamics model including slope resistance within the driving distance domain.

[0160] Optionally, the aforementioned travel distance domain can be understood as a modeling coordinate system with the cumulative travel distance of the vehicle as the independent variable, rather than the travel time. Compared to traditional time-domain modeling, travel distance domain modeling can accurately match the distribution characteristics of road slope as the vehicle's position changes, avoiding errors caused by speed fluctuations in time-domain modeling that lead to mismatches between slope data and travel position, thus improving the accuracy of energy consumption simulation and trajectory planning.

[0161] Optionally, the aforementioned slope resistance can be understood as the driving resistance formed by the component of the vehicle's own weight along the parallel direction of the slope when the vehicle is driving on a slope. It is the core factor affecting the energy consumption of long-distance cruising. It is a positive driving resistance when going uphill and a negative driving assistance when going downhill. This step uses the fused road slope sequence as the core input of the model to achieve accurate calculation of the slope resistance of the entire road section.

[0162] Optionally, the aforementioned vehicle dynamics model can be a discrete state-space model of vehicle longitudinal dynamics in the travel distance domain, a coupled longitudinal dynamics model of electric vehicle power system, or a single-mass longitudinal dynamics model of vehicle, etc. Preferably, the vehicle dynamics model is a discrete state-space model of vehicle longitudinal dynamics in the travel distance domain, which is a digital model used to simulate the changes in force, power output, and energy consumption during the longitudinal travel of a vehicle. The discrete sampling interval of the model corresponds one-to-one with the sampling interval of the fused road slope sequence, ensuring accurate matching between the slope data and the model calculation.

[0163] Optionally, the cloud platform server 101 first uses the fused full-segment road slope sequence as the core input to lock the full distance domain range of the future driving path, sets a discrete sampling interval matching the slope sequence, and constructs a calculation coordinate system for the driving distance domain. Then, within this coordinate system, a discrete state space model of the vehicle's longitudinal dynamics, including the slope resistance of each sampling interval, is established. The braking energy recovery characteristics and coasting characteristics of the target vehicle 1 are integrated into the model simultaneously, and a comprehensive driving resistance model composed of a rolling resistance model and an air resistance model is embedded to complete the simulation framework for the full-dimensional vehicle driving force and energy consumption.

[0164] Optionally, the core parameters of the vehicle dynamics model correspond one-to-one with the actual parameters of the target vehicle 1, including the vehicle's curb weight, gross vehicle weight, frontal area, drag coefficient, rolling resistance coefficient, motor external characteristics, transmission ratio, and regenerative braking efficiency, ensuring a high degree of match between the model simulation results and actual vehicle driving conditions. The cloud platform server 101 can perform online adaptive calibration of the vehicle dynamics model parameters based on the historical driving data of the target vehicle 1. For example, based on historical vehicle energy consumption data and speed change data, it can correct key parameters such as rolling resistance coefficient and drag coefficient in real time, further improving the simulation accuracy of the model.

[0165] S212. Taking the weighted sum of the comprehensive energy consumption and driving time during the vehicle's driving process as the optimization objective, the simulation and deduction of the road segment driving conditions are completed based on the vehicle dynamics model, and the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated.

[0166] Optionally, the above-mentioned comprehensive energy consumption can be understood as the total net energy consumption of the vehicle during the entire road segment, specifically the sum of the driving energy consumption of the vehicle drive system and the fixed energy consumption of the on-board high-voltage accessories, minus the electrical energy returned to the power battery by the braking energy recovery system. It is the core quantitative indicator for measuring the energy-saving effect of cruising.

[0167] Optionally, the core of the above optimization objective is to minimize the weighted sum of the overall energy consumption and driving time throughout the entire vehicle's journey. Each of the overall energy consumption and driving time is configured with corresponding adjustable weight factors. By adjusting these weight factors, the priority of energy saving and the priority of time efficiency during cruising can be flexibly balanced. The above simulation can be understood as performing point-by-point simulation calculations of the driving conditions across the entire road segment based on the established vehicle dynamics model. It iterates through each driving distance sampling point, analyzing the changes in vehicle energy consumption and driving time corresponding to different speed and gear combinations, providing complete operating condition data support for global optimization.

[0168] In this embodiment, the cloud platform server 101 first sets adjustable energy consumption weight factors and time penalty weight factors to construct an optimization objective function with the core of minimizing the weighted sum of comprehensive energy consumption and driving time. Simultaneously, it configures multi-dimensional constraints for vehicle driving, including upper and lower speed limits corresponding to road speed limits, upper and lower acceleration limits corresponding to driving safety and ride comfort, feasible gear constraints corresponding to the vehicle transmission, and efficient operating range constraints corresponding to the drive motor. Then, based on the established vehicle dynamics model, it performs point-by-point simulations of the driving conditions for the entire road segment in the future. Through a global optimization algorithm, it completes the optimization calculation for all driving conditions, ultimately outputting a continuous energy-saving vehicle speed trajectory that corresponds one-to-one with the driving distance domain, and a target gear sequence matching each sampling point of the speed trajectory.

[0169] Optionally, the global optimization algorithm can employ dynamic programming, Pontryagin minimum principle, or other global optimization algorithms. This application does not impose any specific limitations on the algorithm type. The cloud platform server 101 can automatically adjust the weight factors of the optimization target based on the estimated arrival time requirements of the navigation system: when the estimated arrival time is ample, the energy consumption weight factor is increased to prioritize energy saving; when the estimated arrival time is tight, the time penalty weight factor is increased to prioritize meeting the trip's timeliness requirements while ensuring controllable energy consumption deviation; for ultra-long-distance cruising routes, a rolling time-domain optimization strategy can be adopted, dividing the entire path into multiple continuous planning windows for segment-by-segment solution, balancing cloud computing power consumption and planning accuracy.

[0170] In this embodiment, the vehicle cruise control method uses the driving distance domain as the modeling coordinate system, replacing the traditional time domain modeling. It uses the vehicle's cumulative driving distance as the independent variable, fully adapting to the spatial distribution characteristics of the fused road slope sequence. This eliminates the systematic error caused by speed fluctuations in the time domain modeling, which leads to mismatches between slope data and driving position. Simultaneously, slope resistance is incorporated as a core force term into the model, accurately reproducing the core impact of slope driving on vehicle power demand and energy consumption. The model simulation results highly match the actual vehicle driving conditions, avoiding the problem of inaccurate modeling leading to unsatisfactory energy-saving planning effects. Furthermore, this solution uses the weighted sum of comprehensive energy consumption and driving time as the optimization objective. Based on a high-precision dynamic model, it completes the simulation and deduction of the entire road segment's operating conditions. A global optimization algorithm is used to obtain the optimal solution for the entire journey, avoiding the short-sightedness of local optimization. While ensuring transportation timeliness, it maximizes the release of the vehicle's energy-saving potential. Furthermore, this solution simultaneously solves the energy-saving vehicle speed trajectory and the matching target gear sequence during the global optimization process. Unlike the traditional solution that plans the vehicle speed first and then matches the gear, this solution ensures that the vehicle speed and transmission gear are accurately matched at each driving position, so that the drive motor always works in the high-efficiency range. At the same time, it can fully match the slope characteristics to optimize power output and maximize the benefits of braking energy recovery. Compared with the traditional separate planning, it can further improve the energy-saving effect of long-distance cruising and reduce vehicle operating costs.

