An eco-friendly driving method, system, equipment, and medium for a roundabout scenario

Abstract

This invention discloses an eco-driving method, system, device, and medium for a roundabout scenario, comprising: acquiring initial state data of each connected vehicle in the roundabout scenario; constructing an eco-driving model based on a KAN network and setting a composite reward function; the eco-driving model including a control master network and a target network with identical structures; the composite reward function including a safety function, a driving efficiency function, a driving comfort function, a low-speed reward / penalty function, an energy consumption function, and a destination arrival reward / penalty function; training and optimizing the eco-driving model based on the initial state data; inputting the real-time state space into the trained eco-driving model; outputting the optimal safe reference speed; and performing driving control on each connected vehicle based on the optimal safe reference speed. This invention is applicable to intelligent connected environments, focuses on energy conservation and emission reduction, and significantly improves the driving efficiency, safety, driving comfort, and energy-saving effect of connected vehicles passing through roundabouts.

Description

Technical Field

[0001] This invention belongs to the field of traffic driving technology, and in particular relates to an eco-driving method, system, device and medium in a roundabout scenario. Background Technology

[0002] In complex urban traffic environments, traditional intersections are increasingly unable to meet the demands of urban development. While roundabouts offer a viable alternative to alleviate intersection congestion, reducing traffic conflicts and improving efficiency, their common "butterfly effect" limits this advantage: when vehicles inside the roundabout fail to yield to those outside, localized traffic congestion can quickly spread, leading to frequent starts and stops, increased energy consumption, and ultimately, congestion across the entire traffic system. This phenomenon is often difficult to detect and address in a timely manner, thus exacerbating congestion.

[0003] To address roundabout congestion, researchers have proposed centralized traffic management methods. These methods utilize a centralized controller to guide vehicles safely and efficiently through roundabouts in real time. However, with increasing vehicle numbers, centralized management systems are prone to overload due to the increased complexity of information processing, leading to decreased reliability and robustness. In recent years, with the development of connected and automated vehicle (CAV) technology, vehicles can receive surrounding traffic information via V2X (vehicle-to-everything) communication. Researchers have leveraged the V2X communication devices on CAVs to perceive surrounding traffic information and used deep reinforcement learning to train CAVs to control their accelerator, brakes, or steering wheel, effectively solving the problem of safe and efficient roundabout passage for CAVs in multi-vehicle environments. Through deep reinforcement learning training, the system can respond more flexibly to complex traffic environments, improving its reliability and robustness. While these methods have made progress in improving the efficiency, safety, and comfort of vehicles passing through roundabouts, they have not fully considered the needs of eco-driving aspects such as energy conservation and emission reduction. Summary of the Invention

[0004] The purpose of this invention is to provide an eco-driving method, system, device, and medium for roundabout scenarios, in order to solve the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides an eco-driving method in a roundabout scenario, comprising:

[0006] Acquire the initial state data of each connected vehicle in a roundabout scenario. The initial state data includes the longitudinal speed, lateral speed, acceleration, speed difference and distance difference with the vehicles in front and behind each connected vehicle.

[0007] An ecological driving model is constructed based on a KAN network and a composite reward function is set. The ecological driving model includes a control main network and a target network. The control main network and the target network have the same structure, both including an Actor-KAN network and a Critic-KAN network connected in sequence. The composite reward function includes a safety function, a driving efficiency function, a driving comfort function, a low-speed reward and penalty function, an energy consumption function, and a destination arrival reward and penalty function.

[0008] The initial state data is used as the initial state space and input into the eco-driving model for training and optimization to obtain the trained eco-driving model.

[0009] The real-time state space is input into the trained ecological driving model, and the optimal safety reference speed is output. Based on the optimal safety reference speed, driving control is performed on each connected vehicle.

[0010] Optionally, the training and optimization process of the ecological driving model specifically includes:

[0011] Step 1: Input the initial state space into the control main network to obtain the initial action space. Obtain the initial safety reference speed based on the initial action space. Apply the initial safety reference speed to each connected vehicle and obtain the next moment state space of each connected vehicle. Store the initial state space, initial action space, reward function value and next moment state space into the experience pool. Repeat Step 1 until the experience pool reaches the preset saturation state.

