Vehicle path planning method based on traffic signals and related equipment

Abstract

The embodiment of the invention provides a traffic signal-based vehicle path planning method and related equipment, and belongs to the technical field of automatic driving. The method comprises the following steps: acquiring traffic light data of each intersection of a target driving path; according to the green light time interval of the traffic light data, time constraint calculation of multi-intersection continuous passing is carried out, and a time constraint of continuous green wave passing is obtained; performing speed calculation on each road section according to the time constraint to obtain a speed interval of each road section; performing quadratic programming on the speed interval by taking the energy consumption and the driving time as optimization objectives to obtain a global speed sequence; and performing multi-dimensional constraint and multi-target optimization on a speed controller of the vehicle according to the global speed sequence so as to control the vehicle to run on the target running path. According to the embodiment of the invention, the vehicle can be guided to realize continuous green light passing, and the driving efficiency and the driving comfort are improved while the driving energy consumption of the vehicle is reduced.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a vehicle path planning method and related equipment based on traffic signals. Background Technology

[0002] With the acceleration of urbanization, traffic congestion and environmental pollution are becoming increasingly serious problems. Autonomous vehicles, as a key component of future transportation systems, play a crucial role in improving traffic efficiency, ensuring driving safety, and reducing energy consumption through trajectory planning technology. Traditional trajectory planning methods often ignore the impact of traffic signals, leading to frequent vehicle starts and stops, increasing energy consumption and passenger discomfort. Currently, the traffic environment is becoming increasingly complex, requiring vehicles to cope with changing traffic lights, traffic density, and road conditions. Related technologies have limitations in terms of energy efficiency and comfort. Furthermore, the lack of a comprehensive framework in assessing the impact of trajectory on energy consumption and comfort makes it difficult to fully optimize vehicle behavior.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main objective of this application is to propose a vehicle route planning method and related equipment based on traffic signals, which can guide vehicles to pass through green lights continuously, reduce vehicle energy consumption, and improve driving efficiency and driving comfort.

[0005] To achieve the above objectives, one aspect of this application proposes a vehicle route planning method and related equipment based on traffic signals, the method comprising: Obtain traffic light data at each intersection along the target driving route; Based on the green light time interval of the traffic signal data, the time constraint for continuous passage through multiple intersections is calculated to obtain the time constraint for continuous green wave passage. Based on the time constraint, the speed of each road segment is calculated to obtain the speed range of each road segment; With energy consumption and travel time as optimization objectives, a secondary planning is performed on the speed range to obtain a global speed sequence; The vehicle's speed controller is subjected to multi-dimensional constraints and multi-objective optimization based on the global speed sequence in order to control the vehicle to travel on the target driving path.

[0006] In some embodiments, the step of calculating the time constraint for continuous passage through multiple intersections based on the green light time of the traffic signal data to obtain the continuous green wave passage constraint includes: The target travel path is divided according to the intersection, and multiple different road segments are determined; With the goal of the time when a vehicle arrives at each intersection within the green light time interval, the travel time of each road segment is calculated to obtain the passage time interval of each road segment, and the time constraint includes the passage time interval.

[0007] In some embodiments, the step of calculating the speed of each road segment according to the time constraint to obtain the speed range of each road segment includes: Obtain the first speed range of the previous road segment; Speed ​​is calculated based on the first speed range, the length of the current road segment, and the time constraint to obtain the speed range of the current road segment.

[0008] In some embodiments, the step of performing secondary planning on the speed range with energy consumption and travel time as optimization objectives to obtain a global speed sequence includes: The energy consumption under different speed configurations is calculated based on the speed range and the preset vehicle energy consumption model to obtain the total energy consumption of the vehicle. The vehicle's travel time is calculated based on the speed range and the length of each road segment. Based on the total energy consumption and the travel time, a cost function is constructed to obtain the total cost objective function; The global velocity sequence is obtained by minimizing the total cost objective function using a nonlinear constraint optimization method.

[0009] In some embodiments, the vehicle energy consumption model is obtained through the following steps: Based on the vehicle's parameter configuration, a longitudinal dynamics model of the vehicle is obtained by modeling using Newton's second law. The energy consumption model is constructed based on the vehicle's longitudinal dynamics model to obtain the vehicle energy consumption model.

[0010] In some embodiments, the step of performing multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence includes: Obtain real-time vehicle driving parameters; Real-time energy consumption, speed deviation, jerk and real-time position are determined based on the preset vehicle longitudinal dynamics model, the global velocity sequence and the real-time driving parameters, respectively. The speed controller is optimized based on the real-time energy consumption, the speed deviation, the jerk, and the real-time position. The speed controller is subject to multi-dimensional constraints based on actuator constraints, kinematic constraints, and spatiotemporal constraints related to the green light.

[0011] In some embodiments, when a traffic jam is detected ahead of the vehicle, the method further includes: Obtain the length of the congested queue, the traffic density at different positions in the queue, and the traffic flow. The wave velocity of the traffic wave of the queue is obtained by calculating the wave velocity based on the traffic density and the traffic flow. The propagation time of the traffic wave is calculated based on the queue length and the wave speed. The time is modified according to the propagation duration and the actual green light time interval to obtain the virtual time interval of the virtual green light; The speed controller is subject to additional constraints based on the virtual time interval.

[0012] To achieve the above objectives, another aspect of this application proposes a vehicle route planning device based on traffic signals, the device comprising: The acquisition module is used to acquire traffic light data at each intersection along the target driving route; The time planning module is used to calculate the time constraint for continuous passage through multiple intersections based on the green light time interval of the traffic signal data, and obtain the time constraint for continuous green wave passage. The speed planning module is used to calculate the speed of each road segment according to the time constraint, and obtain the speed range of each road segment. The optimization module is used to perform secondary planning on the speed range with energy consumption and travel time as optimization objectives to obtain a global speed sequence; The execution module is used to perform multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence, so as to control the vehicle to travel on the target driving path.

[0013] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0014] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0015] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method.

[0016] The embodiments of this application include at least the following beneficial effects: This application provides a vehicle route planning method, device, electronic device, storage medium, and program product based on traffic signals. The time constraint of this scheme serves as the basis for continuous green light passage of vehicles. It can then enumerate all speed ranges that can be continuously passed, combine energy consumption and travel time, and select a global speed sequence that achieves a good balance between energy consumption and travel time from the speed range. This can reduce the vehicle's driving energy consumption while improving driving efficiency. Based on the global speed sequence, the speed controller is constrained to control the vehicle, enabling the vehicle to drive at a stable speed and improving driving comfort. Attached Figure Description