[0171] Optionally, establishing a vehicle dynamics model including gradient resistance within the driving distance domain includes:

[0172] Within the driving distance domain, a discrete state-space model of vehicle longitudinal dynamics, including gradient resistance, is constructed. This discrete state-space model integrates braking energy recovery characteristics and coasting characteristics, and includes a comprehensive driving resistance model composed of rolling resistance and air resistance.

[0173] Optionally, the cloud platform server 101 first constructs a driving distance domain calculation coordinate system and completes the discretization process to match the fused road slope sequence. The discretization process can be understood as dividing the entire distance domain range of the future driving path into several continuous discrete sampling intervals according to the sampling interval that is completely consistent with the fused road slope sequence, ensuring that the slope value, model calculation value and actual road parameters in each sampling interval correspond one-to-one.

[0174] Optionally, the cloud platform server 101 locks the driving distance range of the entire path based on the navigation information of the future driving path of the target vehicle 1; then, using the sampling interval of the fused road slope sequence as a benchmark, it discretizes the entire driving distance domain at equal intervals and constructs a calculation coordinate system with driving distance as the independent variable, ensuring that the discrete sampling points of the coordinate system correspond one-to-one with the sampling points of the road slope sequence, providing a unified calculation benchmark for subsequent model construction.

[0175] It should be noted that the discrete sampling interval can be flexibly configured according to the driving scenario. For example, a sampling interval of 5m-20m can be set for highway cruising scenarios, while a sampling interval of 1m-5m can be reduced for mountain slope scenarios, improving the simulation accuracy of the model for road sections with drastic slope changes. Optionally, for continuous slope road sections where the slope change exceeds a preset threshold, the cloud platform server 101 can automatically encrypt the sampling points of the road section, improving the resolution of the model calculation and ensuring the simulation accuracy of the slope road section.

[0176] Optionally, the vehicle longitudinal dynamics discrete state-space model is a distance-domain vehicle longitudinal dynamics discrete state-space model. This can be understood as a discretized dynamics calculation model with travel distance as the independent variable, vehicle speed as the core state variable, and vehicle driving force and braking force as control variables. It can simulate the vehicle's force state, speed changes, power output, and energy consumption changes at each travel distance sampling point. The slope resistance can be understood as the driving resistance formed by the component of the vehicle's own weight along the parallel direction of the slope when the vehicle is traveling on a slope.

[0177] In this embodiment, the cloud platform server 101, based on the constructed discrete coordinate system of driving distance domain, uses the slope value of each sampling point in the fused road slope sequence as the core input to calculate the slope resistance of the corresponding sampling point; and uses the vehicle speed as the core state variable to construct the discrete state space equation of vehicle longitudinal dynamics, complete the construction of the core model, and realize the accurate calculation of the vehicle driving force and speed change at each sampling point.

[0178] It should be noted that the core vehicle parameters of the model are matched with the actual parameters of the target vehicle 1, including the vehicle's curb weight, gross vehicle weight, gravitational acceleration, and wheel rolling radius, ensuring that the model simulation results are highly consistent with the actual vehicle driving conditions. Optionally, the cloud platform server 101 can dynamically update the vehicle gross vehicle weight parameters in the model based on the vehicle load data uploaded in real time from the vehicle terminal, adapting to different load conditions such as empty, half-loaded, and fully loaded vehicles, further improving the model's adaptability and simulation accuracy.

[0179] Optionally, the above-mentioned braking energy recovery characteristic can be understood as the characteristic of an electric vehicle switching its drive motor to generator mode during braking or coasting, converting the vehicle's kinetic energy into electrical energy to recharge the power battery. This is the core way for electric vehicles to reduce driving energy consumption. After integrating this characteristic into the model, the energy recovery power and recharge amount under different vehicle speeds and different braking intensities can be accurately simulated.

[0180] Optionally, the above-mentioned coasting characteristic in neutral gear can be understood as the coasting characteristic when the vehicle's transmission is switched to neutral, the drive motor and the transmission system are decoupled, and the vehicle is only subject to driving resistance. Compared with coasting in gear, coasting in neutral gear can eliminate the drag resistance between the motor and the transmission system, extend the coasting distance, and reduce unnecessary driving energy consumption. After integrating this characteristic into the model, the resistance difference and energy consumption difference between coasting in neutral gear and coasting in gear can be accurately simulated.

[0181] In this embodiment, the cloud platform server 101 embeds a braking energy recovery calculation module into the discrete state space model based on the motor external characteristics, braking system control logic, and transmission speed ratio characteristics of the target vehicle 1. This module clarifies the upper limit of energy recovery efficiency and the calculation logic of recovery power under different vehicle speeds and braking intensities. Simultaneously, it embeds a neutral coasting resistance calculation module to clarify the calculation formula for vehicle coasting resistance under neutral conditions. This integration of the two core features enables the model to accurately simulate changes in vehicle energy consumption under different coasting and braking conditions.

[0182] It should be noted that the parameters for regenerative braking characteristics and coasting characteristics are calibrated based on actual vehicle bench test data and road test data of the target vehicle 1, ensuring that the simulated characteristics are completely consistent with the actual operating characteristics of the vehicle. Optionally, the model can be integrated with regenerative braking efficiency correction coefficients under different ambient temperatures to adapt to the impact of changes in battery charging capacity on regenerative braking characteristics under low and high temperature environments, further improving the model's adaptability to all scenarios.

[0183] Optionally, the rolling resistance mentioned above can be understood as the rolling resistance between the tires and the road surface during vehicle operation. The air resistance mentioned above can be understood as the energy consumption to overcome air resistance during vehicle operation. The above comprehensive driving resistance model is an integration of the rolling resistance model and the air resistance model, which, together with the gradient resistance, constitutes the total longitudinal resistance of the vehicle.

[0184] In this embodiment, the cloud platform server 101 constructs a rolling resistance model and an air resistance model matching the target vehicle 1 based on the actual vehicle parameters. The two models are then integrated and embedded into the vehicle's longitudinal dynamic discrete state space model. During the model operation, the rolling resistance and air resistance at each driving distance sampling point can be calculated point by point. Combined with the synchronously calculated slope resistance, the total resistance of the vehicle is obtained, thereby accurately calculating the vehicle's power demand and corresponding energy consumption.

[0185] It should be noted that the air resistance model can simultaneously integrate wind speed and direction correction coefficients. Based on real-time wind speed and direction data obtained from the meteorological platform, the calculated air resistance values ​​are dynamically corrected, further improving the simulation accuracy of the model. Optionally, the rolling resistance model can be configured with a library of rolling resistance coefficients corresponding to different road surface types. Based on the road surface type information obtained from high-precision maps, the corresponding resistance coefficients are automatically matched to adapt to different road surface scenarios such as highways, national roads, and mountain roads.

[0186] In one specific embodiment of this application, a discrete state-space model of vehicle dynamics, including slope resistance, is established in the distance domain:

[0187] .

[0188] in, For the overall vehicle weight, Let be the longitudinal speed of the vehicle at the position or time corresponding to the distance k from the walking distance. For the longitudinal acceleration of the vehicle, Let K be the driving force output by the power system for the vehicle at a step length k. It is the acceleration due to gravity. The slope angle corresponding to the vehicle at step k is determined by the slope sequence obtained from cloud fusion.

[0189] in, The combined driving resistance related to vehicle speed, including rolling resistance and air resistance, can be expressed as:

[0190] .

[0191] in, The rolling resistance coefficient, air density, The air drag coefficient, v represents the vehicle's frontal area and v represents the vehicle speed.

