Intelligent networked electric vehicle energy-saving path and economic vehicle speed collaborative optimization method

Abstract

The invention relates to the technical field of intelligent networked electric vehicles, in particular to an intelligent networked electric vehicle energy-saving path and economic speed collaborative optimization method, which comprises the following steps: 1) acquiring real road information of a target area based on SUMO; 2) constructing a macroscopic passing vehicle speed prediction model based on a Transform neural network; 3) predicting the passing speed of the target vehicle in each road section of the target area; 4) designing an energy consumption prediction method of the intelligent networked electric vehicle in combination with a vehicle longitudinal dynamics theory and a vehicle speed prediction result; 5) designing an efficient layered energy-saving path and economic vehicle speed collaborative optimization architecture; 6) solving an optimal energy-saving path R * by adopting an improved genetic algorithm; and 7) calculating the economic vehicle speed when the target vehicle runs on the path R * in combination with the KKT condition and a gradient descent algorithm. According to the invention, the prediction precision of the passing speed of the target vehicle in each road section can be obviously improved; the precision and the reliability of a solving result are improved; and collaborative optimization of the energy-saving path and the economic speed of the intelligent networked electric vehicle is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent connected electric vehicles, and in particular to a method for collaboratively optimizing energy-saving paths and economical vehicle speeds of intelligent connected electric vehicles. Background Art

[0002] With the growing demands of modern transportation and the advancement of intelligent technologies, intelligent connected vehicles (ICVs) have become a key research focus in the global automotive industry. Combining advanced sensor technology, communication technology, and autonomous driving algorithms, ICVs enable information exchange between vehicles (V2V), vehicles and infrastructure (V2I), vehicles and pedestrians (V2P), and vehicles and networks (V2N). This highly interconnected and autonomous nature offers new solutions for improving traffic safety, alleviating congestion, reducing energy consumption, and minimizing environmental pollution.

[0003] While intelligent connected vehicles (ICVs) demonstrate significant potential for improving traffic efficiency and safety, further reducing vehicle energy consumption and extending driving range remain key challenges. To address this challenge, energy-efficient path planning, leveraging the high-precision sensing and communication capabilities of ICVs, has become a research hotspot. By integrating real-time traffic data, road condition information, and vehicle dynamics models, it is possible to determine the optimal energy-efficient path for ICVs.

[0004] Current research on vehicle energy-saving path planning has achieved certain results, but its effectiveness in practical applications is still constrained by some limitations. First, most existing research on vehicle energy-saving path planning methods ignores the speed fluctuations caused by vehicle switching and changes in traffic light status, resulting in nonlinear accumulation of energy consumption prediction errors along the driving path, making it difficult for path planning methods to obtain a global optimal solution. In addition, most existing research on vehicle energy-saving path planning methods assumes that the vehicle travels at the average speed of each section on the planned path. However, within the allowable range of total travel time, the energy economy and traffic efficiency of the vehicle can be comprehensively improved through the coordinated optimization of energy-saving paths and economical speeds. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention proposes a method for collaboratively optimizing the energy-saving path and economical vehicle speed of an intelligent connected electric vehicle.

[0006] The technical problem to be solved by the present invention is achieved by adopting the following technical solutions:

[0007] A method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle comprises the following steps:

[0008] 1) Collect real road information of the target area based on SUMO;

[0009] 2) Building a macroscopic traffic speed prediction model based on Transformer neural network;

[0010] 3) Combine road parameters with historical traffic data to train a macro-vehicle speed prediction model, enabling it to predict the speed of target vehicles traveling on various road sections in the target area;

[0011] 4) Design an energy consumption prediction method for intelligent connected electric vehicles by combining vehicle longitudinal dynamics theory with vehicle speed prediction results;

[0012] 5) Design an efficient hierarchical energy-saving path and economical speed collaborative optimization architecture;

[0013] 6) Using the collaborative optimization framework designed in step 5), based on the collected real road information, combined with the traffic speed prediction model and the intelligent connected electric vehicle energy consumption prediction method, an improved genetic algorithm is used to solve the optimal energy-saving path R*;

[0014] 7) Using the collaborative optimization framework designed in step 5), based on the planned optimal energy-saving path R*, the KKT condition and the gradient descent algorithm are combined to calculate the economic speed of the target vehicle when traveling on path R*.