[0017] Figure 1 This is a flowchart of a vehicle routing method based on traffic signals provided in an embodiment of this application; Figure 2 This is a schematic diagram of traffic and road information provided in an embodiment of this application; Figure 3 This is a schematic diagram of the speed range for green wave traffic provided in the embodiments of this application; Figure 4 This is a three-dimensional cost graph of the speed range provided in the embodiments of this application; Figure 5 This is a two-dimensional cost graph of the speed range provided in the embodiments of this application; Figure 6 This is a schematic diagram of the SQP algorithm principle provided in the embodiments of this application; Figure 7 This is a schematic diagram of green wave velocity calculation provided in an embodiment of this application; Figure 8 This is a schematic diagram of the green window channel for continuous traffic signal operation provided in the embodiments of this application; Figure 9 This is a schematic diagram illustrating the relationship between flow rate and density in a vehicle queue, provided in an embodiment of this application. Figure 10 This is a schematic diagram of traffic queue passage at a traffic light intersection provided in an embodiment of this application; Figure 11 This is a schematic diagram of the virtual traffic light design provided in an embodiment of this application; Figure 12 This is a schematic diagram of a simple road network provided in an embodiment of this application; Figure 13 This is a schematic diagram of vehicle driving conditions provided in the embodiments of this application. Figure 14 This is a complete implementation flowchart of a vehicle route planning method based on traffic signals provided in this application embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] In some related technologies, the Intelligent Driver Model (IDM) is a representative vehicle trajectory planning model, widely used as a microscopic traffic flow model for vehicle trajectory planning. This model dynamically adjusts the vehicle's acceleration behavior based on the relative speed and distance between the vehicle and the vehicle in front. However, because the IDM model uses fixed parameters and does not integrate traffic light information, its applicability in real urban traffic environments is limited. It cannot adapt to dynamic changes in traffic lights: The IDM model does not consider the impact of traffic light states (such as red and green light durations and phase shifts) on vehicle behavior. In actual traffic, traffic light changes directly cause vehicles to decelerate or accelerate, and the fixed parameter mechanism of the IDM cannot respond to such external events, resulting in significant trajectory planning deviations. It ignores green wave opportunities: The IDM model focuses on car-following behavior and does not fully utilize the cooperative passage potential between consecutive intersections (such as green wave zones), failing to guide vehicles to achieve continuous green light passage between multiple intersections, thus increasing unnecessary waiting times and energy consumption. Relying on local sensor data: IDM makes decisions based solely on information about vehicles ahead obtained from onboard sensors, lacking interaction with traffic signal systems (such as V2I) and unable to obtain global traffic conditions (such as traffic light timing and intersection queuing), resulting in limited planning capabilities in complex multi-intersection scenarios.

[0021] Among other related technologies, the traditional Green Light Optimal Speed ​​Advisory (GLOSA) system aims to suggest that vehicles travel at a specific speed to pass through traffic light intersections during the green light period. This system typically calculates the recommended speed based on the timing of the traffic lights at a single intersection (cycle, green light duration, phase shift) and provides speed suggestions to drivers through onboard terminals or roadside units. Although this technology has some effect in single-intersection scenarios, it has three limitations: (1) Limitations of single-intersection optimization: The traditional GLOSA system only optimizes for a single traffic light and does not consider the cooperative relationship between multiple consecutive intersections. After a vehicle passes through the first intersection, the traffic light status of subsequent intersections may not be able to maintain green wave passage due to dynamic changes in traffic flow, resulting in vehicles still needing to frequently start and stop. (2) Static modeling ignores traffic flow dynamics: The system treats traffic lights as fixed spatiotemporal points and does not consider dynamic phenomena such as vehicle queuing and queue dissipation in traffic flow. In reality, the queue of vehicles ahead will change the effective green light window, which means that the speed recommended by the static model cannot achieve green wave passage in reality. (3) Insufficient comfort: The simplified "acceleration-uniform speed-deceleration" speed curve is used, which causes the acceleration to change abruptly during the stage transition, causing discomfort to the passengers and does not meet the comfort requirements of advanced autonomous driving.

[0022] In view of this, this application provides a vehicle route planning method and related equipment based on traffic signals. Figure 1 This is an optional flowchart of a vehicle routing method based on traffic signals provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0023] Step S101: Obtain traffic light data at each intersection along the target driving route; Step S102: Calculate the time constraint for continuous passage through multiple intersections based on the green light time interval of the traffic signal data to obtain the time constraint for continuous green wave passage. Step S103: Calculate the speed of each road segment according to the time constraint to obtain the speed range of each road segment; Step S104: Using energy consumption and travel time as optimization objectives, perform secondary planning on the speed range to obtain the global speed sequence; Step S105: Perform multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence to control the vehicle to travel on the target driving path.

[0024] Steps S101 to S106 as shown in the embodiments of this application use time constraints as the basis for continuous green light passage of vehicles. Then, all speed ranges that can be continuously passed can be listed. By combining energy consumption and travel time, a global speed sequence that achieves a good balance between energy consumption and travel time is selected from the speed ranges. This can reduce the vehicle's driving energy consumption, reduce travel time, and improve driving efficiency. Based on the global speed sequence, the speed controller's control of the vehicle is constrained, enabling the vehicle to drive at a stable speed and improving driving comfort.

[0025] In step S101 of some embodiments, the traffic signal data includes data such as the traffic signal cycle, green light duration, and phase offset, which can be used to generate time constraints for vehicles passing through each intersection. In step S102 of some embodiments, to achieve trajectory planning that takes traffic signals into account, this embodiment defines the basic rules and constraints for green wave passage. Based on parameters such as the cycle of traffic lights, green light duration, and phase offset, it is clarified that the arrival time of vehicles at each intersection must meet the conditions for green light passage, that is, the arrival time should fall within the green light interval, thus laying the framework for subsequent speed planning.

[0026] In some embodiments, step S102 may include, but is not limited to, steps S201 to S202: Step S201: Divide the target driving route according to the intersection and determine multiple different road segments; Step S202: Taking the time when the vehicle arrives at each intersection within the green light time interval as the objective, calculate the travel time of each road segment to obtain the passage time interval of each road segment. The time constraint includes the passage time interval.

[0027] In steps S201 to S202 of some embodiments, in the green wave traffic problem under continuous traffic lights, the "General-Green Light Optimal Speed ​​Advisory" (G-GLOSA) model aims to set constraints for subsequent speed optimization algorithms. The entire target travel path is divided into several continuous segments according to the location of the traffic light intersections. Each segment consists of an acceleration (or deceleration) interval followed by a cruising interval. The core constraint of the rule is that the time when a vehicle arrives at each intersection must fall within the time interval of the green light phase of that intersection's traffic light (green light time interval). This time interval is determined by the fixed cycle of the traffic light, the duration of the green light, and the phase offset relative to the starting point of the path.

[0028] For example, a traffic light model is first established for a single intersection. The traffic light state can be either green-cycle priority or red-cycle priority, depending on whether the traffic light is green or red at the initial moment. The total cycle of the traffic light is T, during which the green light lasts for... The red light continues The phase offset from the origin of the time axis is Based on this model, it can be extended to traffic models for multiple consecutive intersections. and These represent the time and speed at which the vehicle arrives at the i-th intersection, respectively. and This refers to the time and speed at which the vehicle begins its journey. Taking the green light cycle as an example, for a vehicle to encounter a green light at the i-th intersection, the time required for the vehicle to arrive at that intersection is... Specific conditions must be met, as shown in formulas (1) and (2).

[0029] (1) (2) in, This represents the total cycle of the traffic lights at the i-th intersection. This represents the duration of the green light at the i-th intersection. This represents the phase offset at the i-th intersection. This indicates the time when the vehicle arrives at the i-th intersection. This indicates the time it takes for the vehicle to arrive at the (i-1)th intersection. It is a small positive number representing the minimum time interval between vehicles arriving at consecutive intersections, used to avoid time conflicts. This represents the number of the traffic light cycle at the i-th intersection, i.e., the sequence number of the traffic light cycle when the vehicle arrives.