[0192] In this embodiment, the vehicle cruise control method uses the driving distance domain as the modeling benchmark to construct a discrete state-space model of the vehicle's longitudinal dynamics. The cumulative driving distance is used as the independent variable, aligned with the spatial distribution characteristics of the fused road slope sequence. This avoids the inherent error in traditional time-domain modeling caused by speed fluctuations leading to mismatches between slope data and driving position. Furthermore, the standardized expression of the discrete state space is compatible with global optimization algorithms such as dynamic programming, improving the convergence and accuracy of long-distance optimization calculations. Additionally, this solution specifically integrates the braking energy recovery characteristics and neutral coasting characteristics of electric heavy-duty trucks into the model. Unlike the general dynamics model of traditional fuel vehicles, this accurately reproduces the changing patterns of bidirectional energy flow in the driving and braking systems of electric vehicles. It can accurately simulate the braking energy recovery benefits and the resistance differences between neutral and gear coasting under different speeds, slopes, and gears, solving the problem of planning results being disconnected from the optimal operating conditions of actual vehicles due to the traditional model's failure to fully consider the energy-saving characteristics of electric vehicles. Based on the global planning of this model, energy consumption optimization can be achieved by fully utilizing scenarios such as energy recovery on long downhill slopes and coasting in neutral on flat roads. Compared with traditional general models, it can further amplify the energy-saving benefits of long-distance cruising. In addition, this solution incorporates the core term of slope resistance into the model, while integrating a comprehensive driving resistance model composed of rolling resistance and air resistance models. This achieves comprehensive modeling of multiple resistance terms throughout the entire vehicle driving process, accurately reproducing the real driving resistance of the vehicle under different road surface types, vehicle speeds, and load conditions. This solves the problem of large energy consumption simulation deviations caused by incomplete resistance modeling coverage in traditional models. Whether it is a high-speed flat road, long-distance uphill and downhill slopes, or mountain curves, the model can output energy consumption simulation results that closely match the actual vehicle conditions, ensuring that the final planned vehicle speed and gear sequence have strong real-world executability.

[0193] Please see Figure 4 The optimization objective is to use the weighted sum of the vehicle's overall energy consumption and driving time as the optimization target. Based on the vehicle dynamics model, the simulation and deduction of the road segment driving conditions are completed, and the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated. This includes the following steps:

[0194] S2121. Set adjustable energy consumption weighting factors and time penalty weighting factors to construct an optimization function with the goal of minimizing the weighted sum of comprehensive energy consumption and driving time during vehicle operation.

[0195] Optionally, the aforementioned energy consumption weighting factor can be understood as a quantitative coefficient used to characterize the priority of energy-saving needs. The larger the coefficient value, the higher the priority of energy consumption control during the planning process. The aforementioned time penalty weighting factor refers to a quantitative coefficient used to characterize the priority of travel time needs. The larger the coefficient value, the higher the priority of travel time control during the planning process. Both types of weighting factors are adjustable positive numbers. Operators can manually adjust them through the vehicle-side interactive interface and cloud management platform, or the system can automatically match them according to the driving scenario.

[0196] Optionally, the aforementioned comprehensive energy consumption can be understood as the total net energy consumption of the vehicle during its journey across the entire road segment. Specifically, it is the sum of the driving energy consumption of the vehicle's drive system and the fixed energy consumption of the onboard high-voltage accessories, minus the electrical energy recharged to the power battery by the regenerative braking system. The aforementioned travel time can be understood as the total time it takes for the vehicle to complete the journey along the planned trajectory across the entire road segment, which is a core quantitative indicator for measuring travel efficiency. The aforementioned optimization function can be understood as the objective cost function used for global optimization.

[0197] In this embodiment, the cloud platform server 101 first matches the corresponding energy consumption weight factor and time penalty weight factor according to the cruise mode and navigation estimated arrival time requirements selected by the user; then it clarifies the calculation rules for comprehensive energy consumption and driving time, uses the established vehicle longitudinal dynamics model as the core calculation carrier for energy consumption and duration, and finally constructs an optimization objective function with weighted sum minimization as the core, providing a clear convergence target for subsequent global optimization.

[0198] In one optional embodiment of this application, the vehicle cruise control method provides multiple preset weight combinations for operators to quickly select, including but not limited to: an energy-saving priority mode, for example, with an energy consumption weight factor of 0.7 and a time penalty weight factor of 0.3; a balanced mode, for example, with an energy consumption weight factor of 0.5 and a time penalty weight factor of 0.5; and a time-efficiency priority mode, for example, with an energy consumption weight factor of 0.2 and a time penalty weight factor of 0.8, which can cover the needs of most long-distance cruise scenarios. Optionally, the cloud platform server 101 can automatically adjust the weight factors according to the real-time estimated arrival time of the navigation system: when the real-time estimated arrival time is earlier than the destination arrival time set by the user, the energy consumption weight factor is automatically increased to prioritize optimizing driving energy consumption; when the real-time estimated arrival time is later than the destination arrival time set by the user, the time penalty weight factor is automatically increased to prioritize ensuring the timeliness of the trip under the premise that the energy consumption deviation is controllable.

[0199] S2122. Configure driving constraints, which include at least upper and lower speed limits, upper and lower acceleration limits, and feasible gear constraints.

[0200] Optionally, the above-mentioned upper and lower speed limits can be understood as the boundary between the highest and lowest speeds allowed during vehicle operation. These are mandatory hard constraints, and their values ​​are mainly derived from the legal speed limit information of the corresponding road sections collected by high-precision maps, including the highest and lowest speed limits on highways, and special speed limit requirements for curves, tunnels, and construction sections. The boundaries can be further narrowed according to the highest cruising speed set by the operator.

[0201] Optionally, the above-mentioned acceleration upper and lower limit constraints can be understood as the boundaries of the maximum allowable acceleration and maximum braking deceleration capabilities during vehicle operation, used to ensure the smoothness of vehicle operation, passenger comfort, and cargo safety. Their values ​​are derived from the power system performance parameters and braking system performance parameters of the target vehicle 1, and can be adjusted according to the user's comfort preferences.

[0202] Optionally, the aforementioned feasible gear constraints can be understood as the range of gears that the vehicle transmission is allowed to use, and the gear matching boundaries corresponding to different vehicle speed ranges, in order to avoid problems such as vehicle speed and gear mismatch, transmission system shock, and motor leaving the high-efficiency operating range. The values ​​are derived from the speed ratio characteristics of the transmission of the target vehicle 1 and the gear and vehicle speed matching rules calibrated by bench tests.

[0203] In one optional embodiment of this application, the cloud platform server 101 first obtains the road speed limit information for the entire road segment of the future driving path of the target vehicle 1 from the high-precision map database and configures the upper and lower speed limits. Then, it retrieves the vehicle power and braking system performance parameters from the real vehicle parameter database of the target vehicle 1, and configures the upper and lower acceleration limits in combination with the user's comfort preferences. Then, it configures feasible gear constraints according to the number of gears, speed ratio characteristics and calibration rules of the vehicle's transmission. Finally, a complete multi-dimensional driving constraint system is formed as the boundary for subsequent global optimization.