[0012] Step 2: When the experience pool reaches the preset saturation state, the Critic-KAN network in the control main network obtains the state space from the experience pool for evaluation and obtains the first evaluation result. Based on the first evaluation result, the parameters of the Actor-KAN network in the control main network are updated.

[0013] Step 3: The target network obtains the next state space at the current moment from the experience pool, and calculates the action space and the second evaluation result at the next moment; the parameters of the eco-driving model are updated based on the first evaluation result and the second evaluation result;

[0014] Repeat steps 1 to 3 for iterative updates until the preset number of iterations is reached, then stop iterating to obtain the trained ecological driving model.

[0015] Optionally, the process of obtaining the initial safety reference speed specifically includes:

[0016] The state data is used as the initial state space and input into the Actor-KAN network in the main control network. The initial action is output, and the initial action is scaled to obtain the initial action space. The initial safety reference velocity is obtained based on the initial action space.

[0017] Optionally, step 2 specifically includes:

[0018] When the experience pool reaches a preset saturation state, the current state space and current action space are input into the Critic-KAN network in the main control network for evaluation to obtain a first evaluation result. Based on the first evaluation result, the parameters of the Actor-KAN network in the main control network are updated to obtain the parameter-updated Actor-KAN network.

[0019] Optionally, step 3 specifically includes:

[0020] The Actor-KAN network in the target network obtains the next state space at the current time step from the experience pool, and calculates the action space and reward function value at the next time step based on the obtained next state space;

[0021] The next state space, the next action space, and the corresponding reward function value are input into the Critic-KAN network in the target network for evaluation, and a second evaluation result is obtained.

[0022] With the goal of minimizing the mean square error between the second evaluation result and the first evaluation result, the Critic-KAN network in the control master network is updated to obtain the updated control master network.

[0023] Based on the soft update strategy, the updated parameters in the main control network are applied to the target network to obtain the updated target network.

[0024] An eco-driving system for roundabout scenarios includes:

[0025] The data acquisition module is used to acquire the initial state data of each connected vehicle in the roundabout scenario. The initial state data includes the longitudinal speed, lateral speed, acceleration, speed difference and distance difference with the vehicles in front and behind each connected vehicle.

[0026] The model training module is used to construct an eco-driving model based on the KAN network and set a composite reward function. The eco-driving model includes a control main network and a target network. The control main network and the target network have the same structure, each including an Actor-KAN network and a Critic-KAN network connected in sequence. The composite reward function includes a safety function, a driving efficiency function, a driving comfort function, a low-speed reward / penalty function, an energy consumption function, and a destination arrival reward / penalty function. The initial state data is used as the initial state space and input into the eco-driving model for training and optimization to obtain the trained eco-driving model.

[0027] The eco-driving control module is used to input the real-time state space into the trained eco-driving model, output the optimal safety reference speed, and perform driving control on each connected vehicle based on the optimal safety reference speed.

[0028] An electronic device includes a memory and a processor, the memory storing a computer program and the processor running the computer program to enable the electronic device to perform the eco-driving method in a roundabout scenario.

[0029] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned eco-driving method in a roundabout scenario.

[0030] The technical effects of this invention are as follows:

[0031] This invention provides an eco-driving method based on AC-KAN network in roundabout scenarios. This method is not only applicable to intelligent connected environments, but also focuses on energy conservation and emission reduction, which significantly improves the driving efficiency, safety, driving comfort, and energy-saving effect of connected vehicles passing through roundabouts. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments 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 based on these drawings without creative effort.

[0033] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0034] Figure 1 This is a general framework diagram in an embodiment of the present invention;

[0035] Figure 2This is a diagram of the AC-KAN network structure in an embodiment of the present invention;

[0036] Figure 3 This is a schematic diagram of the key internal structure of the KAN network in an embodiment of the present invention. Detailed Implementation

[0037] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as a limitation of the present invention, but rather as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0038] It should be understood that the terminology used in this invention is merely for describing particular embodiments and is not intended to limit the invention. Furthermore, with respect to numerical ranges in this invention, it should be understood that each intermediate value between the upper and lower limits of the range is also specifically disclosed. Every smaller range between any stated value or intermediate value within a stated range, and any other stated value or intermediate value within said range, is also included in this invention. The upper and lower limits of these smaller ranges may be independently included or excluded from the range.