[0030] Similarly, when a vehicle reaches the i-th intersection, and the current traffic light cycle is red, If equation (3) is satisfied, it is possible to pass through the i-th traffic light cycle.

[0031] (3) When the vehicle is at the i-th intersection The numbering of the traffic light cycle is shown in equation (4).

[0032] (4) The core idea of ​​the G-GLOSA model is to divide the vehicle's travel path into multiple segments according to the traffic light positions, with each segment having a set target speed. In each segment, the vehicle first adjusts its speed to the target speed within the acceleration range. Then, this speed is maintained within the cruising range until the end of the segment; this speed also applies when passing the i-th intersection. Considering that green wave traffic aims to reduce stops at traffic lights, the G-GLOSA model assumes that vehicle acceleration is small and the duration of the acceleration phase in each segment is fixed, denoted as . This assumption helps simplify the driving model, while The specific value will vary depending on the vehicle's acceleration performance and the road speed limit. In this way, the G-GLOSA model provides an effective speed planning framework for green wave traffic.

[0033] The vehicle maintains a constant speed for the remaining road segment, based on the time it takes to reach the i-th intersection. This allows us to obtain the time taken for the cruising section. The acceleration (deceleration) of the i-th road segment's acceleration (deceleration) range is also given. As shown in equation (5).

[0034] (5) in, This indicates the target speed for the previous road segment.

[0035] Distance of acceleration (deceleration) interval As shown in equation (6).

[0036] (6) The travel time of each road segment, consisting of acceleration (deceleration) interval time and cruising interval time. As shown in equation (7).

[0037] (7) in, Indicates the length of the road segment.

[0038] Combining equations (6) and (7), we can obtain equation (8).

[0039] (8) If the traffic light is green when the vehicle arrives at each intersection, the travel time of each road segment is added together to obtain the arrival time of the i-th intersection, as shown in equation (9).

[0040] (9) in, Indicates the initial time of vehicle departure.

[0041] The green window interval selected at the i-th intersection is denoted as [ , According to the green wave traffic rules, it is known that... Combining equation (9), we can obtain equation (10).

[0042] (10) In step S103 of some embodiments, the speed range corresponding to the green window passage is further calculated based on the constraints of green wave traffic. For each road segment, the feasible range of cruising speed for that road segment is derived by combining the segment length, the cruising speed of the previous segment, and the green window interval of the current intersection. If the calculated result exceeds the road speed limit, the speed limit value is used as the boundary of the speed range to ensure the legality of vehicle driving. This step provides a feasible range of alternative speeds for subsequent optimization.

[0043] In some embodiments, step S103 may include, but is not limited to, steps S301 to S302: Step S301: Obtain the first speed range of the previous road segment. Step S302: Calculate the speed based on the first speed range, the length of the current road segment, and the time constraint to obtain the speed range of the current road segment.

[0044] In some embodiments, steps S301 to S302 are used to determine the speed of passing through the selected green window at the i-th intersection. Substituting equation (8) into equation (10) within the range of equation (11) yields equation (11).

[0045] (11) in, The length of the current road segment. This is the first speed range.

[0046] In equation (11), the denominator of the fraction must be greater than 0, as shown in equations (12) and (13).

[0047] (12) (13) in The range of values ​​is determined by the first... i The green window interval at -1 intersection is used to determine the location. The lower limit of the green window section at the previous intersection , The upper limit of the green window section at the previous intersection The upper and lower limits of the selected green window interval are determined by the intersection of equations (1) and (11). Therefore, the selection of the green window under the G-GLOSA model needs to satisfy the following conditions: and Without considering speed limits, the first iThe first intersection j Each green window interval is a subset of the green light cycle of the traffic lights, as shown in equation (14).

[0048] (14) Assume a driving route is divided into N According to the G-GLOSA model, the speed at which each traffic light arrives is as shown in equation (15).

[0049] (15) To facilitate visualization of the results, we take a path that passes through two traffic lights consecutively as an example, illustrating the green wave traffic flow achieved under the G-GLOSA model. The road parameters are shown in Table 1. Based on these parameters, a traffic light signal diagram for the experimental road was drawn. To narrow the effective solution range, a minimum speed limit of 4 m / s was added. Figure 2 As shown, from Figure 2 As you can see, there are two green windows at the first traffic light and three green windows to choose from at the second traffic light.

[0050] Table 1 Road Parameters

[0051] At the initial moment, The initial velocity is given. The solution obtained from the G-GLOSA model is... Figure 3 The horizontal axis represents the speed at which vehicles pass through the first intersection. The vertical axis represents the speed at which vehicles pass through the second intersection. The green areas correspond to situations where no stops were made at either the first or second intersection.

[0052] In step S104 of some embodiments, after obtaining all feasible green window channels (i.e., all possible cruise speed sequences), the optimal channel needs to be selected. This embodiment designs a green window channel selection algorithm. This algorithm uses energy consumption and travel time as optimization objectives, constructs a cost function incorporating both, and solves it using a Sequential Quadratic Programming (SQP) algorithm to select the optimal solution from the feasible green window channels and determine the cruise speed sequence for each road segment. By reasonably setting the weights of energy consumption and time, a balance between energy saving and efficiency is achieved.

[0053] In some embodiments, step S104 may include, but is not limited to, steps S401 to S404: Step S401: Calculate the energy consumption under different speed configurations based on the speed range and the preset vehicle energy consumption model to obtain the total energy consumption of the vehicle. Step S402: Calculate the travel time of the vehicle based on the speed range and the length of each road segment; Step S403: Construct a cost function based on total energy consumption and travel time to obtain the total cost objective function; Step S404: The total cost objective function is solved by minimizing the cost using a nonlinear constraint optimization method to obtain the global velocity sequence.

[0054] In steps S401 to S404 of some embodiments, different speed configurations (speed ranges) are set for the same driving path based on the constructed electric vehicle model and the G-GLOSA model. Calculate energy consumption. The speed range determines whether the vehicle needs to stop at the traffic light. For example... Figure 3 As shown, the velocity regions corresponding to different green windows are marked in green, and these are selected as the research objects. This is based on the objective function (total cost objective function). Given the road speed limit constraints, we perform optimization on these areas. The total cost objective function is shown in Equation (16), and the road speed limit constraints are shown in Equation (17).

[0055] (16) (17) In formula (16), This represents the energy consumption of the path, while Representing travel time, both parameters can be adjusted based on speed. Sure. and This is used as a weighting coefficient. As described in the previous example, each path segment includes two intervals: an acceleration interval and a cruising interval. i The energy consumption of the road segment during acceleration or deceleration is shown in equation (18).

[0056] (18) in, This can be obtained from the vehicle energy consumption model.

[0057] During the acceleration (or deceleration) range, the vehicle's motion can be approximated as uniformly accelerated (or uniformly decelerated) linear motion. The length of the cruising range can be calculated using formula (6). Within the cruising range of each road segment, the vehicle maintains a constant cruising speed and zero acceleration until the end of the segment. According to formula (8), the cruising time can be expressed as... The energy consumed within this cruising range is denoted as... Specifically, as shown in equation (19).