[0204] Optionally, all driving constraints correspond one-to-one with discrete sampling points in the driving distance domain and can be dynamically adjusted according to the characteristics of different road sections. For example, the upper and lower speed limits are automatically narrowed in tunnels and construction sections, and the upper and lower acceleration limits are adjusted in mountainous and sloping road sections to ensure that the constraints accurately match the actual driving scenarios. Optionally, in addition to the three core types of constraints, the cloud platform server 101 can also be configured with constraints such as the high-efficiency working range of the drive motor, the charging and discharging power constraints of the power battery, and the braking intensity constraints corresponding to the road surface adhesion coefficient, to further improve the matching degree between the planning results and the actual vehicle conditions. In one specific embodiment of this application, for the high-speed cruising scenario of a 49-ton electric heavy-duty long-haul truck, the driving constraints configured on the cloud platform server 101 are as follows: the upper and lower limits of vehicle speed are 60km / h-90km / h, matching the legal speed limit of the highway section; the upper and lower limits of acceleration are -2m / s² to 1m / s², ensuring the smoothness of the heavy-duty truck's driving and the safety of the cargo; the feasible gear constraints are 1st gear and 2nd gear, matching the 2nd gear transmission installed in the vehicle, and it is also calibrated that the 2nd gear is used preferentially at speeds above 60km / h to ensure that the motor works in the high-efficiency range.

[0205] S2123. Call the intelligent simulation model of the vehicle power system, and within the boundary of the driving constraints, perform optimization through dynamic programming algorithm to obtain the energy-saving vehicle speed trajectory that matches the optimization function and the target gear sequence that matches the energy-saving vehicle speed trajectory.

[0206] Optionally, the aforementioned intelligent simulation model of the vehicle power system can be understood as a digital simulation model pre-trained based on the actual vehicle bench test data and actual vehicle road test data of the target vehicle 1. It can accurately simulate the working state, energy conversion efficiency and energy consumption level of the vehicle drive motor, transmission, power battery and braking energy recovery system under different vehicle speeds, gears, slopes, driving and braking conditions.

[0207] Further optionally, the intelligent simulation model of the vehicle power system is an intelligent simulation model of the efficiency MAP of the power system of an electric commercial vehicle, or a fully coupled digital twin simulation model of the power system of an electric heavy truck, or an intelligent simulation model of the efficiency interpolation of a lightweight power system, or an intelligent simulation model of an adaptive and self-calibrating vehicle power system, or other types of intelligent simulation models.

[0208] Optionally, the dynamic programming algorithm described above can be understood as a global optimization algorithm applicable to multi-stage decision-making processes. Its core advantage lies in its ability to break down the global planning problem of long-distance cruise into several consecutive discrete decision-making stages. By recursively solving the optimal decision for each stage, the optimal solution for the entire road segment is finally obtained, thus avoiding the problems of short-sightedness and easy entrapment in local optima in traditional local optimization algorithms.

[0209] In one optional embodiment of this application, the cloud platform server 101 first calls the pre-trained intelligent simulation model of the target vehicle 1's power system and connects it with the established discrete state space model of the vehicle's longitudinal dynamics to form a complete working condition simulation and energy consumption calculation system. Then, the entire driving distance domain of the target vehicle 1's future driving path is divided into N consecutive decision stages corresponding to discrete sampling points. Each stage uses the vehicle's driving speed as the state variable and acceleration and gear as control variables. Then, using the constructed optimization objective function as the convergence objective, within the configured driving constraint boundary, the minimum cumulative target cost is calculated by recursively tracing from the end point to the starting point of the driving path using a dynamic programming algorithm, completing the simulation deduction and optimal decision solution for the entire working condition. Finally, through forward backtracking, the optimal vehicle speed and optimal gear corresponding to each discrete sampling point of the entire road segment are obtained, ultimately generating a continuous energy-saving vehicle speed trajectory and a target gear sequence that matches the vehicle speed trajectory one by one.

[0210] Optionally, the discrete decision-making stage of the dynamic programming algorithm is consistent with the sampling interval of the driving distance domain, ensuring that the planning results accurately match the road slope sequence and constraints. For ultra-long-distance cruising routes, a rolling time-domain dynamic programming strategy can be adopted, dividing the entire path into multiple continuous rolling planning windows, solving for the optimal trajectory window by window, balancing cloud computing power consumption and planning accuracy. Optionally, the cloud platform server 101 can perform smoothing filtering on the solved vehicle speed trajectory and gear sequence to avoid frequent speed fluctuations and frequent gear shifts, further improving the smoothness of vehicle driving and reducing mechanical losses in the transmission system.

[0211] In one specific embodiment of this application, based on the discrete state-space model of vehicle longitudinal dynamics, the objective function is the combination of fuel consumption and driving time:

[0212] .

[0213] in, To comprehensively optimize the target cost, To predict the total number of steps, This is a fuel consumption weighting factor used to balance the importance of energy-saving targets. This is a time penalty weighting factor used to balance driving efficiency. The instantaneous fuel consumption rate of the vehicle at step k is calculated by the engine or powertrain energy consumption model. This represents the time interval corresponding to the step size k.

[0214] Furthermore, the aforementioned driving constraints include:

[0215] ;

[0216] Among them, V min (s)≤V(s)≤V max (s) represents the upper and lower limits of vehicle speed, a min ≤a(s)≤a max For acceleration upper and lower limit constraints, g(s)∈G is a feasible gear constraint. And, s is the cumulative distance traveled by the vehicle, V(s) is the vehicle speed at distance s, V... min (s) represents the minimum permissible vehicle speed at a distance s, V max a(s) represents the maximum permissible speed at a distance s, and a(s) represents the longitudinal acceleration of the vehicle at a distance s. a is the maximum permissible braking deceleration of the vehicle. max Let g(s) be the maximum permissible positive acceleration of the vehicle, g(s) be the gear of the vehicle's transmission at a distance s, and G be the set of feasible gears of the vehicle's transmission.

[0217] Furthermore, the energy-saving vehicle speed trajectory can be obtained through dynamic programming or equivalent numerical optimization methods. .

[0218] Furthermore, based on the aforementioned energy-saving vehicle speed trajectory, an allowable energy consumption deviation ratio is introduced. The energy-saving reference trajectory envelope is defined as:

[0219] ;

[0220] in, This refers to the energy-saving reference speed range that the vehicle is allowed to travel at position s. To achieve, without significantly increasing energy consumption, relative to Maximum permissible deceleration deviation To achieve, without significantly increasing energy consumption, relative to Maximum allowable acceleration deviation The allowable energy consumption deviation ratio is used to limit the envelope width.

[0221] Furthermore:

[0222] ;

[0223] in, This refers to the energy consumption per unit distance or unit time when the vehicle travels at speed v within a location s or the corresponding travel distance. This is the energy consumption tolerance coefficient, used to determine the acceptable range of energy consumption growth near the optimal energy-saving solution. To represent the speed deviation that minimizes the energy consumption difference, v is relative to Speed ​​deviation.