[0015] As a further improvement of the present invention, the specific process of step 2) is as follows:

[0016] 2-1) Designing a Transformer-based encoder-decoder architecture;

[0017] 2-2) Configure the multi-head self-attention mechanism;

[0018] 2-3) Use mean squared error as the main loss function to calculate the mean squared difference between the predicted speed and the true value. 2-4) Add L2 regularization to prevent overfitting.

[0019] 2-5) Select the Adam optimizer, set the initial learning rate, and configure the cosine annealing learning rate scheduling strategy.

[0020] As a further improvement of the present invention, the specific process of step 4) is as follows:

[0021] 4-1) Based on the theory of vehicle longitudinal dynamics, derive the energy consumption calculation formula for intelligent connected electric vehicles;

[0022] 4-2) Combine the energy consumption calculation formula of the intelligent connected electric vehicle in step 4-1) with the vehicle speed prediction result in step 3) to predict the electric energy consumed by the vehicle when passing.

[0023] As a further improvement of the present invention, the energy consumption calculation formula of the intelligent connected electric vehicle in step 4-1) is derived as follows:

[0024] The basic equation for vehicle longitudinal dynamics is:

[0025] ma(t)=F t -(F f +F w +F i ) (1)

[0026] Where m is the total mass of the vehicle (kg); a(t) is the instantaneous acceleration of the vehicle (m / s 2 );F t is the driving force of the vehicle; F f 、F w 、F i These are the rolling resistance, air resistance, and slope resistance that a vehicle overcomes when driving on the road; the specific formulas for each force are as follows:

[0027] F f =fmgcosθ (2)

[0028] F w =0.5C d Aρv(t) 2 (3)

[0029] F i =mgsinθ (4)

[0030] Where g is the gravitational acceleration constant (9.81m / s 2 ); θ is the road inclination angle (°); f is the rolling resistance coefficient; C d is the air resistance coefficient; A is the frontal area of ​​the vehicle (m 2 ); ρ is the air density, generally taken as 1.2258 kg / m 3 ; v(t) is the instantaneous speed of the vehicle;

[0031] To calculate the instantaneous mechanical power P required to drive the wheels m (t)(W), multiplying Equation (1) by the instantaneous speed of the vehicle v(t) yields:

[0032] P m (t)=ma(t)v(t)+fmgv(t)cosθ+0.5C d Aρv(t) 3 +mgv(t)sinθ (5);

[0033] In order to estimate the total mechanical energy e required by the vehicle's power system when it accelerates, decelerates or travels at a constant speed on a certain road m (W·h), integrate Equation (5) over time; assuming the vehicle accelerates or decelerates at a constant acceleration a, the following formula can be obtained:

[0034] v(t)=at+v o (6)

[0035]

[0036] Among them, v o is the initial velocity of the vehicle (m / s), v f is the final speed of the vehicle (m / s); d, t f The vehicle has a speed v i Accelerate or decelerate to speed v f Distance traveled (m) and time (s);

[0037] In order to calculate the total electrical energy e (W h) provided by the vehicle battery pack, the efficiency of the power system must also be considered. The efficiency of the vehicle power system is represented by η, and the following formula is obtained:

[0038]

[0039] In summary, the total electric energy e consumed by the vehicle traveling from node p to node q is pq (W·h) is:

[0040]

[0041] in, The electric energy consumed by the vehicle when accelerating, running at a constant speed, and decelerating respectively; The distances traveled by the vehicle during acceleration, constant speed, and deceleration (m) respectively; θ pq is the inclination angle of section pq; R = 0.5C d Aρ;v pq is the speed of vehicles traveling on road section pq (m / s).

[0042] As a further improvement of the present invention, in step 4-2), a macro traffic speed prediction model is used to obtain the predicted value of the vehicle's traffic speed on each road section in the target area. At the same time, the vehicle parameters and the collected road data are combined and input into the energy consumption calculation formula of the intelligent connected electric vehicle to predict the electric energy consumed by the vehicle when passing.

[0043] As a further improvement of the present invention, the efficient hierarchical energy-saving path and economical vehicle speed collaborative optimization architecture in step 5) includes an energy-saving path planning decision layer and an economical vehicle speed optimization execution layer.