[0039] Unless otherwise stated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. While only preferred methods have been described herein, any methods similar or equivalent to those described herein may be used in the implementation or testing of this invention. All references to this specification are incorporated by way of citation to disclose and describe the methods associated with those references. In the event of any conflict with any incorporated reference, the content of this specification shall prevail.

[0040] Various modifications and variations can be made to the specific embodiments described in this specification without departing from the scope or spirit of the invention, as will be apparent to those skilled in the art. Other embodiments derived from this specification will also be obvious to those skilled in the art. This application specification and embodiments are merely exemplary.

[0041] The terms “include,” “including,” “have,” “contain,” etc., used in this article are all open-ended terms, meaning that they include but are not limited to.

[0042] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0043] Example 1

[0044] like Figure 1 - Figure 3As shown, this embodiment provides an eco-driving method in a roundabout scenario, including: acquiring initial state data of each connected vehicle in the roundabout scenario, the initial state data including the longitudinal speed, lateral speed, acceleration, speed difference and distance difference with the vehicles in front and behind each connected vehicle; constructing an eco-driving model based on a KAN network and setting a composite reward function, the eco-driving model including a control master network and a target network, the control master network and the target network having the same structure, both including an Actor-KAN network and a Critic-KAN network connected in sequence; the composite reward function including a safety function, a driving efficiency function, a driving comfort function, a low-speed reward / penalty function, an energy consumption function, and a destination arrival reward / penalty function; inputting the initial state data as the initial state space into the eco-driving model for training and optimization, obtaining a trained eco-driving model; inputting the real-time state space into the trained eco-driving model, outputting the optimal safety reference speed, and performing driving control on each connected vehicle based on the optimal safety reference speed.

[0045] This embodiment provides an eco-driving method based on AC-KAN network in a roundabout scenario. This method can not only improve vehicle driving efficiency, safety and driving comfort, but also effectively reduce energy consumption.

[0046] This embodiment mainly includes the following steps:

[0047] Step 1: Design the state space. In the roundabout scenario, the longitudinal speed, lateral speed, acceleration of the connected vehicle, as well as the speed difference and distance with the vehicles in front and behind are set as the state space.

[0048] Step 2: Design an eco-driving model. Design an eco-driving model based on the AC-KAN network, and use the state space from Step 1 as the input to the model.

[0049] Step 3: Design a composite reward function. Develop a reward function for the eco-driving model from Step 2. This function includes a safety function, a driving efficiency function, a driving comfort function, a low-speed reward / penalty function, an energy consumption function, and a destination arrival reward / penalty function.

[0050] Step 4: Design safety actions. The acceleration of the connected vehicle is set as the action space and used as the output of the eco-driving model in Step 2. The actions are scaled in conjunction with safe speed and anti-idle strategies to obtain a safe reference speed, thus achieving safe control of the connected vehicle.

[0051] First, in the intelligent connected roundabout scenario, this embodiment utilizes V2X communication technology to acquire traffic information of connected vehicles and parse the state space. Before the experience pool is saturated, the eco-driving model based on the AC-KAN network is in the experience accumulation period. It only uses the initial network to calculate the input state space to obtain the action output. Combined with safe speed and anti-idle strategy, the action is scaled to obtain a safe reference speed, realizing intelligent control of the connected vehicle. At the same time, the state space, action space, reward function, next state space, round end flag, and other information parsed from the traffic information are stored in the experience pool to accumulate experience. When the experience pool is saturated, the eco-driving model based on the AC-KAN network enters the sampling learning period. Through deep calculation of input features such as state space, action space, reward function, and next state space, the network is continuously updated to generate the action space with the highest evaluation value. Combined with safe speed and anti-idle strategy, the optimal safe reference speed of the connected vehicle in the roundabout scenario is obtained. The above process is repeated to finally obtain the optimal eco-driving model.