[0224] In this embodiment, the vehicle cruise control method constructs a customizable multi-objective optimization function through adjustable energy consumption weight factors and time penalty weight factors, replacing the rigid optimization objectives of traditional schemes that focus solely on energy saving or time efficiency. Users can flexibly adjust the ratio of these two types of weights based on transportation scenario attributes, estimated navigation arrival time, and vehicle remaining range. This adapts to both energy-saving operational needs such as general freight transportation and time-sensitive transportation scenarios such as cold chain and express delivery. Simultaneously, this solution defines the boundaries of the global optimization process by configuring three core driving constraints: upper and lower speed limits, upper and lower acceleration limits, and feasible gears. Specifically, the upper and lower speed limits match legal speed limits to ensure the planning results fully comply with traffic regulations; the upper and lower acceleration limits match vehicle power performance and driving smoothness requirements, avoiding additional energy consumption from rapid acceleration and deceleration while ensuring cargo safety and driving stability for heavy-duty trucks; and the feasible gear constraints match the physical characteristics of the transmission, avoiding transmission losses and smoothness issues caused by speed and gear mismatch and frequent gear shifting. Furthermore, this solution utilizes a pre-trained intelligent simulation model of the vehicle's powertrain system to accurately reproduce the real operating characteristics of the vehicle's drive motor, transmission, power battery, and regenerative braking system. It can not only accurately calculate energy consumption levels under constant-speed steady-state conditions but also precisely simulate energy flow changes during transient conditions such as acceleration / deceleration, gear shifting, and coasting on inclines, ensuring that cost calculations in the global optimization process closely match the actual energy consumption characteristics of real vehicles. In addition, this solution employs a dynamic programming algorithm to complete multi-stage global optimization across the entire road segment within preset constraint boundaries. It can combine long-distance road gradient information to pre-plan global speed strategies for uphill inertial acceleration and downhill coasting energy recovery, while simultaneously matching the target gear sequence to achieve deep collaborative optimization of speed and gear. This avoids the short-sightedness of traditional methods such as rule-based control and local model predictive control, ensuring that energy-saving effects throughout the entire journey are not sacrificed for local road condition optimization. Furthermore, the energy-saving vehicle speed trajectory and the matching target gear sequence output by this solution serve as the core benchmark for subsequent energy-saving reference trajectory envelope construction, lane changing, and speed coordination planning. This ensures that subsequent planning stages do not deviate from the core goal of overall energy saving and can quickly adapt to the energy-saving cruising needs of commercial vehicles with different models, loads, and operating scenarios, effectively reducing the energy consumption costs and range anxiety of long-distance commercial vehicle operations.

[0225] Please see Figure 5 The process of searching for the target lane-changing sequence and speed adjustment sequence, using the energy-saving reference trajectory envelope as a constraint and combining multi-lane traffic flow information with real-time vehicle status information uploaded by the target vehicle's on-board platform device, includes the following steps:

[0226] S31. Obtain multi-lane traffic flow information and real-time vehicle status information uploaded by the target vehicle's vehicle-side platform device. The real-time vehicle status information includes the vehicle's own driving status data and side and rear vehicle target information collected by the vehicle-mounted multi-source perception system.

[0227] Optionally, the aforementioned multi-lane traffic flow information can be understood as the traffic flow status and prediction information of all lanes covered by the target vehicle 1's future preset planning time domain. This information can be obtained through vehicle-road cooperative roadside equipment, crowdsourced data of connected vehicles on the same road segment, big data platform of traffic management department, and third-party traffic prediction platform. Specifically, it includes, but is not limited to, the average vehicle speed, traffic density, average following distance, location of slow-moving road section ahead, location of accident road section ahead, location of construction road section ahead, and traffic flow speed prediction data within the future preset time period for each lane.

[0228] Optionally, the aforementioned vehicle driving status data can be understood as real-time parameters representing the current driving status of the target vehicle 1. These parameters are collected and uploaded in real time by the vehicle-side platform device 102 through on-board sensors and the vehicle controller. Specifically, these parameters include, but are not limited to, the vehicle's current real-time speed, current gear, power battery SOC value, vehicle total mass, current driving position, throttle opening, brake opening, and estimated navigation arrival time.

[0229] Optionally, the aforementioned side-rear vehicle target information can be understood as traffic participant data in the area behind and to the side of the adjacent lane of the target vehicle 1, which is collected and uploaded in real time by the vehicle-mounted multi-source perception system. This data includes, but is not limited to, the relative position, relative speed, acceleration, and collision time of vehicles behind and to the vehicle in the adjacent lane. The vehicle-mounted multi-source perception system includes, but is not limited to, millimeter-wave radar, vehicle-mounted cameras, or lidar.

[0230] In one optional embodiment of this application, the cloud platform server 101 establishes a data connection with the traffic big data platform and the vehicle-road cooperative system through the vehicle network communication link to obtain real-time traffic flow information of all lanes and multiple lanes within the planning time domain of the target vehicle 1 in the next 1-3km; at the same time, it receives real-time vehicle status information uploaded by the vehicle-side platform device 102 of the target vehicle 1 at a preset frequency through the 5G / C-V2X low latency communication link, including the vehicle's own driving status data and side and rear vehicle target information; and performs standardized preprocessing such as time synchronization, spatial coordinate alignment and outlier filtering on all the acquired data to complete the preparation of input data for collaborative planning.

[0231] Furthermore, the length of the planning time domain can be flexibly configured based on cloud computing power and driving scenarios. For highway cruising scenarios, a planning time domain of 1-3 km can be set, while for urban expressway scenarios, it can be shortened to 500m-1km, balancing the overall planning perspective with real-time performance. Optionally, the cloud platform server 101 can perform short-term predictions of multi-lane traffic flow information, outputting the traffic flow change trends of each lane within the next 5-10 seconds, improving the adaptability of the planning results to dynamic changes in traffic flow. Simultaneously, it can predict the trajectories of vehicles uploaded from the vehicle's side and rear, anticipating the acceleration, deceleration, and lane-changing intentions of vehicles in adjacent lanes, further enhancing the safety of lane-changing decisions.

[0232] S32. Using the forward behavior tree search algorithm, with the energy-saving reference trajectory envelope interval as a constraint, and combining the multi-lane traffic flow information and the real-time vehicle status information, lane changing and speed coordination planning is performed.

[0233] Optionally, the aforementioned forward behavior tree search algorithm can be understood as a global path search algorithm applicable to scenarios with multiple decision branches and multiple constraints. Its core is to decompose the vehicle's driving behavior into a hierarchical behavior tree structure: the root node of the behavior tree represents the vehicle's current driving state, intermediate nodes are condition judgment nodes (e.g., whether lane-changing safety conditions are met, whether the energy-saving envelope is exceeded), and leaf nodes are executable candidate behavior nodes, including lateral and longitudinal behavior nodes. Lateral behavior nodes include changing lanes to the left, changing lanes to the right, and maintaining the current lane; longitudinal behavior nodes include accelerating, decelerating, and maintaining speed.

[0234] Optionally, forward search can be understood as starting from the root node of the vehicle's current state and expanding layer by layer along the future planning time domain to generate all feasible behavior combination branches, covering all possible combinations of lane changing and speed adjustment. Compared with traditional rule-based lane changing decisions, it has stronger global optimization capabilities and scenario adaptability.

[0235] In one optional embodiment of this application, the cloud platform server 101 first locks the planning time domain and discrete planning step size based on the acquired input data. The planning step size is consistent with the sampling interval of the aforementioned vehicle dynamics model to ensure that the planning results correspond one-to-one with the sampling points of the energy-saving reference trajectory envelope interval. Then, a behavior tree framework for lane-changing and speed-adjustment planning is constructed, and combination rules for lateral and vertical behavior nodes are set. Finally, with the energy-saving reference trajectory envelope interval as the core hard constraint, starting from the current state of the vehicle, the forward behavior tree search algorithm is used to expand layer by layer along the planning time domain to generate all candidate lane-changing behavior and speed adjustment behavior combination branches, thus completing the global deduction of the collaborative planning.

[0236] Optionally, the behavior tree unfolding process employs a rolling temporal optimization strategy, re-executing the forward search in each control cycle to update the planning results, ensuring that the planning results can adapt to dynamically changing traffic flow and vehicle status in real time. Optionally, the minimum execution interval for lane-changing behavior can be set according to the vehicle type characteristics of target vehicle 1 to avoid frequent lane changes affecting driving safety and smoothness.

