A vehicle reference speed generation method based on dynamic programming

Abstract

The present application relates to a kind of vehicle reference speed generation methods based on dynamic programming, method includes the following steps: S1, constructs state transition model, cost function and cost function, obtains terminal state, from terminal state reverse traversal to initial state, calculate the minimum cost of all states and optimal control input as DP optimization result in traversal process;S2, obtain actual road segment, extract slope data from actual road segment, identify slope top position and slope bottom position based on slope data, calculate slope bottom rush slope parameter based on DP optimization result;S3, based on slope bottom rush slope parameter, calculate slope top speed, left boundary speed calculation case and right boundary speed calculation case, carry out smooth interpolation and obtain the reference speed corresponding to the road working condition in front. Compared with prior art, the present application has the advantages of balancing energy consumption optimization and power output.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to a method for generating a vehicle reference speed based on dynamic programming. Background Art

[0002] With the development of intelligent driving technology, vehicle energy efficiency and safety have become core concerns in the automotive industry. Properly planning the vehicle's reference speed during driving can not only effectively reduce fuel consumption but also improve driving safety and ride comfort. Dynamically adjusting the vehicle's reference speed is particularly important for energy-efficient driving, especially in scenarios such as long-distance driving and complex slopes. However, existing reference speed generation methods have significant limitations when dealing with complex driving conditions.

[0003] Existing reference speed generation technologies typically use fixed parameters or simple rule-based models. These methods can be broadly categorized as fixed speed planning, rule-based speed generation, and static optimization-based speed planning. These methods often fail to adjust the reference speed based on real-time road conditions and vehicle status when handling dynamic driving conditions, resulting in increased fuel consumption, insufficient power, and reduced driving comfort.

[0004] Optimizing power demand and fuel economy on slopes is particularly complex. For example, improperly planning power demand at the bottom of a slope can lead to difficulty climbing, while excessively high or low speeds at the top of the slope can lead to inefficient energy utilization. Furthermore, the smoothness of speed adjustments significantly impacts the driving experience and the stability of the autonomous driving system.

[0005] The reference vehicle speed generated by the existing method cannot balance energy consumption optimization and power output, resulting in high energy consumption and low power utilization. Summary of the Invention

[0006] The purpose of the present invention is to provide a vehicle reference speed generation method based on dynamic programming in order to balance energy consumption optimization and power output.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A method for generating a vehicle reference speed based on dynamic programming, the method comprising the following steps:

[0009] S1. Construct a state transition model, cost function, and price function, obtain the terminal state, traverse backward from the terminal state to the initial state, and calculate the minimum cost and optimal control input of all states as the DP optimization result during the traversal process;

[0010] S2. Obtaining an actual road condition segment, extracting slope data from the actual road condition segment, identifying the top and bottom of the slope based on the slope data, and calculating the slope bottom impact parameter based on the DP optimization result;

[0011] S3. Calculate the top speed, left boundary speed, and right boundary speed based on the slope bottom attack parameters, and perform smooth interpolation to obtain a reference speed corresponding to the road condition ahead.

[0012] Furthermore, the state transition model is:

[0013]

[0014] v t+1 =v t +a t ·Δt

[0015] Among them, s t+1 is the vehicle position at time t+1, s t is the vehicle position at time t, v t is the vehicle speed at time t, Δt is the time change, a t is the vehicle acceleration at time t;

[0016] The cost function is:

[0017]

[0018] Among them, F represents the driving force, L represents the cost function, and J * represents the cost function.

[0019] Furthermore, the identification of the top and bottom of the slope is specifically as follows:

[0020] The top of the slope in the slope data is extracted by extracting the maximum value. The identification threshold is defined as 50m. If the maximum altitude in the position interval [s(i)-50, s(i)+50] is h(i), then the position of s(i) is considered to be a top of the slope. t (k), the corresponding slope top altitude is h(i), let the number of identified slope tops be N, and the slope top position and slope top altitude be s respectively t (k) and h t (k), where i represents the index of the location point, i = 1, 2, ..., 250, k = 1, 2, ..., N;

[0021] For the 2nd, 3rd, ..., Nth uphill conditions, the corresponding slope bottom is obtained by identifying the minimum altitude between two adjacent slope top positions, that is: for k = 2, 3, ..., N, if the position interval [s t (k-1),s t (k)] is h(i), then the position of s(i) is considered to be the kth slope bottom, that is, the slope bottom position s b (k) = s(i), slope bottom elevation h b (k)=h(i),

[0022] For the first uphill condition, the first slope bottom is obtained by identifying the minimum altitude between the 5km starting point and the first slope top position, that is: if the position interval [s(1),s t (1)] The maximum altitude is h(i), then the position of s(i) is considered to be the kth slope bottom, that is, the slope bottom position s b (1) = s(i), slope bottom elevation h b (1) = h(i).