[0058] (19) When calculating the total energy consumption, the energy consumption during the cruising and acceleration phases needs to be added together. Based on formulas (18) and (19), the following can be derived: i The energy consumption cost of the first road segment, i.e., the energy consumption cost of the first road segment i The total energy consumption of each road segment consists of two parts: energy consumption during the acceleration phase and energy consumption during the cruising phase. Therefore, the energy consumption of the first segment is obtained. i The energy cost of each road segment is expressed as Specifically, as shown in equation (20).

[0059] (20) pass Equation (16) can be obtained. .

[0060] Based on the energy consumption calculation method and G-GLOSA model described above, the cost within the solution domain was calculated, such as... Figure 4 As shown. Reuse Figure 3 The green wave velocity is used to find the minimum cost within the velocity domain. ,like Figure 5 As shown.

[0061] In summary, by using the total cost objective function of equation (16) and combining it with the constraints of the green window channel, a solution is obtained. N dimensional vector In addressing the optimization problem of continuous traffic light paths, the SQP algorithm is employed to optimize all green window channels. The SQP algorithm is an efficient nonlinear constrained optimization method. In each iteration step, it constructs and solves a quadratic programming subproblem to determine the search direction, and updates the iteration points through a one-dimensional search until convergence to the optimal solution of the nonlinear programming problem, such as... Figure 6 As shown.

[0062] In some embodiments, the vehicle energy consumption model can be implemented through steps S410 to S420: Step S410: Based on the vehicle's parameter configuration, a model is constructed using Newton's second law to obtain the vehicle's longitudinal dynamics model; Step S420: Construct an energy consumption model based on the vehicle's longitudinal dynamics model to obtain the vehicle energy consumption model.

[0063] In steps S410 to S420 of some embodiments, relevant vehicle parameters are first defined to construct a longitudinal dynamics model for the pure electric vehicle. This model, based on Newton's second law, quantifies the relationship between driving force, braking force, rolling resistance, air resistance, and acceleration during vehicle operation. The driving force is determined by motor torque, transmission ratio, transmission efficiency, and wheel radius. Rolling resistance considers the influence of road slope and rolling resistance coefficient. Air resistance is related to air density, drag coefficient, frontal area, and the square of vehicle speed, enabling the model to accurately reflect the longitudinal motion characteristics of the vehicle.

[0064] According to Newton's second law, the longitudinal dynamics model of the vehicle is shown in equation (21).

[0065] ,(twenty one) in, Indicates the longitudinal speed of the vehicle. Indicates the total mass of the vehicle. Indicates braking force. The driving force can be expressed through transmission efficiency. transmission ratio Motor torque and wheel radius The result is shown in equation (22).

[0066] ,(twenty two) For slope resistance, it is determined by the road slope angle. The decision is as shown in equation (23).

[0067] ,(twenty three) in, It is the gravitational constant. This is the rolling resistance coefficient.

[0068] The air resistance of a vehicle can be determined by air density. drag coefficient The frontal area of ​​the vehicle The specific details are shown in equation (24).

[0069] ,(twenty four) For example, the vehicle parameters are shown in Table 2. Table 2 Vehicle Parameter Table

[0070] Based on the vehicle's longitudinal dynamics model, a vehicle energy consumption model is constructed. This model employs a two-dimensional polynomial fitting approach to establish the mapping relationship between motor power, motor speed, and torque. The polynomial coefficients are calibrated using real vehicle data, enabling the model to accurately calculate energy consumption under different driving conditions. Energy consumption calculations are divided into acceleration and cruising phases, with power and time accumulated for each phase to achieve a precise assessment of the vehicle's overall energy consumption, providing a quantitative basis for subsequent energy-saving optimization.

[0071] The vehicle energy consumption model uses a two-dimensional polynomial fitting to represent the motor power, as shown in equation (25).

[0072] (25) in These are the fitting coefficients. This refers to the motor speed. The values ​​of the motor torque are shown in Table 3.

[0073] Table 3. Fitted numerical table

[0074] motor speed Due to transmission ratio Wheel radius and vehicle speed The specific details are shown in equation (26).

[0075] (26) In some embodiments, traffic state maps exhibit different characteristics under different operating conditions due to variations in road speed limits and traffic light timing cycles. An actual driving route often includes multiple road sections with traffic lights, significantly increasing the number of green window lanes and making it difficult to determine their exact number through visual observation. This embodiment proposes a statistical method for G-GLOSA green window lanes, accurately calculating the upper and lower speed limits for each green window lane based on different speed limit requirements and traffic light timing cycle characteristics.

[0076] At each intersection, the number of green windows that allow passage varies; these are denoted as follows: .like Figure 7 As shown, the vehicle starts from the beginning of the road segment at the... The green light begins at the start of the cycle. The first average speed at each traffic light intersection is denoted as . The calculation formula is shown in equation (27).

[0077] (27) Among them, Represents the traffic light intersection number. To sequence the traffic light cycles, It is the current moment. For vehicles and the The distance is equal to the distance between several intersections. Similarly, the distance from the starting point of the road segment is equal to the distance between several intersections. The initial time of the red light cycle reaches the [number]th cycle. The second average speed at each traffic light intersection is denoted as... The calculation formula is shown in equation (28).

[0078] (28) To ensure that the vehicle speed is non-negative, it is specified as shown in equation (29).

[0079] (29) Therefore, it can be concluded that at the traffic light number j The green window passing speed range for each cycle is: If this section coincides with the lane speed limit section intersection If the set is non-empty, then That is, the traffic light number The green window interval for each cycle is shown in Equation (30).

[0080] (30) This leads to the first A collection of green windows at each intersection . No. The green window speed range at each intersection is affected by the arrival of vehicles at the first intersection. The time difference at each intersection, assuming the vehicle's acceleration and deceleration intervals... The magnitude is negligible, and each road segment is considered to be moving at a constant speed. The initial moment of the green window where the vehicle is located is selected. Find the upper limit of the green window channel.

[0081] Based on the obtained traffic light cycle and speed limit information, the steps for selecting the green window channel according to the above method are as follows: When the vehicle is in its initial position, the set of available green windows at the first intersection is denoted as: ; When choosing When a green window is used as the first intersection for passage, the set of optional green windows for the second intersection to be passed is: ; This method continues until the first path is selected. intersection This completes the selection process for a green window channel.

[0082] Similarly, all feasible green window passages along the vehicle's current path can be obtained, but the number is too large to list them all here. To improve the efficiency of solving for green window passages, this embodiment only selects passage from the first two green windows and introduces constraints. This significantly narrowed down the range of options available through the green window.

[0083] It should be noted that the selected green window It is determined by the non-empty set obtained from formula (30). Obviously, the number of green windows in different green window channels is not the same, mainly depending on the distance from the starting intersection to the... The green window selected at each intersection, therefore, the above embodiment adopts... To represent the green window.

[0084] When determining the green window for passage, it should start from the first green window. The initial time of the vehicle's current green window is calculated according to formula (30), thereby clarifying the upper limit of the green window interval. Similarly, using the reverse method, starting from the end green window of the selected green window channel, the lower limit of each green window interval is obtained. In this way, all green window channels can be obtained. Finally, using the energy-saving green window channel selection scheme introduced above, the optimal selection is made from many green window channels based on the cost function, such as... Figure 8 As shown.