[0237] S33. During the behavior tree unfolding and deduction process, if a candidate lane-changing behavior or speed adjustment behavior causes the vehicle speed to exceed the envelope range of the energy-saving reference trajectory, then the branch corresponding to the behavior will be penalized or pruned.

[0238] Optionally, the above pruning can be understood as directly removing candidate branches that do not meet the constraints during the behavior tree expansion process from the search tree, excluding them from subsequent comprehensive cost calculations and optimal solution selection, thus avoiding behaviors that violate energy-saving hard constraints, reducing invalid calculations, and lowering cloud computing power consumption. In this step, if the vehicle speed directly exceeds the upper or lower speed limit of the energy-saving reference trajectory envelope after the candidate behavior is executed, completely exceeding the preset allowable energy consumption deviation range, then direct pruning is performed on that branch.

[0239] Optionally, the aforementioned penalty can be understood as adding an additional penalty cost to the comprehensive cost calculation for candidate branches that have not completely broken through the constraint boundary but are close to it. The penalty coefficient is positively correlated with the magnitude of the vehicle speed deviation from the optimal energy-saving trajectory, guiding the algorithm to prioritize behavior branches that are closer to the global optimal energy-saving trajectory, maximizing energy-saving effects while maintaining control flexibility. In this step, if after the candidate behavior is executed, the vehicle speed is still within the envelope range, but the deviation from the optimal energy-saving trajectory exceeds a preset warning threshold, such as exceeding 80% of the maximum allowable deviation, then the corresponding cost penalty is applied to that branch.

[0240] In one optional embodiment of this application, when the cloud platform server 101 expands the deduction at each node of the behavior tree, it first simulates and calculates the final vehicle speed at the corresponding driving position after the candidate lane-changing behavior or speed adjustment behavior is executed based on the vehicle dynamics model; it then compares the simulated vehicle speed with the upper and lower limits of the energy-saving reference trajectory envelope interval corresponding to that position and executes the corresponding branch processing rules: if the vehicle speed exceeds the upper and lower limits of the envelope interval, the branch is pruned directly, and the subsequent expansion of the branch is terminated; if the vehicle speed is within the envelope interval but close to the boundary, the corresponding penalty coefficient is calculated and included in the comprehensive cost calculation of the branch; if the vehicle speed is within the envelope interval and close to the optimal trajectory, no penalty is applied, and the subsequent nodes of the branch continue to expand.

[0241] Optionally, the penalty coefficient can be dynamically adjusted based on the cruise mode selected by the user. In energy-saving priority mode, the penalty coefficient can be increased to strengthen the priority of energy-saving constraints; in time-priority mode, the penalty coefficient can be appropriately reduced to retain more behavioral flexibility. Optionally, for sudden safety scenarios, constraint exemption rules can be set. If the candidate behavior is an emergency braking or emergency avoidance behavior to avoid collision risks, the envelope interval constraint can be temporarily exempted to prioritize driving safety.

[0242] S34. Complete collaborative planning with the goal of minimizing the overall cost, and search to obtain the target lane change sequence and vehicle speed adjustment sequence. The overall cost includes the cost of vehicle speed deviating from the energy-saving vehicle speed trajectory, the cost of travel time change, and the cost of lane change safety risk.

[0243] Optionally, the aforementioned comprehensive cost can be understood as a core indicator used to quantitatively evaluate the overall performance of candidate behavioral branches, with the optimization objective being the minimum of the comprehensive cost.

[0244] Optionally, the cost of deviating from the energy-saving speed trajectory is positively correlated with the magnitude and duration of the deviation; the greater the deviation and the longer the duration, the higher the cost. The cost of changes in travel time is positively correlated with the deviation between the total travel time of the candidate planning result and the estimated arrival time of the navigation; the greater the deviation, the higher the cost. The cost of lane-changing safety risks is negatively correlated with the number of lane changes, the collision time with surrounding vehicles during lane changes, and the following distance of vehicles to the side and rear; the more lane changes and the shorter the collision time, the higher the cost. The weight coefficient of lane-changing safety risk costs is set to a basic high weight in all modes to ensure that the safety baseline is not breached.

[0245] In one optional embodiment of this application, the cloud platform server 101 calculates the cumulative comprehensive cost over the entire planning time domain for each of the valid candidate branches retained after the behavior tree is expanded; compares the cumulative comprehensive cost values ​​of all branches, and selects the branch with the smallest cumulative comprehensive cost as the global optimal solution; decomposes the behavior sequence of the optimal solution, extracts the lane-changing timing, lane-changing target lane, and number of lane changes within the planning time domain, and generates a target lane-changing sequence; at the same time, it extracts the target vehicle speed corresponding to each driving distance sampling point and generates a vehicle speed adjustment sequence that matches the lane-changing sequence one by one.

[0246] Optionally, the final output target lane-changing sequence and speed adjustment sequence ensure that the corresponding vehicle speeds remain within the energy-saving reference trajectory envelope throughout the entire process, guaranteeing that the energy consumption deviation is always controlled within a preset allowable threshold. Optionally, the cloud platform server 101 can perform smoothing filtering on the output speed adjustment sequence to avoid frequent speed jumps and ensure smooth vehicle operation; simultaneously, it can perform a secondary safety check on the lane-changing sequence to ensure that the lane-changing behavior meets the minimum safe collision time requirement throughout the entire process.

[0247] In one specific embodiment of this application, the vehicle cruise control method employs a collaborative search method with a forward behavior tree as the structure and deviation from the energy-saving reference trajectory as the pruning condition. During the behavior tree expansion process, if a candidate lane change or vehicle speed behavior causes the vehicle speed to exceed the envelope of the energy-saving reference trajectory, then that branch is penalized or pruned. Under the premise of satisfying the lane change safety distance constraint, the objective is to minimize the following comprehensive cost function:

[0248] ;

[0249] in, The combined cost of lane-changing and speed-coordinated planning. This is the squared term of the deviation of the candidate vehicle speed from the center of the energy-saving reference trajectory envelope. t represents the change in travel time caused by the behavior. For lane change safety risk indicators, ∑ represents the corresponding weight coefficients, and ∑ is the summation symbol.

[0250] Furthermore, the final output is the target lane-changing sequence and the fine-tuning speed adjustment sequence under executable constraints.

[0251] In this embodiment, the vehicle cruise control method simultaneously acquires multi-lane traffic flow information covering a long-distance forward-looking path, as well as real-time vehicle status and side / rear vehicle target information uploaded by the vehicle. This provides a global traffic flow view of several kilometers ahead for lane-changing decisions, allowing for advance prediction of slow-moving, construction, and congested road sections, and planning the globally optimal lane-changing timing and lane selection. This avoids the problems of frequent lane changes and missed optimal passage opportunities caused by traditional short-sightedness decision-making. Simultaneously, using the aforementioned generated energy-saving reference trajectory envelope as the core constraint for lane-changing and speed-coordinated planning, during the unfolding and deduction of each node in the behavior tree, it verifies in real-time whether the vehicle speed corresponding to the candidate lane-changing and speed adjustment behaviors exceeds the envelope. Branches that exceed the boundary are directly pruned and removed, while branches approaching the boundary are penalized, ensuring that all candidate solutions do not deviate from the preset energy-saving allowable range. This solves the problem that traditional lane-changing decisions lead to the failure of the global energy-saving target and a significant reduction in actual energy-saving effects. Furthermore, a comprehensive cost function is constructed that includes the cost of vehicle speed deviating from the optimal energy-saving trajectory, the cost of changes in travel time, and the cost of lane-changing safety risks. With the goal of minimizing the comprehensive cost, the energy-saving effect of long-distance cruising can be maximized without exceeding the safety limit or affecting transportation timeliness.