[0023] Furthermore, the slope bottom parameters are calculated based on the DP optimization results as follows:

[0024] The uphill working condition parameters are calculated based on the top position, top altitude, bottom position and bottom altitude of the slope. The uphill working condition parameters include the horizontal displacement change Δs(k) and the elevation change Δh(k), where k = 1, 2, ..., N. Then, the slope bottom rushing parameters are calculated based on the uphill working condition parameters. The slope bottom rushing parameters include the rushing vehicle speed v under different uphill working conditions. b (k) and slope Δv b , the slope impact amount is obtained by fitting the DP optimization result.

[0025] Furthermore, the slope speed v b (k) is:

[0026] v b (k)=v set +max[0.9ω v (v max -v set ),Δv b (k)]

[0027] Among them, the horizontal displacement change Δs(k), the elevation change Δh(k), the expected speed v set by the driver set , maximum vehicle speed v max , vehicle dynamic fuel consumption weight ω v , slope Δv b .

[0028] Furthermore, the top speed is:

[0029]

[0030] Among them, the speed of the vehicle going uphill v b (k), the horizontal displacement change Δs(k), the resultant force F acting on the vehicle during the k-th uphill process up (k), vehicle mass M.

[0031] Furthermore, the resultant force F acting on the vehicle during the kth uphill process is up(k) is:

[0032]

[0033] Where g is the acceleration due to gravity, f is the wheel rolling resistance coefficient, and C D is the air resistance coefficient, A is the truck's frontal area, ρ is the air density, the horizontal displacement change Δs(k), the elevation change Δh(k), and the desired speed v set by the driver set .

[0034] Furthermore, the left boundary speed is calculated as follows:

[0035] If the first slope bottom is not the left boundary value, the left boundary speed calculation case F Vstart =1; otherwise, the left boundary speed calculation situation F Vstart =0.

[0036] Furthermore, the right boundary speed is calculated as follows:

[0037] If the last slope top is not the right boundary value, the right boundary speed calculation case F Vend =1; otherwise, the right boundary speed calculation situation F Vend =0.

[0038] Furthermore, the reference vehicle speed is obtained by interpolation based on the feature point position sequence and the vehicle speed sequence, which are:

[0039] If F Vstart =1, F Vend =1, then the feature point position sequence S interp With the speed sequence V interp for:

[0040] S interp =[s(1);s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N);s(250)] (2N+2)×1

[0041] V interp =[v start ;v b (1);v t (1);v b (2);v t (2);…;v b (N);v t (N);v end ] (2N+2)×1

[0042] Among them, s(250) is the end point, v start is the left boundary speed, v end is the right boundary speed;

[0043] If F Vstart =1, F Vend = 0, then the feature point position sequence S interp With the speed sequence V interp for:

[0044] S interp =[s(1);s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N)] (2N+1)×1

[0045] V interp =[v start ;v b (1);v t (1);v b (2);v t (2);…;v b (N);v t (N)] (2N+1)×1

[0046] If F Vstart =0, F Vend =1, then the feature point position sequence S interp With the speed sequence V interp for:

[0047] S interp =[s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N);s(250)] (2N+1)×1

[0048] V interp =[v b (1);v t (1);v b (2);v t (2);…;v b (N);v t (N);v end ] (2N+1)×1

[0049] If F Vstart =0, F Vend = 0, then the feature point position sequence S interp With the speed sequence V interp for

[0050] S interp =[s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N)] 2N×1

[0051] V interp =[v b (1);v t (1);v b (2);v t (2);…;v b (N); v t (N)] 2N×1 .