[0085] In step S105 of some embodiments, to transform the planned speed sequence into actual vehicle control commands, this embodiment constructs a speed optimization controller framework based on MPC (Model Predictive Control). This framework establishes a discretized vehicle state-space model, using displacement and velocity as state variables and driving force and braking force as control variables. It employs a rolling optimization mechanism, solving for the optimal control sequence in the prediction time domain at each sampling time. Only the first control variable is executed, and re-optimization is performed at the next time step, resulting in good adaptability to disturbances during driving.

[0086] In some embodiments, step S105 may include, but is not limited to, steps S501 to S504: Step S501: Obtain real-time vehicle driving parameters; Step S502: Determine the real-time energy consumption, speed deviation, jerk and real-time position based on the preset vehicle longitudinal dynamics model, global speed sequence and real-time driving parameters respectively; Step S503: Perform multi-objective optimization of the speed controller based on real-time energy consumption, speed deviation, jerk and real-time position; Step S504: Apply multi-dimensional constraints to the speed controller based on actuator constraints, kinematic constraints, and spatiotemporal constraints related to the green light.

[0087] In steps S501 to S504 of some embodiments, the core of the MPC controller (speed controller) lies in the design of its multi-objective optimization function and constraints. The optimization function comprehensively considers energy consumption, speed tracking accuracy, comfort, and position tracking accuracy, achieving a balance among multiple objectives by setting different weights. Simultaneously, multi-dimensional constraints are applied to the speed controller based on actuator physical constraints, kinematic constraints, and green window constraints, ensuring that the vehicle's control commands remain within a safe and legal range, satisfying both energy-saving and comfort requirements while allowing the vehicle to pass through intersections according to the green wave time constraint.

[0088] For example, the control variable is set as The design of the MPC controller includes key technical aspects such as system modeling, construction of multi-objective optimization functions, constraint design, and real-time solution strategies.

[0089] (1) System dynamics modeling and discretization: Considering the longitudinal dynamic characteristics (based on the vehicle longitudinal dynamics model), as described in the previous embodiment, a continuous time state space model is established, as shown in equation (31).

[0090] (31) in, For displacement, For speed, For the total vehicle mass. Resistance term. and As shown in equations (32) and (33).

[0091] (32) (33) Discretization is performed using the forward Euler method, with a time step of . The prediction model is shown in Equation (34), which can predict the driving parameters at the next moment based on the current real-time driving parameters.

[0092] (34) (2) Construction of multi-objective optimization function: The comprehensive cost function in the prediction time domain is shown in equation (35).

[0093] (35) in, This represents the comprehensive cost function in the prediction time domain. This indicates the number of steps in the prediction time domain. This represents the weighting coefficient for the energy consumption item. This represents the energy consumption value at time step k. This represents the weighting coefficient of the speed tracking term. This represents the actual speed of the vehicle at time step k. Indicates reference speed. This represents the weighting coefficient for the comfort item. This represents the jerk at time step k, i.e., the rate of change of acceleration. This represents the weighting coefficient of the location tracking term. This indicates the actual position of the vehicle at time step k. Indicates a reference position.

[0094] The energy consumption item adopts a data-driven motor power model, as shown in equation (36).

[0095] (36) in, This refers to the motor speed. This represents the motor torque.

[0096] The speed tracking term introduces a global reference speed (global speed sequence). Specifically, as shown in equation (37).

[0097] (37) The comfort factor is based on jerk, as shown in equation (38).

[0098] (38) Position tracking item, based on the spatiotemporal reference trajectory generated by SQP. Tracking is performed using piecewise linear interpolation, as shown in equation (39).

[0099] (39) (3) Multi-dimensional constraint design: Three types of constraints are applied in the prediction time domain, including actuator physical constraints (actuator constraints), kinematic constraints, and green window spatiotemporal constraints.

[0100] The physical constraints of the actuator are as shown in equation (40).

[0101] (40) The kinematic constraints are as shown in equation (41).

[0102] (41) The specific spatiotemporal constraints of the green window are shown in equation (42).

[0103] (42) In some embodiments, when a traffic jam is detected in front of the vehicle, the method may also include, but is not limited to, steps S601 to S605: Step S601: Obtain the length of the congested queue, the traffic density at different positions in the queue, and the traffic flow. Step S602: Calculate the wave velocity based on traffic density and traffic flow to obtain the wave velocity of the traffic wave in the queue. Step S603: Calculate the propagation time based on the queue length and wave speed to obtain the propagation time of the traffic wave; Step S604: Modify the time according to the propagation duration and the actual green light time interval to obtain the virtual time interval of the virtual green light; Step S605: Apply additional constraints to the speed controller based on the virtual time interval.

[0104] In steps S601 to S605 of some embodiments, traffic congestion often occurs on roads equipped with traffic lights, leading to vehicle queues and reduced traffic efficiency. During speed planning, to balance safety and comfort, a queue model needs to be introduced to accurately characterize congestion conditions. Based on the LWR (Lighthill-Whitham-Richard) traffic flow theory, the vehicle queue characteristics at traffic light intersections are quantified, providing constraints for congestion scenarios. Through sensors such as roadside units, the number of vehicles currently waiting at traffic lights on the road can be collected in real time. Combined with pre-set vehicle length and inter-vehicle gap parameters, the queue length can be accurately calculated, providing crucial congestion status information for subsequent speed optimization.

[0105] This embodiment comprehensively considers the phase and cycle of traffic lights, as well as traffic queues under congestion conditions, and introduces virtual traffic light technology. The queue length is linked to the location of the virtual traffic light, and its signal phase and cycle correspond to the formation and dissipation time of the queue. Subsequently, the virtual traffic light information is used as an inequality constraint in the speed optimization problem to achieve optimized traffic flow management.

[0106] use The traffic flow model describes the dynamics of vehicle queues at signalized intersections, and its continuity equation is shown in equation (43).

[0107] (43) In the formula, For position With time Traffic density and flow rate at the location. Wave velocity is calculated based on the difference in traffic conditions between upstream and downstream areas, as shown in equation (44).

[0108] (44) and For upstream regional flow and density, and For the downstream region's flow rate and density. Flow rate and density have a typical concave functional relationship, such as... Figure 9 As shown, changes in traffic light status can trigger the generation and propagation of traffic waves.

[0109] Changes in traffic light status cause convoy waves and starting waves to form in traffic queues, thus affecting the movement of vehicles. When a traffic queue arrives at an intersection and the traffic light is red, a convoy wave is generated at the front of the queue. Subsequently, when the traffic light turns green, a starting wave is formed.

[0110] When the traffic light turns red, vehicles stop and form a rallying wave. The upstream traffic state changes from free flow (state A) to congested flow (state C). The wave speed of the rallying wave is shown in equation (45).

[0111] (45) in, Traffic flow before the red light For the initial density, This represents congestion density. After the traffic light turns green, the queue begins to dissipate, forming a starting wave, and the state changes from congestion flow (state). ) transforms into saturated flow (state) The wave velocity of the starting wave is as shown in equation (46).

[0112] (46) In the formula, This refers to the saturation flow rate after the green light. It is the saturation density. When the accretion wave meets the initiation wave, a deviation wave is generated. Its wave velocity is determined by the difference in state between the upstream and downstream, as shown in equation (47).