[0252] Optionally, the control commands sent from the cloud platform server 101 to the vehicle-side platform device 102 of the target vehicle 1 further include a graded degradation control mechanism, so that the target vehicle 1 adopts the graded degradation control mechanism to perform vehicle cruise control. The graded degradation control mechanism includes: in normal control mode, the vehicle-side platform device 102 executes the target lane change sequence and the vehicle speed adjustment sequence sent from the cloud within the energy-saving reference trajectory envelope; when the target lane change sequence cannot be executed, the vehicle-side platform device 102 degrades to the energy-saving cruise control mode, in which cruise control is only constrained by the energy-saving reference trajectory envelope; when the constraint of the energy-saving reference trajectory envelope cannot be met, the vehicle-side platform device 102 degrades to the safety fallback control mode, in which traditional adaptive cruise control is switched.

[0253] In this embodiment, the hierarchical degradation control rules are used to guide the vehicle-side platform device 102 in providing safety fallback control in emergency scenarios, ensuring the safety and continuity of vehicle operation.

[0254] In one possible embodiment, after completing global planning, the cloud platform server 101 sends control commands, including the energy-saving reference trajectory envelope, target lane-changing sequence, and speed adjustment sequence, to the on-board domain controller of the electric heavy-duty truck via a 5G network. Upon receiving the commands, the on-board domain controller executes the lane-changing and speed adjustment commands within the energy-saving reference trajectory envelope under normal driving conditions. When a sudden accident ahead prevents the lane-changing sequence from being executed, it automatically downgrades to energy-saving cruise control mode, only performing longitudinal speed control under the envelope constraint. When an extreme emergency occurs and the envelope constraint cannot be met, it automatically downgrades to traditional adaptive cruise control to ensure driving safety. Throughout this process, the cloud completes global energy-saving planning and collaborative decision-making, while the vehicle completes execution and safety safeguards, achieving both globally optimal energy-saving effects and ensuring driving safety and anti-disturbance capabilities.

[0255] Optionally, after the control command, which includes the energy-saving reference trajectory envelope interval, the target lane change sequence, and the vehicle speed adjustment sequence, is sent to the vehicle-side platform device 102 of the target vehicle 1 so that the vehicle-side platform device 102 performs vehicle cruise control based on the received control command, the method further includes: the cloud platform server 101, according to the real-time driving position of the target vehicle 1, repeatedly performs the operations of acquiring road slope information, generating energy-saving reference trajectory envelope interval, generating target lane change sequence and vehicle speed adjustment sequence according to a preset period or event triggering method, and sends the updated control command to the vehicle-side platform device 102 of the target vehicle 1 to realize closed-loop optimization control.

[0256] Please see Figure 6 and Figure 7This application also provides a vehicle cruise control system 100, including a cloud platform server 101 and a vehicle-end platform device 102 connected to the cloud platform server 101 via vehicle network communication, wherein the cloud platform server 101 is used to execute the steps in the vehicle cruise control method provided in the embodiments of this application.

[0257] Optionally, the cloud platform server 101 is the global planning and decision-making core of the system, used to execute all the steps of the vehicle cruise control method described in any of the foregoing embodiments of this application. The vehicle-side platform device 102 is the front-end data acquisition terminal and control execution terminal of the system, used to upload the vehicle and environmental data required for planning to the cloud platform server 101 in real time, receive the control commands issued by the cloud platform server 101, and complete the implementation of vehicle cruise control based on the commands.

[0258] Optionally, the cloud platform server 101 is deployed on the vehicle network cloud platform and has the capabilities of high-performance parallel computing, massive data storage, multi-source data interaction, and global path planning. It can provide parallel cruise planning and decision-making services for multiple connected vehicles at the same time.

[0259] Optionally, the vehicle-side platform device 102 is installed on the target vehicle 1 and electrically connected to the vehicle's bus, on-board sensing system, and actuator system, possessing the capabilities of data acquisition, communication interaction, vehicle control, and safety backup. The vehicle-side platform device 102 includes at least a vehicle-side communication interaction module, an on-board sensing module, a vehicle control unit, and vehicle actuator components. These modules are interconnected via bidirectional communication through on-board Ethernet and CAN bus to collaboratively complete vehicle-side data acquisition and control execution.

[0260] Optionally, the vehicle-side communication interaction module serves as the communication hub between the vehicle and the cloud. It adopts an in-vehicle communication unit that supports 5G, C-V2X, and 4G LTE, and has low latency and high reliability bidirectional communication capabilities. It is used to upload real-time data collected by the in-vehicle perception module and the vehicle control unit to the cloud platform server 101, and at the same time receive cruise control commands issued by the cloud platform server 101 and forward them to the vehicle control unit.

[0261] Optionally, the vehicle-mounted perception module is the core of the vehicle's environment and status perception, including vehicle-mounted inertial measurement unit, vehicle speed sensor, millimeter-wave radar, vehicle-mounted camera, lidar and other vehicle-mounted sensors, used to collect real-time vehicle driving status data, longitudinal acceleration data, vehicle speed change data, and vehicle target information in the surrounding area and to the side and rear of the vehicle, providing real-time vehicle and environmental data input for cloud planning.

[0262] Optionally, the vehicle control unit is the core of vehicle-side control decision-making, including an autonomous driving domain controller, a vehicle controller, a motor controller, and a brake controller. It is used to receive cruise control commands sent from the cloud, analyze the energy-saving reference trajectory envelope, the target lane-changing sequence, and the vehicle speed adjustment sequence, and generate corresponding longitudinal and lateral control signals for the vehicle. At the same time, it has built-in hierarchical degradation control rules, which can automatically switch between normal control mode, energy-saving cruise control mode, and safety fallback control mode according to real-time road conditions and vehicle status to ensure driving safety and control continuity.

[0263] Further optionally, the vehicle actuator component is a vehicle-side control execution terminal, including a drive motor, an electric power steering system, and a hydraulic braking system, used to receive control signals from the vehicle control unit, execute corresponding drive acceleration, braking deceleration, and steering lane-changing actions, and ultimately realize the cruise control strategy planned in the cloud.

[0264] In this embodiment, by employing a collaborative architecture between the cloud platform server 101 and the vehicle-side platform device 102, the high-computing-power aspects of global energy-saving planning and lane-changing collaborative decision-making are moved to the cloud. This addresses the issues of insufficient local computing power, limited planning scope, and inability to achieve optimal global energy saving on the vehicle side. The vehicle side focuses on data acquisition and execution control, while incorporating a built-in tiered degradation safety fallback mechanism. The overall system architecture is adaptable to long-distance cruising scenarios for various vehicle types, including commercial vehicles and passenger cars, demonstrating strong scenario adaptability and effectively balancing overall energy saving and driving safety.

[0265] In this application, the terms "embodiment" and "implementation" mean that a specific feature, structure, or characteristic described in connection with an embodiment can be included in at least one embodiment of this application. The appearance of these phrases in various locations throughout the specification does not necessarily refer to the same embodiment, nor are they independent or alternative embodiments mutually exclusive with other embodiments. Those skilled in the art will understand, explicitly and implicitly, that the embodiments described in this application can be combined with other embodiments. Furthermore, it should be understood that the features, structures, or characteristics described in the various embodiments of this application can be arbitrarily combined to form another embodiment that does not depart from the spirit and scope of the technical solution of this application, provided there is no contradiction between them.