[0044] As a further improvement of the present invention, in step 6), the optimal energy-saving path R* is solved by the energy-saving path planning decision layer. The specific process is as follows:

[0045] 6-1) Use Dijkstra's algorithm to calculate the initial solution:

[0046] Based on the spatial distance of each road section, the Dijkstra algorithm is used to solve the path R0 with the minimum energy consumption for the target vehicle from the starting point to the end point, combining the vehicle speed prediction model with the energy consumption prediction method of intelligent connected electric vehicles under ideal conditions ignoring the dynamic change characteristics of vehicle speed.

[0047] 6-2) Initialize the population:

[0048] Taking R0 as the initial path, four new paths are generated through mutation operation, thus forming the first generation population, which contains five individual paths in total. Each individual path is assigned four attributes, namely travel distance, speed profile, travel time and total energy consumption.

[0049] 6-3) Fitness evaluation:

[0050] The total energy consumption of the trip is used as the basis for evaluating the fitness of the individual. The fitness function is defined as the inverse of the total energy consumption of the trip, ensuring that individuals with lower energy consumption have higher fitness values.

[0051] 6-4) Make a selection:

[0052] From the five individuals of the first generation, the individual with the highest fitness is selected according to the fitness value as the basis for the subsequent crossover operation;

[0053] 6-5) Perform crossover operation:

[0054] Taking the aforementioned high-fitness individuals as the main body, cross operations are performed with other individuals to generate the first individual of the second generation;

[0055] 6-6) Perform mutation operation:

[0056] The first individual of the second generation is used as the initial path, and four more paths are generated through mutation operations to complete the construction of the second generation population.

[0057] 6-7) Repeat the above steps to generate a new generation of population each time until the 20th generation;

[0058] 6-8) In the 20th generation, the individual with the highest fitness is selected as the final solution, and its corresponding path is the optimal energy-saving path R* for the target vehicle to travel from the starting point to the end point.

[0059] As a further improvement of the present invention, the specific calculation method of the travel distance, speed profile, travel time and total travel energy consumption in step 6-2) is as follows:

[0060] ① Travel distance

[0061] The trip distance is directly calculated from the road condition data collected above;

[0062] ② Velocity profile

[0063] The target vehicle's initial speed is set, and it is assumed that upon entering each road segment, the vehicle accelerates or decelerates at a constant rate to the average speed of that section, and then continues to travel at that speed until it enters the next road segment or stops. Based on this assumption, once the path is determined and combined with the collected traffic light position and phase information, the speed profile of the target vehicle along that path can be drawn.

[0064] ③ Travel time and total energy consumption

[0065] Based on the designed energy consumption prediction method, the travel time and total energy consumption of the target vehicle when traveling along the path are calculated according to the travel distance and speed profile of the above path.

[0066] As a further improvement of the present invention, in step 7), the economic speed optimization execution layer is used to calculate the economic speed of the vehicle when traveling on the path R*. The specific process is as follows:

[0067] 7-1) Set constraints:

[0068] Assume that the speed range of each road section is [v imin ,v imax ],v imin =max[0.7v i ,v ilow ],v imax =min[1.3v i ,v iup ], where v i is the speed of the target vehicle on each road section, v ilow , v iup are the minimum speed limit and the maximum speed limit of each road section respectively; the total travel time limit of the target vehicle from the starting point to the end point is T lim ;

[0069] 7-2) Ignore the road slope of each section and assume that the target vehicle is moving at v i (i=1,2,…n) driving at a constant speed, based on the above-mentioned intelligent connected electric vehicle energy consumption prediction method and the set constraints, the following objective function is constructed:

[0070]

[0071] Where n represents the number of road segments on the path R*; d i represents the distance of road segment i;

[0072] 7-3) Combine the KKT condition and the gradient descent algorithm to solve the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint. * ;

[0073]

[0074]

[0075] HΔ=-J (21)

[0076] In the above, according to the KKT condition, the Lagrangian function shown in formula (17) is constructed to find the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint. * , equation (18) should be solved. Since equation (18) is a nonlinear equation group, the Newton iteration method is used to continuously iteratively update the velocity vector v so that it approaches the optimal solution. After constructing the first-order partial derivative matrix (Jacobian matrix) J and the second-order partial derivative matrix (Hessian matrix) H of the Lagrangian function, the iteration step matrix Δ of the Newton iteration method is solved by equation (21). The step matrix Δ is then used to iteratively update the velocity vector v so that it continuously approaches the optimal solution. When the step matrix Δ is less than the set critical value, the iteration terminates. At this time, the velocity vector solution v * That is, it is the optimal economic speed vector that minimizes the objective function while satisfying the travel time constraint.