[0113] (47) The propagation characteristics of the traffic waves directly affect the movement of vehicles within the queue. The queue length is obtained, denoted as... Due to vehicle queues, the actual waiting time will be longer than the red light countdown. When the traffic light turns green, a starting wave will be generated in the queue. The time required for the starting wave to travel from the front to the back of the queue is shown in equation (48).

[0114] (48) The time required from the arrival of the initiating wave at the end of the queue to the complete dissipation of the queue is The calculation formula (49) is shown.

[0115] (49) By integrating actual traffic lights with queue status, a virtual traffic light sequence is generated to solve the problem of green wave traffic constraint failure under congestion conditions. The introduced virtual traffic light concept, based on traffic flow and queue characteristics, and using data such as queue length, number of vehicles, road parameters, and traffic light sequence, designs virtual traffic light sequence and location parameters suitable for the current road segment.

[0116] The queue completely dissipates during the next green light cycle, as... Figure 11 As shown. The queue reaches its maximum length and the flow is zero when it is formed. The moment the queue is formed is denoted as... ,satisfy The moment the traffic light turns green is recorded as follows: The moment when the traffic light turns red is recorded as At this point, a starting wave begins to appear in the queue, with a wave speed of... The moment the initiation wave reaches the tail of the queue is It can be obtained through equation (50) (50) Time of arrival of the last vehicle in the queue at the intersection The calculation is shown in equation (51).

[0117] (51) in, This is the current intersection number.

[0118] To ensure the starting wave reaches the end of the queue before the controlled vehicle arrives, the actual traffic light position needs to be virtually moved to the end of the queue. Then, the initial green light time of the virtual traffic light cycle needs to be determined. and the initial moment of the red light As shown in equations (52) and (53).

[0119] (52) (53) Depend on Figure 10 It can be seen that the positional deviation between the virtual traffic light and the actual traffic light is... To improve driving safety, the buffer distance is increased. Take half the length of the controlled vehicle body. Specifically, as shown in equation (54).

[0120] (54) The virtual traffic light estimation method integrates the actual traffic light status and queue length information, and uses the green window interval of the virtual traffic light as a new constraint for the MPC controller, improving the applicability of the solution in complex traffic environments. Through the organic combination of the above series of technical steps, this embodiment forms a complete energy-saving and comfortable trajectory planning scheme for autonomous vehicles that considers traffic signals, realizing energy-saving, comfortable, and efficient driving of vehicles in continuous traffic light scenarios.

[0121] To simulate a real urban traffic environment, this embodiment uses SUMO software to construct a road network and traffic flow model. SUMO is an open-source microscopic traffic simulation tool capable of accurately simulating vehicle behavior, traffic light control, and traffic congestion. First, real-world traffic data, such as road geometry, traffic light cycles, and traffic density, are integrated to create a multi-vehicle road network model in SUMO. Then, SUMO is connected to MATLAB via the TraCI interface to form a co-simulation platform. TraCI allows MATLAB to read the traffic light status, vehicle positions, and speeds from SUMO in real time and provide feedback control commands.

[0122] The MPC speed optimization algorithm was applied and verified in a joint simulation traffic scenario. Figure 10 A simple road network was set up on SUMO. Vehicles started from road 1 and traveled along path 1, 2, 3, 4, 5, passing intersections 1, 2, 3, 4, and 5 in sequence. A co-simulation model was used to obtain the traffic light timing sequence for this path and the distribution of each traffic light relative to the vehicle's initial position. Specific measurements showed that the distances of the five traffic lights from the vehicle were 300 meters, 600 meters, 900 meters, 1200 meters, and 1800 meters, respectively. During the speed optimization solution process, to narrow the solution range and improve computational efficiency, a minimum speed limit of 3.5 m / s (i.e., 12.6 km / h) was set based on the road speed limit.

[0123] Based on the defined road speed limits and vehicle current positions, and according to the energy-saving green window lane selection scheme, green window constraints are determined. Subsequently, using a time-domain-based MPC speed planning algorithm, the driving curve for green wave traffic is calculated by considering both energy consumption and comfort, as shown below. Figure 12 As shown in the figure, the MPC speed optimization scheme effectively solves the problem of frequent starts and stops during vehicle operation, generating a relatively smooth and more comfortable reference speed curve.

[0124] To verify the advantages of the methods described in the above embodiments in terms of comfort and economy, a comparative experiment was conducted between the MPC-GLOSA speed optimization control model and the IDM following model built into SUMO. The IDM model comprehensively considers factors such as the vehicle's desired speed, distance to the vehicle in front, and relative speed, and calculates the acceleration of the following vehicle through mathematical formulas, thereby simulating the vehicle's acceleration and deceleration. Through joint simulation using MATLAB and SUMO, under the conditions of six consecutive traffic lights and a road length of 2100m, traffic flows of 200 vehicles / h, 300 vehicles / h, 400 vehicles / h, 500 vehicles / h, and 600 vehicles / h were set. The control results of the MPC-GLOSA speed optimization control model and the IDM following model are shown in Tables 4 and 5, respectively, including records of maximum acceleration, maximum jerk, average acceleration, energy consumption, and time consumption data. The energy consumption and average acceleration of the vehicle controlled by the MPC-GLOSA speed optimization control model are much lower than those of the vehicle controlled by the IDM following model. The MPC-GLOSA speed optimization control model reduces energy consumption by an average of 52%, significantly reduces the three acceleration indicators, and keeps the time almost the same, resulting in a significant improvement in both economy and comfort.

[0125] Table 4. Control Results of the MPC-GLOSA Model under Six Consecutive Traffic Lights

[0126] Table 5. Control Results of the IDM Model under Six Consecutive Traffic Lights

[0127] In some embodiments, please refer to Figure 13 The first step is to establish a longitudinal dynamics model for the electric vehicle. Traffic flow simulation is performed using SUMO (Simulation of Urban Mobility) software. Since vehicles in SUMO are based on cellular automata models and lack dynamic attributes, the relevant vehicle parameters are first defined to construct a longitudinal dynamics model for the pure electric vehicle. This model is based on Newton's second law; specifically, it quantifies the dynamic relationship between driving force, braking force, rolling resistance, air resistance, and vehicle acceleration. The calculation of driving force integrates multiple factors, including motor output torque, the overall transmission ratio and efficiency of the transmission system, and the wheel rolling radius. The modeling of rolling resistance considers both the road slope angle and the rolling resistance coefficient between the tire and the road surface. The calculation of air resistance incorporates parameters such as air density, vehicle drag coefficient, frontal projected area, and the square of the current speed. This model accurately simulates the acceleration, deceleration, and cruising behavior of the vehicle at different slopes and speeds, laying the foundation for subsequent speed optimization planning and tracking control.

[0128] The second step involves constructing a data-driven energy consumption model for electric vehicles. Building upon the established longitudinal dynamics model, to further quantify the impact of different driving behaviors on energy consumption and achieve energy-saving optimization goals, a computational model that accurately reflects the energy consumption characteristics of electric vehicles needs to be constructed. This model employs a two-dimensional polynomial fitting method based on real-vehicle operating data to establish a nonlinear mapping relationship between motor output power and real-time motor speed and output torque. The fitting coefficients of the polynomials in the model are all obtained through calibration using a large amount of test data, improving the model's accuracy. This high-precision energy consumption model can achieve real-time and accurate evaluation of the vehicle's energy consumption under different operating conditions, providing a quantitative basis for green window channel selection and multi-objective optimization control aimed at energy saving.