[0266] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the above preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of this application should not depart from the spirit and scope of the technical solutions of this application.

Claims

1. A vehicle cruise control method, characterized in that, include: Obtain multi-source slope data of the future driving path of the target vehicle, and perform fusion processing on the multi-source slope data to obtain the fused road slope sequence; Based on the fused road slope sequence, an energy-saving reference trajectory envelope interval centered on the energy-saving vehicle speed trajectory is generated; Using the energy-saving reference trajectory envelope as a constraint, and combining multi-lane traffic flow information with real-time vehicle status information uploaded by the target vehicle's on-board platform device, the target lane-changing sequence and speed adjustment sequence are searched. The control command, which includes the energy-saving reference trajectory envelope, the target lane change sequence, and the vehicle speed adjustment sequence, is sent to the vehicle-side platform device of the target vehicle, so that the vehicle-side platform device can perform vehicle cruise control based on the received control command; The step of generating an energy-saving reference trajectory envelope centered on the energy-saving vehicle speed trajectory based on the fused road slope sequence includes: calculating the energy-saving vehicle speed trajectory and the matching target gear sequence based on the fused road slope sequence; calculating the maximum allowable acceleration deviation and the maximum allowable deceleration deviation relative to the energy-saving vehicle speed trajectory using a preset allowable energy consumption deviation threshold as a limiting condition, combined with the vehicle power system working boundary corresponding to the target gear sequence; constructing a vehicle speed allowable driving range based on the energy-saving vehicle speed trajectory, the maximum allowable acceleration deviation, and the maximum allowable deceleration deviation, and using the vehicle speed allowable driving range as the energy-saving reference trajectory envelope; The process of searching for target lane-changing sequences and speed adjustment sequences, constrained by the energy-saving reference trajectory envelope and combined with multi-lane traffic flow information and real-time vehicle status information uploaded by the target vehicle's on-board platform device, includes: acquiring multi-lane traffic flow information and real-time vehicle status information uploaded by the target vehicle's on-board platform device, wherein the real-time vehicle status information includes the vehicle's own driving status data and side and rear vehicle target information collected by the onboard multi-source perception system; employing a forward behavior tree search algorithm, constrained by the energy-saving reference trajectory envelope and combined with the multi-lane traffic flow information and the real-time vehicle status information, to perform lane-changing and speed coordination planning; during the behavior tree unfolding and deduction process, if a candidate lane-changing behavior or speed adjustment behavior causes the vehicle's driving speed to exceed the energy-saving reference trajectory envelope, then the branch corresponding to that behavior is penalized or pruned; completing the coordination planning with minimizing the comprehensive cost as the optimization objective, and searching for target lane-changing sequences and speed adjustment sequences, wherein the comprehensive cost includes the cost of vehicle speed deviating from the energy-saving speed trajectory, the cost of travel time changes, and the cost of lane-changing safety risks. The process of acquiring multi-source slope data for the future driving path of the target vehicle and fusing the multi-source slope data to obtain a fused road slope sequence includes: classifying and acquiring multi-source slope data corresponding to the future driving path of the target vehicle, wherein the multi-source slope data includes at least high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data; performing preprocessing on the high-precision map slope data, the real-time vehicle sensor inversion slope data, and the historical vehicle trajectory statistical slope data respectively; performing multi-source fusion calculation on the preprocessed high-precision map slope data, real-time vehicle sensor inversion slope data, and historical vehicle trajectory statistical slope data, and outputting the fused road slope sequence.

2. The vehicle cruise control method as described in claim 1, characterized in that, The calculation of the energy-saving vehicle speed trajectory and the matching target gear sequence based on the fused road slope sequence includes: Based on the fused road slope sequence, a vehicle dynamics model including slope resistance is established within the driving distance domain; Using the weighted sum of the vehicle's overall energy consumption and driving time as the optimization objective, the vehicle dynamics model is used to simulate and extrapolate the driving conditions of the road segment, and the energy-saving vehicle speed trajectory and the matching target gear sequence are calculated.

3. The vehicle cruise control method as described in claim 2, characterized in that, The establishment of a vehicle dynamics model including gradient resistance within the driving distance domain includes: Within the driving distance domain, a discrete state-space model of vehicle longitudinal dynamics, including gradient resistance, is constructed. This discrete state-space model integrates braking energy recovery characteristics and coasting characteristics, and includes a comprehensive driving resistance model composed of rolling resistance and air resistance.

4. The vehicle cruise control method as described in claim 2, characterized in that, The optimization objective is a weighted sum of the vehicle's overall energy consumption and driving time. Based on the vehicle dynamics model, simulations of road segment driving conditions are performed to calculate the energy-saving speed trajectory and the matching target gear sequence, including: Set adjustable energy consumption weighting factors and time penalty weighting factors, and construct an optimization function with the goal of minimizing the weighted sum of comprehensive energy consumption and driving time during vehicle operation. Configure driving constraints, which include at least upper and lower limits for vehicle speed, upper and lower limits for acceleration, and feasible gear constraints. By calling the intelligent simulation model of the vehicle power system, within the boundaries of the driving constraints, a dynamic programming algorithm is used to find the optimal energy-saving vehicle speed trajectory that matches the optimization function and the target gear sequence that matches the energy-saving vehicle speed trajectory.

5. The vehicle cruise control method as described in claim 1, characterized in that, The classification process obtains multi-source slope data corresponding to the future driving path of the target vehicle, including: Access a high-precision map database to obtain discrete elevation data on the future driving path of the target vehicle, and calculate the slope data of the corresponding road segment based on the elevation difference and distance difference between adjacent sampling points, which is used as the slope data of the high-precision map. The longitudinal acceleration data and vehicle speed change data collected by the vehicle inertial measurement unit are acquired. Based on the longitudinal dynamics of the vehicle, the longitudinal acceleration data and the vehicle speed change data are inverted and calculated to obtain a real-time slope estimate, which is used as the real-time vehicle sensor inverted slope data. Historical vehicle trajectory data collected when driving on the target road segment is obtained, the trajectory data is statistically analyzed, and slope feature information is extracted to generate historical vehicle trajectory statistical slope data.

6. The vehicle cruise control method as described in claim 5, characterized in that, The process involves multi-source fusion calculation of preprocessed high-precision map slope data, real-time vehicle sensor-derived slope data, and historical vehicle trajectory statistical slope data, outputting a fused road slope sequence, including: Initial weight coefficients are assigned to the slope data of high-precision map, the slope data inverted by real-time vehicle sensors, and the slope data statistically obtained from historical vehicle trajectories, and the sum of the initial weight coefficients of the three types of slope data is a set value. For the sampling interval along the future driving path of the target vehicle, the confidence level of the three types of slope data in the corresponding sampling interval is calculated in real time; The weight coefficients of the three types of slope data are updated based on the confidence scores obtained in real time. The weighted Kalman filter algorithm is used to complete the fusion calculation of multi-source slope data and output the fused road slope sequence.

7. A vehicle cruise control system, characterized in that, It includes a cloud platform server and a vehicle-side platform device connected to the cloud platform server via vehicle-to-everything (V2X) communication, wherein the cloud platform server is used to perform the steps in the method as described in any one of claims 1 to 6.

Images