[0052] The specific implementation process of this embodiment includes:

[0053] Step 1: Design the state space. In the roundabout scenario, the longitudinal velocity, lateral velocity, acceleration of the connected vehicle, and the speed difference and distance with the vehicles in front and behind are defined as the state space. This state space is represented as follows:

[0054]

[0055] in, For the longitudinal speed of the connected vehicle, For the lateral speed of connected vehicles, Accelerating connected vehicles The speed difference between the connected vehicle and the vehicle in front. The speed difference between the connected vehicle and the vehicle behind it. This refers to the distance difference between the connected vehicle and the vehicle in front. Let n represent the distance difference between the connected vehicle and the vehicle behind it, where n represents the number of model iterations and i represents the current time step. This information can be obtained through V2X technology and GPS (Global Positioning System), ensuring the real-time nature of the data.

[0056] Step 2: Design an eco-driving model. Design an eco-driving model based on an AC-KAN network, using the state space from Step 1 as input. (See below.) Figure 2 The diagram shown is a schematic of the AC-KAN network structure in this embodiment. The specific steps are as follows:

[0057] (1) Experience accumulation period. The Actor-KAN network obtains the state space of the connected vehicle in the current time step 1 of the roundabout scenario. The original action values ​​are then calculated using a KAN network. To ensure the action values ​​are within a suitable range, they are scaled to obtain the action space in step 4. A safe reference speed is obtained by combining safe speed and anti-idle speed strategies, and then applied to connected vehicles in a roundabout scenario. At this point, the experience pool... Store the state space obtained from the island scene. Action space reward function value r i n Next state space And the iteration count flag "done". Before the experience pool is saturated, make the current state space equal to the next state space, i.e. To achieve a cyclical process of accumulating experience.

[0058] in, Indicates the model performs actions Then, the state space obtained again from the roundabout scene is denoted as... The "done" flag indicates the number of model iterations. When "done = True", the next iteration begins; when "done = Flash", experience continues to accumulate.

[0059] (2) Sampling learning period. When the experience pool is saturated, the Critic-KAN network obtains the state space of the current time step from the experience pool. Obtaining the action space from the Actor-KAN network Calculate the state-action pair The goal of the Critic-KAN network is to evaluate the value in the state space. Lower execution action space The value of the Critic-KAN network. Based on the value feedback information of the Critic-KAN network, the Actor-KAN network uses L... actor The function updates its own parameters in order to anticipate the action space generated in the next time step. The value is greater.

[0060] in, The value is the Critic-KAN network's estimate of the value of the current state and action, reflecting the expected cumulative reward that can be obtained by performing the action under the current policy. The formula for calculating the value is as follows: A schematic formula for the output of the Critic-KAN network:

[0061]

[0062] Among them, L actor The formula is as follows, where N is the training batch size, and its purpose is to maximize... value:

[0063]

[0064] (3) Target network learning period. When the experience pool is saturated, the target Actor-KAN network obtains the next state space for the current time step from the experience pool. The scaled target action space is obtained by calculating using a KAN network. The goal is for the Critic-KAN network to obtain the next state space for the current time step from the experience pool. and reward function value Obtain the target action space from the target Actor-KAN network. Calculate the goal of this state-action-reward pair value.

[0065] Among them, the target The value is calculated based on the actual reward received and the target network's estimate. It represents a more accurate expected reward and is used to guide the Critic-KAN network's learning, making its estimate closer to the true value. Target The formula for calculating the value is as follows: target A schematic formula for the output of the target Critic-KAN network:

[0066]

[0067] (4) Network update period. The ecological driving model uses L... critic Function minimization Values ​​and Targets The mean squared error of the values ​​is used to update the parameters of the Critic-KAN network, thereby improving its accuracy in evaluating the quality of the action space. A soft update strategy is used for both the target Actor-KAN and Critic-KAN networks, gradually updating their parameters to the target network with small step sizes τ to ensure training stability. Repeating these steps yields the optimal eco-driving model.