[0077] The beneficial effects of the present invention are:

[0078] Compared to traditional neural networks like CNN / RNN, Transformer neural networks can effectively capture long-range temporal correlations through their multi-head self-attention mechanism. Therefore, building a speed prediction model based on Transformer neural networks can significantly improve the accuracy of predicting the target vehicle's speed on various road sections.

[0079] The proposed improved genetic algorithm can fully consider the dynamic characteristics of the target vehicle's speed when solving the vehicle energy-saving path planning problem. At the same time, it uses the result of the Dijkstra algorithm under ideal conditions as a high-quality initial solution. Compared with other traditional solution methods, it significantly improves the accuracy and reliability of the solution.

[0080] A hierarchical architecture for collaborative optimization of energy-saving paths and economic speed was designed. The economic speed optimization execution layer, based on the optimal energy-saving path R* output by the energy-saving path planning decision layer, combines the KKT condition with a gradient descent algorithm to calculate the economic speed of the target vehicle along path R*, thereby achieving collaborative optimization of the energy-saving path and economic speed of intelligent connected electric vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0082] Figure 1 This is a schematic diagram of the main structure of the present invention;

[0083] Figure 2 Import map for a certain area;

[0084] Figure 3 Schematic diagram of the optimal energy-saving path R* to be solved;

[0085] Figure 4 is the speed profile of the target vehicle along the optimal energy-saving path R*. DETAILED DESCRIPTION

[0086] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below with reference to the accompanying drawings and embodiments.

[0087] like Figure 1 As shown, a method for collaboratively optimizing energy-saving paths and economical vehicle speeds for intelligent connected electric vehicles includes the following steps:

[0088] 1) Collect real road information of the target area based on SUMO.

[0089] Specifically, in this embodiment, the map data and road network information of the target area are imported into SUMO using OpenStreetMap, and information such as the location of each node, the location of traffic lights and their phases are collected. Figure 2 As shown, this is the imported map of a certain area.

[0090] 2) Construct a macro traffic speed prediction model based on Transformer neural network.

[0091] The specific process is as follows:

[0092] 2-1) Design a Transformer-based encoder-decoder architecture.

[0093] 2-2) Configure the multi-head self-attention mechanism.

[0094] 2-3) The mean square error (MSE) is used as the main loss function to calculate the mean square difference between the predicted vehicle speed and the true value.

[0095] 2-4) Add L2 regularization term to prevent overfitting.

[0096] 2-5) Select the Adam optimizer, set the initial learning rate, and configure the cosine annealing learning rate scheduling strategy.

[0097] 3) Combine historical traffic data with time series characteristics to train a macro-vehicle speed prediction model so that it can predict the speed of target vehicles traveling on various sections of the target area.

[0098] Specifically, in this embodiment, the historical traffic data with time series characteristics includes the current cycle (such as weekdays / holidays), time period (such as morning and evening peak hours) and seasonal patterns.

[0099] 4) Design an energy consumption prediction method for intelligent connected electric vehicles by combining vehicle longitudinal dynamics theory and vehicle speed prediction results.

[0100] The specific process is as follows:

[0101] 4-1) Based on the theory of vehicle longitudinal dynamics, derive the energy consumption calculation formula for intelligent connected electric vehicles.

[0102] The basic equation for vehicle longitudinal dynamics is:

[0103] ma(t)=F t -(F f +F w +F i ) (1)

[0104] Where m is the total mass of the vehicle (kg); a(t) is the instantaneous acceleration of the vehicle (m / s 2 );F t is the driving force of the vehicle; F f 、F w 、F i They are respectively the rolling resistance, air resistance and slope resistance that a vehicle overcomes when driving on the road. The specific formulas for each force are as follows:

[0105] F f =fmgcosθ (2)

[0106] F w =0.5C d Aρv(t) 2 (3)

[0107] F i =mgsinθ (4)

[0108] Where g is the gravitational acceleration constant (9.81m / s 2 ); θ is the road inclination angle (°); f is the rolling resistance coefficient; C d is the air resistance coefficient; A is the frontal area of ​​the vehicle (m 2 ); ρ is the air density, generally taken as 1.2258 kg / m 3 ; v(t) is the instantaneous speed of the vehicle.