[0129] The third step is to define the generalized green wave traffic rules and constraints applicable to consecutive intersections. To ensure vehicles pass through green light intervals as uninterruptedly as possible when crossing multiple consecutive traffic light intersections (i.e., achieving "green wave traffic"), a clear set of traffic rules and constraints is needed. In the green wave traffic problem under consecutive traffic lights, the "Generalized Green Light Optimal Speed ​​Advisory" (G-GLOSA) model aims to set constraints for subsequent speed optimization algorithms. The core constraint of the rule is that the time a vehicle arrives at each intersection must fall within the time interval of the green light phase of that intersection's traffic light. This time interval is determined by the fixed cycle of the traffic light, the duration of the green light, and the phase offset relative to the starting point of the path. To simplify the model, it is assumed that the acceleration process of each road segment lasts for a fixed time interval. Based on this, the mathematical relationship between the vehicle's travel time through each road segment, the time to arrive at each intersection, and the cruising speed of each road segment is derived using kinematic formulas. This transforms the goal of green wave traffic into constraints on a series of speed values, providing a clear mathematical framework and constraint boundaries for subsequent optimization algorithms.

[0130] The fourth step is to calculate the green window lanes and their feasible speed intervals for each road segment. Under the constraints defined by the G-GLOSA rules, for each segment, it is necessary to calculate all possible cruising speeds that allow the vehicle to encounter a green light at the next intersection—that is, the feasible speed interval. This calculation requires considering multiple parameters, including the length of the current road segment, the cruising speed value derived from the previous segment, and the green light time interval at the current target intersection. By solving a series of time-constrained inequalities, the upper and lower limits of the cruising speed for the current road segment can be derived. If the calculated upper speed limit exceeds the legal speed limit of the road, the road speed limit must be used as the actual upper limit of the speed interval to ensure the legality of the planned speed. Each feasible speed interval represents a potential passage plan (i.e., a "green window"), and the process of a vehicle traversing the entire path is equivalent to sequentially selecting a green window for each road segment. All these selections combined form a complete "green window lane." The purpose of this step is to calculate all feasible green windows (speed intervals) for each road segment, thereby enumerating all possible green window lanes and providing the necessary input options for global optimization.

[0131] The fifth step involves designing an optimal green window lane selection algorithm oriented towards energy conservation. After obtaining all feasible green window lanes (i.e., all possible cruising speed sequences), the optimal lane needs to be selected. This embodiment designs a decision-making algorithm with total energy consumption and total travel time as common optimization objectives. The algorithm constructs a multi-objective cost function, which is a linear combination of the weighted sum of energy consumption and the weighted sum of time. The energy consumption term is accurately calculated using the energy consumption model constructed in the second step, while the time term is obtained by accumulating the travel time of each road segment. The cost function is solved using the efficient nonlinear constrained optimization algorithm Sequential Quadratic Programming (SQP). The SQP algorithm iteratively solves a series of quadratic programming subproblems, gradually approaching the optimal solution, and finally selects the optimal cruising speed sequence from all feasible green window lanes that achieves the best balance between energy consumption and travel time. This algorithm effectively combines high-level traffic strategies (which window to choose) with low-level vehicle energy consumption characteristics (how much energy is consumed), achieving global energy conservation optimization.

[0132] Step 6: Design an efficient statistical algorithm for green window channels in multi-intersection scenarios. In real urban roads, a path often contains multiple traffic light intersections, which leads to an exponential increase in the number of feasible solutions for green window channels, resulting in a "combinatorial explosion" problem. To address this challenge, this embodiment proposes an efficient method for managing and calculating green window channels. The algorithm dynamically calculates the upper and lower speed limits for each intersection within a potential green window channel based on the periodic characteristics and phase shift of each traffic light at each intersection, as well as the road's speed limit requirements. Constraint propagation and pruning techniques are employed: if the selection of a current intersection results in no feasible speed range (i.e., a green window) for a subsequent intersection, further calculation of that branch is terminated early. This method reduces the number of green window channels requiring precise calculation and optimization, solves the problem of high computational complexity in multi-intersection scenarios, and improves the real-time performance of the optimization algorithm.

[0133] Step 7: Construct an MPC speed tracking controller based on time-domain rolling optimization. After obtaining the globally optimal cruise speed sequence through the preceding steps, it is necessary to transform the planned speed sequence into actual vehicle control commands and handle various disturbances encountered during real-time driving. This embodiment uses a Model Predictive Control (MPC) framework to design this speed tracking controller. First, based on the longitudinal dynamics model from Step 1, a discretized state-space prediction model is established with vehicle displacement and velocity as state variables and driving force and braking force as control variables. The core of the controller is an online iterative finite-time domain optimization problem, whose optimization function comprehensively considers four key objectives: energy consumption, speed tracking accuracy, comfort, and position tracking accuracy. Simultaneously, the optimization problem is also subject to multiple constraints, including the physical limits of actuators (motors, brakes), the limits of vehicle kinematics (maximum speed, maximum acceleration), and the green window time constraint that must be satisfied, ensuring that the vehicle's control commands are within a safe and legal range, meeting both energy-saving and comfort requirements while allowing the vehicle to pass through intersections according to the green wave time constraint. In each control cycle, the MPC controller re-solves the optimization problem based on the current vehicle state, executes the optimal control command for the current moment, and then re-optimizes the solution based on the new state in the next moment.

[0134] Step 8: Establish a dynamic model of traffic flow queues at intersections. On roads with traffic lights, vehicle queues often lead to traffic congestion and reduced traffic efficiency. In speed planning, to balance safety and comfort, a queue model needs to be introduced to accurately depict congestion conditions. Considering the impact of congestion in the planning, this embodiment introduces the Lighthill-Whitham-Richard (LWR) model from traffic flow theory to dynamically describe the formation and dissipation of queues at intersections. The LWR model correlates traffic density and flow rate through a continuity equation. By analyzing the propagation speed and direction of traffic waves (such as rallies, initiation waves, and dissipation waves) generated when traffic light states change (red to green, green to red), the start time of queue formation, the time when the queue reaches its maximum length, and the time when it dissipates can be derived. Combined with real-time information such as the number of queued vehicles, average vehicle length, and vehicle distance obtained through roadside sensors (such as cameras and radar) or vehicle-to-everything (V2X) communication, the current real-time queue length and the time required for queue dissipation can be estimated, providing dynamic input for adjusting speed planning under congestion conditions.

[0135] Step nine involves designing a virtual traffic light estimation and mapping method that integrates real-time queue status. This method merges actual traffic lights with queue status to generate virtual traffic light timing sequences, addressing the issue of green wave traffic constraint failure under congestion conditions. The introduced virtual traffic light concept, tailored to traffic flow and queue characteristics, utilizes data obtained from co-simulation, including queue length, vehicle quantity, road parameters, and traffic light timing sequences, to design virtual traffic light timing and location parameters suitable for the current road segment.