[0068] Among them, L critic The formula is as follows, where N is the training batch size, y is the intermediate variable, and γ is the discount factor, typically taken as 0.99:

[0069]

[0070]

[0071] The formula for soft parameter update is as follows: θ target Let θ be the target network parameter.current Given the current network parameters, step size τ = 0.005:

[0072] θ target ←τθ current +(1-τ)θ target (7)

[0073] The internal key structure of the KAN network, which is included in all the above networks, is illustrated below. Figure 3 As shown. The specific implementation steps of the environment-adaptive KAN network construction method in this embodiment are as follows:

[0074] (1) Design adaptive nodes: Obtain the state space from step 1 before constructing the KAN network. and the motion space of step 4 The dimension. In the Actor-KAN network, only the state space is used. The dimension of the input layer is used as the number of nodes, and the state space is used as the number of nodes. As input values, the state space is used in the Critic-KAN network. Add movement space The dimensions and the number of nodes used as input layers, state space and action space All are used as input values. The number of nodes in the intermediate layers is the number of nodes in the input layers plus 1. A 4-layer structure is used for the intermediate layers, which is more consistent with the construction of multi-layer KAN networks. The number of nodes in the output layer is determined by the action space. The number of dimensions.

[0075] (2) Defining spline basis functions: The KAN network uses B-spline basis functions as spline basis functions. Some of the B-spline basis functions are as follows: Figure 3 As shown, the function is composed of a set of piecewise polynomials and has good local control and continuity. Due to the high complexity of the state space in step 1 of this embodiment, a third-order B-spline basis function is used, and its formula is as follows:

[0076]

[0077] Where j is the number of B-spline basis functions, k is the order of the B-spline basis functions, x is the function input value, t is the node of the B-spline function, and B... j,k (x) represents the j-th B-spline basis function.

[0078] (3) KAN network learning and updating: In the KAN network, the output y of each layer KAN It is calculated through a nonlinear combination of the input value x, and each layer f in the network is... m The transformation is performed using a spline basis function, and the formula is as follows:

[0079]

[0080] Where m represents the number of layers in the KAN, and n KAN The dimension representing the weight, ω j This represents the weights of the spline basis functions.

[0081] In a KAN network, the loss function L is defined. KAN It depends on the output of the KAN network and the mean square error of the true values ​​in the experience pool, and its formula is as follows:

[0082]

[0083] Where, N KAN This indicates the batch size, and h represents the sample index value. The final output value of equation (9) is... The actual values ​​in the experience pool, here representing the action space.

[0084] The gradient descent method is used to update the weights ω in the KAN network. j The nodes t of the spline basis functions j The formula is as follows:

[0085]

[0086]

[0087] Where ρ is the network learning rate, typically 0.001.

[0088] When the Actor-KAN and Critic-KAN networks, as well as their corresponding target networks, are updated, the parameters in the KAN network are simultaneously trained and updated, ensuring that the network learning process is mutually reinforcing. Each training session typically includes 30 update steps, which can be adjusted according to the complexity of the task. During training, the KAN network optimizes its parameters by introducing regularization and sparsity techniques. Regularization helps constrain the size of the network weights, thereby preventing overfitting, while sparsity makes the network more concise and efficient by reducing unnecessary connections or nodes. This optimization method not only improves the network's generalization ability but also enhances the computational efficiency when deploying the eco-driving model.

[0089] Step 3: Design a composite reward function: Develop a reward function for the eco-driving model in Step 2. The formula for this function is as follows:

[0090]

[0091] in, For security functions, For driving efficiency function, For driving comfort function, Let be the energy consumption function. This is a low-speed reward / penalty function. The reward / penalty function for reaching the destination.