[0109] To calculate the instantaneous mechanical power P required to drive the wheels m (t)(W), multiplying Equation (1) by the instantaneous speed of the vehicle v(t) yields:

[0110] P m (t)=ma(t)v(t)+fmgv(t)cosθ+0.5C d Aρv(t) 3 +mgv(t)sinθ (5).

[0111] In order to estimate the total mechanical energy e required by the vehicle's power system when it accelerates, decelerates or travels at a constant speed on a certain road m (W·h), time-integrate Equation (5). Assuming the vehicle accelerates or decelerates at a constant acceleration a, the following formula can be obtained:

[0112] v(t)=at+v o (6)

[0113]

[0114] Among them, v o is the initial velocity of the vehicle (m / s), v f is the final speed of the vehicle (m / s); d, t f The vehicle has a speed v i Accelerate or decelerate to speed v f Distance traveled (m) and time (s).

[0115] In order to calculate the total electrical energy e (W h) provided by the vehicle battery pack, the efficiency of the power system must also be considered. The efficiency of the vehicle power system is represented by η, and the following formula is obtained:

[0116]

[0117] In summary, the total electric energy e consumed by the vehicle traveling from node p to node q is pq (W·h) is:

[0118]

[0119] in, The electric energy consumed by the vehicle when accelerating, running at a constant speed, and decelerating respectively; The distances traveled by the vehicle during acceleration, constant speed, and deceleration (m) respectively; θ pq is the inclination angle of section pq; R = 0.5C d Aρ;v pq is the speed of vehicles traveling on road section pq (m / s).

[0120] 4-2) Combine the energy consumption calculation formula of the intelligent connected electric vehicle in step 4-1) with the vehicle speed prediction result in step 3) to predict the electric energy consumed by the vehicle when passing.

[0121] The macro-vehicle speed prediction model is used to obtain the predicted value of the vehicle's travel speed on each road section in the target area. At the same time, the vehicle parameters (such as the total mass, frontal area and power system efficiency of the target vehicle) and the collected road data (such as the length of each road section and the road inclination) are combined and input into the energy consumption calculation formula of the intelligent connected electric vehicle to predict the electric energy consumed by the vehicle when passing.

[0122] 5) Design an efficient hierarchical energy-saving path and economical speed collaborative optimization architecture. By considering the coupling relationship between road traffic characteristics and travel speeds on energy-saving paths, the energy economy and travel efficiency of vehicles can be comprehensively improved. This specifically includes an energy-saving path planning decision layer and an economical speed optimization execution layer.

[0123] 6) Solve the optimal energy-saving path R* through the energy-saving path planning decision layer. Figure 3 As shown in FIG, this is a schematic diagram of the optimal energy-saving path R* solved.

[0124] The energy-saving path planning decision layer uses an improved genetic algorithm to solve the optimal energy-saving path R* of the target vehicle based on the collected real road information, combined with the traffic speed prediction model and the energy consumption prediction method of intelligent connected electric vehicles.

[0125] The specific process is as follows:

[0126] 6-1) Use Dijkstra's algorithm to calculate the initial solution:

[0127] Based on the spatial distance of each road section, combined with the traffic speed prediction model and the energy consumption prediction method of intelligent connected electric vehicles, under ideal conditions that ignore the dynamic changes in vehicle speed (including speed adjustments when switching between road sections and speed fluctuations caused by changes in traffic light status), the Dijkstra algorithm is used to solve the path R0 that minimizes the energy consumption of the target vehicle from the starting point to the end point.

[0128] 6-2) Initialize the population:

[0129] Using R0 as the initial path, four new paths are generated through mutation, forming the first generation population, which consists of five individuals. Each individual (i.e., path) is assigned four attributes: travel distance, speed profile, travel time, and total energy consumption. These attributes are calculated as follows:

[0130] ① Travel distance

[0131] The travel distance is directly calculated from the road condition data collected above.

[0132] ② Velocity profile

[0133] The target vehicle's initial speed is set, and it is assumed that upon entering each road segment, the vehicle accelerates or decelerates at a constant rate to the average speed of that section, then continues to travel at that speed until it enters the next road segment or stops. Based on this assumption, once a path is determined and combined with the collected traffic light position and phase information, the speed profile of the target vehicle along that path can be plotted.