[0136] To incorporate the impact of traffic congestion at the intersection into the core green wave traffic framework, the concept and estimation method of "virtual traffic lights" are introduced. The actual physical traffic lights are merged with the dynamic queue information estimated in step eight to generate a new set of traffic light timing and location parameters suitable for this vehicle. Specifically, the start time of the virtual traffic light's green light corresponds to the expected time when the queue is completely cleared (or a slightly earlier safe time), while its start time of the red light may be earlier than the start time of the actual traffic light's red light (to prompt the vehicle to slow down in advance). The virtual traffic light's geographical location is set at the end of the current queue, rather than the actual intersection stop line. Subsequently, this set of virtual traffic light green light time intervals is introduced as a new, more stringent constraint into the MPC controller optimization problem in step seven. This allows the vehicle to smoothly decelerate in advance when it senses congestion ahead, waiting for the queue to clear, thereby reducing the energy-intensive and uncomfortable behavior of sudden braking and restarting.

[0137] This embodiment offers at least the following advantages: Regarding energy consumption optimization, by combining the green wave traffic strategy with Model Predictive Control (MPC), this embodiment achieves global energy-saving optimization of the vehicle's trajectory. The solution reduces the high-energy-consuming operations of traditional following models, such as frequent starts and stops, and rapid acceleration and deceleration, caused by ignoring traffic light status. It guides vehicles through multiple consecutive intersections within the green light window with a smooth speed curve. Comparative experimental data from specific implementations show that, under different traffic flow rates (200–600 vehicles / h), compared to the traditional IDM following model, this model reduces energy consumption by an average of 52%, improving the economy of autonomous vehicles and aligning with the development trend of green transportation. Regarding improved ride comfort, this embodiment improves the smoothness of vehicle operation by optimizing the speed curve and control strategy. The model incorporates jerk constraints during design, controlling the rate of acceleration change within a reasonable range and reducing ride discomfort caused by sudden speed changes. Compared to the IDM model, the average acceleration is reduced by 79%, and the speed and acceleration curves are smoother, improving the comfort of passengers. This enhanced comfort makes the autonomous driving system feel more like natural human driving, helping to increase user acceptance and trust in the technology. Regarding traffic efficiency and capacity improvement, this embodiment improves vehicle throughput at intersections with consecutive traffic lights through green window lane selection and virtual traffic light technology. The green window lane selection strategy plans routes and speeds for vehicles, allowing them to pass through multiple intersections consecutively during green light periods, reducing waiting time; virtual traffic light technology adapts to traffic congestion scenarios and dynamically adjusts traffic constraints.

[0138] This application also provides a vehicle routing device based on traffic signals, which can implement the above method. The device includes: The acquisition module is used to acquire traffic light data at each intersection along the target driving route; The time planning module is used to calculate the time constraint for continuous passage through multiple intersections based on the green light time interval of the traffic signal data, and obtain the time constraint for continuous green wave passage. The speed planning module is used to calculate the speed of each road segment according to the time constraint, and obtain the speed range of each road segment. The optimization module is used to perform secondary planning on the speed range with energy consumption and travel time as optimization objectives to obtain a global speed sequence; The execution module is used to perform multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence, so as to control the vehicle to travel on the target driving path.

[0139] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0140] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A vehicle route planning method based on traffic signals, characterized in that, The method includes the following steps: Obtain traffic light data at each intersection along the target driving route; Based on the green light time interval of the traffic signal data, the time constraint for continuous passage through multiple intersections is calculated to obtain the time constraint for continuous green wave passage. Based on the time constraints, the speed of each road segment is calculated to obtain the speed range of each road segment; With energy consumption and travel time as optimization objectives, a secondary planning is performed on the speed range to obtain a global speed sequence; The vehicle's speed controller is subjected to multi-dimensional constraints and multi-objective optimization based on the global speed sequence in order to control the vehicle to travel on the target driving path.

2. The method according to claim 1, characterized in that, The step of calculating the time constraint for continuous passage through multiple intersections based on the green light time of the traffic signal data to obtain the continuous green wave passage constraint includes: The target travel path is divided according to the intersection, and multiple different road segments are determined; With the goal of ensuring that the time when a vehicle arrives at each intersection falls within the green light time interval, the travel time for each road segment is calculated to obtain the travel time of the vehicle in each road segment. The time constraint includes the travel time.

3. The method according to claim 1, characterized in that, The step of calculating the speed of each road segment based on the time constraint to obtain the speed range of each road segment includes: Obtain the first speed range of the previous road segment; Speed ​​is calculated based on the first speed range, the length of the current road segment, and the time constraint to obtain the speed range of the current road segment.

4. The method according to claim 1, characterized in that, The process of performing secondary planning on the speed range with energy consumption and travel time as optimization objectives yields a global speed sequence, including: The energy consumption under different speed configurations is calculated based on the speed range and the preset vehicle energy consumption model to obtain the total energy consumption of the vehicle. The vehicle's travel time is calculated based on the speed range and the length of each road segment. Based on the total energy consumption and the travel time, a cost function is constructed to obtain the total cost objective function; The global velocity sequence is obtained by minimizing the total cost objective function using a nonlinear constraint optimization method.

5. The method according to claim 4, characterized in that, The vehicle energy consumption model is obtained through the following steps: Based on the vehicle's parameter configuration, a longitudinal dynamics model of the vehicle is obtained by modeling using Newton's second law. The energy consumption model is constructed based on the vehicle's longitudinal dynamics model to obtain the vehicle energy consumption model.

6. The method according to claim 1, characterized in that, The step of performing multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence includes: Obtain real-time vehicle driving parameters; Real-time energy consumption, speed deviation, jerk and real-time position are determined based on the preset vehicle longitudinal dynamics model, the global velocity sequence and the real-time driving parameters, respectively. The speed controller is optimized based on the real-time energy consumption, the speed deviation, the jerk, and the real-time position. The speed controller is subject to multi-dimensional constraints based on actuator constraints, kinematic constraints, and spatiotemporal constraints related to the green light.

7. The method according to claim 1, characterized in that, When a traffic jam is detected ahead of the vehicle, the method further includes: Obtain the length of the congested queue, the traffic density at different positions in the queue, and the traffic flow. The wave velocity of the traffic wave of the queue is obtained by calculating the wave velocity based on the traffic density and the traffic flow. The propagation time of the traffic wave is calculated based on the queue length and the wave speed. The time is modified according to the propagation duration and the actual green light time interval to obtain the virtual time interval of the virtual green light; The speed controller is subject to additional constraints based on the virtual time interval.

8. A vehicle route planning device based on traffic signals, characterized in that, The device includes: The acquisition module is used to acquire traffic light data at each intersection along the target driving route; The time planning module is used to calculate the time constraint for continuous passage through multiple intersections based on the green light time interval of the traffic signal data, and obtain the time constraint for continuous green wave passage. The speed planning module is used to calculate the speed of each road segment according to the time constraint, and obtain the speed range of each road segment. The optimization module is used to perform secondary planning on the speed range with energy consumption and travel time as optimization objectives to obtain a global speed sequence; The execution module is used to perform multi-dimensional constraints and multi-objective optimization on the vehicle's speed controller based on the global speed sequence, so as to control the vehicle to travel on the target driving path.

9. An electronic device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images