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The present invention obtains the slope data of the road ahead of the vehicle in real time, analyzes it in combination with historical data, and dynamically generates a reference speed for the future road section. In the speed optimization calculation, the speed at the bottom of the slope and the speed at the top of the slope are generated based on the dynamic programming algorithm to meet the power and fuel economy requirements. The generated reference speed actively reduces the speed at the top of the slope in uphill conditions to avoid insufficient power and reduce fuel consumption; in downhill conditions, the utilization efficiency of gravitational potential energy is optimized by adjusting the speed at the bottom of the slope in advance, thereby further reducing energy consumption. Therefore, the present invention is particularly suitable for complex slope conditions. By optimizing the vehicle reference speed in real time, fuel economy can be significantly improved while ensuring driving safety and comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Reconstructing the schematic diagram for the map;

[0055] Figure 2 This is a schematic diagram for slope top identification;

[0056] Figure 3 This is a schematic diagram for identifying the bottom of the slope;

[0057] Figure 4 The speed calculation for the vehicle to rush up the slope indicates the amount of advance rushing up the slope;

[0058] Figure 5 The MAP indicates the vehicle speed when climbing a hill.

[0059] Figure 6 This is a schematic diagram for calculating vehicle speed at the top of a slope;

[0060] Figure 7 This is a schematic diagram of the left boundary speed calculation;

[0061] Figure 8 It is a schematic diagram of vehicle speed at the right boundary;

[0062] Figure 9 Obtain reference vehicle speed diagram for smooth interpolation;

[0063] Figure 10 is a flow chart of the present invention;

[0064] Figure 11 Prepare a general design plan for speed planning;

[0065] Figure 12 This is a diagram of the real-time calculation steps for the reference vehicle speed. DETAILED DESCRIPTION

[0066] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0067] The present invention proposes a method for generating a vehicle reference speed based on dynamic programming. By acquiring the slope data of the road ahead of the vehicle in real time and analyzing it in combination with historical data, the reference speed of the future road section is dynamically generated. The method includes the following steps: real-time map data reconstruction, ramp condition identification, vehicle speed optimization calculation and reference speed generation. In real-time map data reconstruction, the slope data for the next 2 kilometers is acquired and combined with the historical 3-kilometer data to accurately identify the top and bottom of the slope, providing reliable input for subsequent optimization. In ramp condition identification, the local extreme value method is used to extract the top and bottom of the slope, and calculate ramp parameters such as horizontal displacement and elevation change. In vehicle speed optimization calculation, the ramp speed at the bottom of the slope and the speed at the top of the slope are generated based on the dynamic programming algorithm to meet the power and fuel economy requirements. Finally, a continuous and smooth reference speed curve is generated by the feature point interpolation method. Figure 10 is a flow chart of the present invention; Figure 11 Prepare a general design plan for speed planning; Figure 12 This is a diagram of the real-time calculation steps for the reference vehicle speed.

[0068] The present invention is particularly suitable for complex slope conditions. By optimizing the vehicle reference speed in real time, it can significantly improve fuel economy while ensuring driving safety and comfort. The method includes the following steps:

[0069] S1. Construct a state transition model, cost function, and price function, obtain the terminal state, traverse backward from the terminal state to the initial state, and calculate the minimum cost and optimal control input of all states as the DP optimization result during the traversal process;

[0070] S2. Obtaining an actual road condition segment, extracting slope data from the actual road condition segment, identifying the top and bottom of the slope based on the slope data, and calculating the slope bottom impact parameter based on the DP optimization result;

[0071] S3. Calculate the top speed, left boundary speed, and right boundary speed based on the slope bottom attack parameters, and perform smooth interpolation to obtain a reference speed corresponding to the road condition ahead.

[0072] In the actual experiment, through map reconstruction, the slope data is obtained and stored in real time, including the position data and elevation data of the previous 2km and the historical 3km. The data storage interval is 20m, that is. Therefore, the dimensions of the position data and elevation data are both. The reason for storing the historical 3km data is to more accurately identify the bottom and top of the slope near the current position in the future. Figure 1 As shown. In this case, the method of the present invention comprises the following steps:

[0073] S1. Dynamic programming

[0074] The dynamic programming module calculates the optimal vehicle control strategy under different operating conditions through state transition, cost function calculation and cost function recursion, with the goal of minimizing vehicle fuel consumption:

[0075] State transition model: This describes the process of the vehicle transitioning from its current position and velocity to the next state. The control input is the driving force, and the state transition equation comprehensively considers the vehicle's gravity, air resistance, and rolling resistance.

[0076] Cost function calculation: Calculate the vehicle's instantaneous fuel consumption based on speed and driving force interpolation;

[0077] Cost function calculation and reverse recursion: By recursively calculating the cumulative cost from the terminal state to the initial state, reversely traversing each state, and generating the optimal control strategy.

[0078] Finally, the module outputs the optimal driving force and vehicle speed data for working conditions such as from the bottom of the slope to the top of the slope and from the top of the slope to the boundary, providing basic data for subsequent calculations.