[0092] The specific descriptions of each function are as follows:

[0093] (1) Safety Function: In this embodiment, the safety of the current action is represented by the sum of the lane departure of the connected vehicle and the TTC (Time to Collision). The formula for the safety function is as follows:

[0094]

[0095]

[0096]

[0097] Where, α safe α is a scaling factor, calculated and manually adjusted. safe =10, For lane departure reward / penalty functions, This is the TTC reward / penalty function. L is the lateral distance of the vehicle relative to the center of the lane. width This refers to the lane width. and Let D be the distance and speed difference between the connected vehicle and the vehicle in front in the state space of step 1, and let D be the distance threshold between the connected vehicle and the vehicles in front and behind. Since the diameter of a roundabout generally does not exceed 100m, setting D=50 is more appropriate here. When the distance to the vehicle in front exceeds 50m, Approaching 0, therefore for No effect. Conversely, when the distance to the preceding vehicle is less than or equal to 50m, the result is calculated using equation (14).

[0098] (2) Driving efficiency function: This embodiment encourages connected vehicles to drive as efficiently as possible within the speed limit based on their speed. The formula for the driving efficiency function is as follows:

[0099]

[0100] Where, α efficient α is a scaling factor, calculated and manually adjusted. efficient =40, v limit This refers to the road speed limit. Let be the longitudinal velocity of the connected vehicle in the state space of step 1.

[0101] (3) Driving comfort function: In this embodiment, the degree of bumpiness caused to passengers by the vehicle's acceleration change is represented by the square of the change in acceleration. The formula for the driving comfort function is as follows:

[0102]

[0103] Where, α comfort α is a scaling factor, calculated and manually adjusted. comfort =1, The acceleration in the state space of step 1, Let t(i) represent the acceleration at the previous time step in the state space, and let t(i) represent the size of the time step.

[0104] (4) Energy Consumption Function: Since the vehicle used in this embodiment is a new energy electric vehicle with energy recovery function, this embodiment calculates the reward value based on the power consumption and travel distance of the connected vehicle, using the quotient of the power consumed per unit time step and the travel distance. If the vehicle's travel distance within this time step is less than 0.01m, i.e., the vehicle is moving slowly, a fixed negative reward value of 50 is set, which can effectively solve the problem of vehicle idling. The formula for the energy consumption function is as follows:

[0105]

[0106] Where, α energy α is a coefficient of variation, set through calculation and manual adjustment. energy =15, This represents the amount of electricity consumed within the current time step. This represents the distance traveled within the current time step.

[0107] (5) Low-speed reward / penalty function: In this embodiment, the penalty is applied based on the speed of the connected vehicle. When When the vehicle is clearly traveling at a low speed, the penalty increases with the speed. The formula for the low-speed reward / penalty function is as follows:

[0108]

[0109] Where, α idle α is a coefficient of variation, set through calculation and manual adjustment. idle =120, Let be the longitudinal velocity of the connected vehicle in the state space of step 1.

[0110] (6) Destination Reward / Penalty Function: In this embodiment, the driving route of the connected vehicle is first obtained, and then the last lane in the route, i.e., the target lane, is locked. When the distance traveled by the connected vehicle in the target lane is greater than or equal to the length of the target lane, the connected vehicle is considered to have reached the destination, and at this time, arrive_flash = True is set; otherwise, arrive_flash = False. After the connected vehicle reaches the destination, a fixed positive reward value is given, set to 50; otherwise, it is 0. The formula for the destination reward / penalty function is as follows:

[0111]

[0112] Here, arrival_flash represents the marker value indicating that the connected vehicle has arrived at its destination.

[0113] Step 4: Designing Safety Actions: In this embodiment, the acceleration of the connected vehicle is set as the action space of the ecological driving model in Step 2. Considering safety and the physical meaning of speed, the acceleration output by the algorithm cannot be directly applied to the vehicle's motion in some situations and needs to be constrained. Due to the limitations on the output of the ecological driving model... Therefore, the action values ​​output by the eco-driving model need to be scaled first, as shown below:

[0114]

[0115] The scaled action space is defined as a one-dimensional vector:

[0116]

[0117] The formula for safe action after constraints is as follows:

[0118]

[0119] in, a is the original output value of the eco-driving model. min a max These are the vehicle's maximum deceleration and maximum acceleration, respectively. limit For road speed limits, v carfollow The speed is calculated using the default car-following model (Krauss model). The reference speed value for the intelligent driving model when finally applied to the intelligent connected vehicle scenario is the acceleration multiplied by the unit time, i.e., t(i). and The connected vehicle is identified as being in an idling state, and the default car-following speed v is adopted. carfollow In other cases, a safe speed limit is applied. Repeating the above steps will train the optimal intelligent driving model.