[0134] ③ Travel time and total energy consumption

[0135] Based on the designed energy consumption prediction method, the travel time and total energy consumption of the target vehicle when traveling along the path are calculated according to the travel distance and speed profile of the above path.

[0136] Through the above method, each individual in the population is given a complete attribute value, thereby completing the initialization of the population.

[0137] 6-3) Fitness evaluation:

[0138] The total energy consumption of the trip is used as the basis for evaluating individual fitness, and the fitness function is defined as the inverse of the total energy consumption of the trip, ensuring that individuals with lower energy consumption have higher fitness values.

[0139] 6-4) Make a selection:

[0140] From the five individuals of the first generation, the individual with the highest fitness is selected according to the fitness value as the basis for the subsequent crossover operation.

[0141] 6-5) Perform crossover operation:

[0142] Taking the aforementioned high-fitness individuals as the main body, crossover operations are performed with other individuals to generate the first individual of the second generation.

[0143] 6-6) Perform mutation operation:

[0144] The first individual of the second generation is used as the initial path, and four other paths are generated through mutation operations to complete the construction of the second generation population.

[0145] 6-7) Repeat the above steps, generating a new generation of population each time, until the 20th generation.

[0146] 6-8) In the 20th generation, the individual with the highest fitness is selected as the final solution, and its corresponding path is the optimal energy-saving path R* for the target vehicle to travel from the starting point to the end point.

[0147] 7) Use the economic speed optimization execution layer to calculate the economic speed of the vehicle when traveling on path R*.

[0148] Based on the planned optimal energy-saving path R*, the economic speed of the target vehicle when traveling on the path R* is calculated by combining the KKT condition and the gradient descent algorithm. Figure 4 As shown in Figure 2, it is the speed profile of the target vehicle along the optimal energy-saving path R*.

[0149] The specific process is as follows:

[0150] 7-1) Set constraints:

[0151] Assume that the speed range of each road section is [v imin ,v imax ],v imin =max[0.7v i ,v ilow ],v imax =min[1.3v i ,v iup ], where v i is the speed of the target vehicle on each road section, v ilow , v iup are the minimum speed limit and the maximum speed limit of each road section respectively; the total travel time limit of the target vehicle from the starting point to the end point is T lim .

[0152] 7-2) Ignore the road slope of each section and assume that the target vehicle is moving at v i (i=1,2,…n) driving at a constant speed. Based on the above-mentioned intelligent connected electric vehicle energy consumption prediction method and the set constraints, the following objective function is constructed:

[0153]

[0154] Where n represents the number of road segments on the path R*; d i Represents the distance of road segment i.

[0155] 7-3) Combine the KKT condition and the gradient descent algorithm to solve the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint. * ;

[0156]

[0157] HΔ=-J (21)

[0158] According to the KKT condition, the Lagrangian function shown in formula (17) is constructed. To find the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint *, equation (18) should be solved. Since equation (18) is a nonlinear equation system, the Newton iteration method is used to continuously iteratively update the velocity vector v so that it approaches the optimal solution. After constructing the first-order partial derivative matrix (Jacobian matrix) J and the second-order partial derivative matrix (Hessian matrix) H of the Lagrangian function, the iteration step matrix Δ of the Newton iteration method is solved by equation (21). The step matrix Δ is then used to iteratively update the velocity vector v so that it continuously approaches the optimal solution. When the step matrix Δ is less than the set critical value, the iteration terminates. At this time, the velocity vector solution v * That is, it is the optimal economic speed vector that minimizes the objective function while satisfying the travel time constraint.

[0159] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and description merely illustrate the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for collaboratively optimizing energy-saving paths and economical vehicle speed for intelligent connected electric vehicles, characterized by: The following steps are involved: 1) Collect real road information of the target area based on SUMO; 2) Building a macroscopic traffic speed prediction model based on Transformer neural network; 3) Combine road parameters with historical traffic data to train a macro-vehicle speed prediction model, enabling it to predict the speed of target vehicles traveling on various road sections in the target area; 4) Design an energy consumption prediction method for intelligent connected electric vehicles by combining vehicle longitudinal dynamics theory with vehicle speed prediction results; 5) Design an efficient hierarchical energy-saving path and economical speed collaborative optimization architecture; 6) Using the collaborative optimization framework designed in step 5), based on the collected real road information, combined with the traffic speed prediction model and the intelligent connected electric vehicle energy consumption prediction method, an improved genetic algorithm is used to solve the optimal energy-saving path R*; 7) Using the collaborative optimization framework designed in step 5), based on the planned optimal energy-saving path R*, the KKT condition and the gradient descent algorithm are combined to calculate the economic speed of the target vehicle when traveling on path R*.

2. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 1, characterized in that: Step 2) The specific process is as follows: 2-1) Designing a Transformer-based encoder-decoder architecture; 2-2) Configure the multi-head self-attention mechanism; 2-3) Use mean squared error as the main loss function to calculate the mean squared difference between the predicted speed and the true value. 2-4) Add L2 regularization to prevent overfitting. 2-5) Select the Adam optimizer, set the initial learning rate, and configure the cosine annealing learning rate scheduling strategy.

3. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 1, characterized in that: Step 4) The specific process is as follows: 4-1) Based on the theory of vehicle longitudinal dynamics, derive the energy consumption calculation formula for intelligent connected electric vehicles; 4-2) Combine the energy consumption calculation formula of the intelligent connected electric vehicle in step 4-1) with the vehicle speed prediction result in step 3) to predict the electric energy consumed by the vehicle when passing.

4. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 3, characterized in that: The energy consumption calculation formula for intelligent connected electric vehicles in step 4-1) is derived as follows: The basic equation for vehicle longitudinal dynamics is: at(t)=F t -(F f +F w +F i ) (1) Where m is the total mass of the vehicle (kg); a(t) is the instantaneous acceleration of the vehicle (m / s 2 );F t is the driving force of the vehicle; F f 、F w 、F i These are the rolling resistance, air resistance, and slope resistance that a vehicle overcomes when driving on the road; the specific formulas for each force are as follows: F f =fmgcosθ (2) F w =0.5C d Aρv(t) 2 (3) F i =mgsinθ (4) Where g is the gravitational acceleration constant (9.81m / s 2 ); θ is the road inclination angle (°); f is the rolling resistance coefficient; C d is the air resistance coefficient; A is the frontal area of ​​the vehicle (m 2 ); ρ is the air density, generally taken as 1.2258 kg / m 3 ; v(t) is the instantaneous speed of the vehicle; To calculate the instantaneous mechanical power P required to drive the wheels m (t)(W), multiplying Equation (1) by the instantaneous speed of the vehicle v(t) yields: P m (t)=ma(t)v(t)+fmgv(t)cosθ+0.5C d Aρv(t) 3 +mgv(t)sinθ (5); In order to estimate the total mechanical energy e required by the vehicle's power system when it accelerates, decelerates or travels at a constant speed on a certain road m (W·h), integrate Equation (5) over time; assuming the vehicle accelerates or decelerates at a constant acceleration a, the following formula can be obtained: v(t)=at+v o (6) Among them, v o is the initial velocity of the vehicle (m / s), v f is the final speed of the vehicle (m / s); d, t f The vehicle has a speed v i Accelerate or decelerate to speed v f Distance traveled (m) and time (s); In order to calculate the total electrical energy e (W h) provided by the vehicle battery pack, the efficiency of the power system must also be considered. The efficiency of the vehicle power system is represented by η, and the following formula is obtained: In summary, the total electric energy e consumed by the vehicle traveling from node p to node q is pq (W·h) is: in, The electric energy consumed by the vehicle when accelerating, running at a constant speed, and decelerating respectively; The distances traveled by the vehicle during acceleration, constant speed, and deceleration (m) respectively; θ pq is the inclination angle of section pq; R = 0.5C d Aρ;v pq is the speed of vehicles traveling on road section pq (m / s).

5. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds of an intelligent connected electric vehicle according to claim 3, characterized in that: In step 4-2), the macro-vehicle speed prediction model is used to obtain the predicted value of the vehicle's speed on each road section in the target area. At the same time, the vehicle parameters and the collected road data are combined and input into the energy consumption calculation formula of the intelligent connected electric vehicle to predict the electric energy consumed by the vehicle when passing.

6. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 1, characterized in that: The efficient hierarchical energy-saving path and economical speed collaborative optimization architecture in step 5) includes an energy-saving path planning decision layer and an economical speed optimization execution layer.

7. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 6, characterized in that: In step 6), the optimal energy-saving path R* is solved through the energy-saving path planning decision layer. The specific process is as follows: 6-1) Use Dijkstra's algorithm to calculate the initial solution: Based on the spatial distance of each road section, the Dijkstra algorithm is used to solve the path R0 with the minimum energy consumption for the target vehicle from the starting point to the end point, combining the vehicle speed prediction model with the energy consumption prediction method of intelligent connected electric vehicles under ideal conditions ignoring the dynamic change characteristics of vehicle speed. 6-2) Initialize the population: Taking R0 as the initial path, four new paths are generated through mutation operation, thus forming the first generation population, which contains five individual paths in total. Each individual path is assigned four attributes, namely travel distance, speed profile, travel time and total energy consumption. 6-3) Fitness evaluation: The total energy consumption of the trip is used as the basis for evaluating the fitness of the individual. The fitness function is defined as the inverse of the total energy consumption of the trip, ensuring that individuals with lower energy consumption have higher fitness values. 6-4) Make a selection: From the five individuals of the first generation, the individual with the highest fitness is selected according to the fitness value as the basis for the subsequent crossover operation; 6-5) Perform crossover operation: Taking the aforementioned high-fitness individuals as the main body, cross operations are performed with other individuals to generate the first individual of the second generation; 6-6) Perform mutation operation: The first individual of the second generation is used as the initial path, and four more paths are generated through mutation operations to complete the construction of the second generation population. 6-7) Repeat the above steps to generate a new generation of population each time until the 20th generation; 6-8) In the 20th generation, the individual with the highest fitness is selected as the final solution, and its corresponding path is the optimal energy-saving path R* for the target vehicle to travel from the starting point to the end point.

8. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 7, characterized in that: The specific calculation method for the travel distance, speed profile, travel time and total travel energy consumption in step 6-2) is as follows: ① Travel distance The trip distance is directly calculated from the road condition data collected above; ② Velocity profile The target vehicle's initial speed is set, and it is assumed that upon entering each road segment, the vehicle accelerates or decelerates at a constant rate to the average speed of that section, and then continues to travel at that speed until it enters the next road segment or stops. Based on this assumption, once the path is determined and combined with the collected traffic light position and phase information, the speed profile of the target vehicle along that path can be drawn. ③ Travel time and total energy consumption Based on the designed energy consumption prediction method, the travel time and total energy consumption of the target vehicle when traveling along the path are calculated according to the travel distance and speed profile of the above path.

9. The method for collaboratively optimizing energy-saving paths and economical vehicle speeds for an intelligent connected electric vehicle according to claim 6, characterized in that: In step 7), the economic speed optimization execution layer is used to calculate the economic speed of the vehicle when traveling on path R*. The specific process is as follows: 7-1) Set constraints: Assume that the speed range of each road section is [v imin ,v imax ],v imin =max[0.7v i ,v ilow ],v imax =min[1.3v i ,v iup ], where v i is the speed of the target vehicle on each road section, v ilow , v iup are the minimum speed limit and the maximum speed limit of each road section respectively; the total travel time limit of the target vehicle from the starting point to the end point is T lim ; 7-2) Ignore the road slope of each section and assume that the target vehicle is moving at v i (i=1,2,…n) driving at a constant speed, based on the above-mentioned intelligent connected electric vehicle energy consumption prediction method and the set constraints, the following objective function is constructed: Where n represents the number of road segments on the path R*; d i represents the distance of road segment i; 7-3) Combine the KKT condition and the gradient descent algorithm to solve the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint. * ; HΔ=-J (21) In the above, according to the KKT condition, the Lagrangian function shown in formula (17) is constructed to find the optimal economic speed vector v that minimizes the objective function G(v) while satisfying the travel time constraint. * , equation (18) should be solved. Since equation (18) is a nonlinear equation group, the Newton iteration method is used to continuously iteratively update the velocity vector v so that it approaches the optimal solution. After constructing the first-order partial derivative matrix (Jacobian matrix) J and the second-order partial derivative matrix (Hessian matrix) H of the Lagrangian function, the iteration step matrix Δ of the Newton iteration method is solved by equation (21). The step matrix Δ is then used to iteratively update the velocity vector v so that it continuously approaches the optimal solution. When the step matrix Δ is less than the set critical value, the iteration terminates. At this time, the velocity vector solution v * That is, it is the optimal economic speed vector that minimizes the objective function while satisfying the travel time constraint.

Images