[0120] This embodiment provides an eco-driving method based on AC-KAN network in a roundabout scenario. This method is not only applicable to intelligent connected environments, but also focuses on energy conservation and emission reduction, which significantly improves the driving efficiency, safety, driving comfort, and energy-saving effect of connected vehicles passing through roundabouts.

[0121] An eco-driving system for roundabout scenarios includes:

[0122] The data acquisition module is used to acquire the initial state data of each connected vehicle in the roundabout scenario. The initial state data includes the longitudinal speed, lateral speed, acceleration, speed difference and distance difference with the vehicles in front and behind each connected vehicle.

[0123] The model training module is used to construct an eco-driving model based on the KAN network and set a composite reward function. The eco-driving model includes a control main network and a target network. The control main network and the target network have the same structure, each including an Actor-KAN network and a Critic-KAN network connected in sequence. The composite reward function includes a safety function, a driving efficiency function, a driving comfort function, a low-speed reward / penalty function, an energy consumption function, and a destination arrival reward / penalty function. The initial state data is used as the initial state space and input into the eco-driving model for training and optimization to obtain the trained eco-driving model.

[0124] The eco-driving control module is used to input the real-time state space into the trained eco-driving model, output the optimal safety reference speed, and perform driving control on each connected vehicle based on the optimal safety reference speed.

[0125] An electronic device includes a memory and a processor, the memory storing a computer program and the processor running the computer program to enable the electronic device to perform the eco-driving method in a roundabout scenario.

[0126] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned eco-driving method in a roundabout scenario.

[0127] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An ecological driving method in a roundabout scene, characterized in that, The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene.

2. The method according to claim 1, wherein, The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene.

3. The method according to claim 1, wherein, The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene.

4. The method according to claim 1, wherein, The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The application relates to an ecological driving method and device for a ring island scene. The Actor-KAN network in the target network obtains a next state space at a current time in an experience pool, and calculates an action space at a next time and a reward function value based on the obtained next state space; The next state space, the action space at the next time and the corresponding reward function value are input into a Critic-KAN network in the target network for evaluation to obtain a second evaluation result; The Critic-KAN network in the control main network is updated to obtain an updated control main network, with the objective of minimizing the mean square error between the second evaluation result and the first evaluation result; The parameters in the updated control main network are updated into the target network based on a soft update strategy to obtain an updated target network.

5. An eco-driving system in a roundabout scenario, applying the method according to any one of claims 1-4, characterized in that, Comprise: A data acquisition module is configured to acquire initial state data of each connected vehicle in a roundabout scene, wherein the initial state data includes the longitudinal speed, the lateral speed, the acceleration, the speed difference and the distance difference of each connected vehicle relative to the front and rear vehicles; A model training module is configured to construct an ecological driving model according to a KAN network and set a composite reward function, wherein the ecological driving model comprises a control main network and a target network, the control main network and the target network have the same structure and each comprises an Actor-KAN network and a Critic-KAN network connected in sequence; the composite reward function comprises a safety function, a driving efficiency function, a driving comfort function, a low-speed reward and punishment function, an energy consumption function and a destination reward and punishment function; the initial state data is input into the ecological driving model as an initial state space for training and optimization to obtain a trained ecological driving model; An ecological driving control module is configured to input a real-time state space into the trained ecological driving model to output an optimal safe reference speed, and perform driving control on each connected vehicle based on the optimal safe reference speed.

6. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory is configured to store a computer program, and the processor is configured to run the computer program to enable the electronic device to perform the ecological driving method in a roundabout scene according to any one of claims 1-4.

7. A computer readable storage medium characterized by The computer program is stored in the memory and is executed by the processor to implement the ecological driving method in a roundabout scene according to any one of claims 1-4.

Images