[0079] S2. Slope working condition identification

[0080] S2.1. Extracting slope data

[0081] Combine the real-time collected 2 km forward slope data with the historical 3 km elevation data to generate a unified road condition segment.

[0082] S2.2. Identify the top of the slope

[0083] The slope top position is identified by the local extreme value method, and the effective threshold (altitude difference greater than 50 meters) is set to extract the slope top position and its corresponding altitude value.

[0084] S2.3. Identify the bottom of the slope

[0085] According to the positions of adjacent slope tops, the local minimum method is used to identify the slope bottom position and its parameters, and the horizontal displacement (Δs) and elevation change (Δh) of the slope are calculated.

[0086] S2.4. Calculate the slope bottom parameters

[0087] Based on the slope parameters Δs and Δh, combined with vehicle mass, driving mode, dynamic weight and expected vehicle speed, dynamic programming optimization is used to generate the vehicle speed at the bottom of the slope. Nine typical slope conditions are classified (such as short slopes, long slopes, etc.) to ensure data accuracy.

[0088] S3, vehicle speed optimization calculation

[0089] S3.1. Calculate the speed at the top of the slope

[0090] Based on the speed at the bottom of the slope, the speed at the top of the slope is calculated using the kinetic energy theorem, combined with the vehicle's gravity component, rolling friction, and wind resistance, to minimize the use of driving force.

[0091] S3.2. Calculate the left boundary speed

[0092] When the starting point is downhill within 5 kilometers, calculate the left boundary speed at the starting point.

[0093] S3.3. Calculate the left and right boundary speeds

[0094] When the end point within 5 kilometers is a downhill condition, calculate the right boundary speed at the end point.

[0095] According to different working conditions and combined with boundary conditions, the optimal vehicle speed range for each state is calculated.

[0096] S3.4. Reference speed generation

[0097] Define vehicle speed feature points: Use the speed at the bottom of the slope, the speed at the top of the slope, the speed at the left boundary, and the speed at the right boundary as interpolation feature points;

[0098] Generate a smooth reference speed curve: Use linear interpolation to generate a smooth reference speed curve within a 5-kilometer range based on feature points.

[0099] Limiting processing: Limiting the maximum and minimum speeds of the generated reference speed curve to ensure safety and comfort;

[0100] Output reference vehicle speed: ultimately generate a reference vehicle speed curve for the next 2 kilometers, and record road condition analysis data within a 5-kilometer range to provide real-time input for the vehicle control module.

[0101] Dynamic programming is used to calculate the optimal vehicle control strategy under different operating conditions to minimize total fuel consumption. The core of the DP algorithm consists of three parts: state transition, cost function calculation, and price function calculation. Specifically:

[0102] S1.1 State Transition Model

[0103] The state transition model is the core of dynamic programming. It describes how the system transitions from one state to the next. In this algorithm, the state consists of the vehicle's position s and velocity v, and the control input is the driving force F.

[0104] The state transition equation is as follows:

[0105]

[0106] v t+1 =v t +a t ·Δt

[0107] The acceleration a is calculated by the following formula:

[0108]

[0109] -F is the driving force.

[0110] -M is the vehicle mass.

[0111] -θ is the road slope.

[0112] -f is the rolling resistance coefficient.

[0113] -g is the acceleration due to gravity.

[0114] -ρ is the air density.

[0115] -A f is the frontal area of ​​the vehicle.

[0116] -C d is the coefficient of air resistance.

[0117] -v is the vehicle speed.

[0118] These equations describe how the vehicle moves from its current state (s t ,v t )Transfer to the next state (s t+1 ,v t+1 ).

[0119] S1.2. Cost function calculation

[0120] The cost function L represents the fuel consumption rate of the vehicle in each state. The fuel consumption rate is determined by the vehicle's speed and driving force. We calculate the fuel consumption rate mf through interpolation:

[0121] L(s,v,F)=mf

[0122] The cost function calculates the fuel consumption at each state and control input combination, and the ultimate goal is to find a control policy that minimizes the total fuel consumption.

[0123] S1.3. Cost function calculation

[0124] The cost function J represents the cumulative cost from the initial state to the terminal state. To minimize the total fuel consumption, we use the following recursive formula for calculation:

[0125]

[0126] Initial condition: The cost of the terminal state is set to 0:

[0127] J * (s T ,v T )=0

[0128] In this formula, J * (s t ,v t ) is in the state at time t (s t ,v t ) under the minimum cost. By minimizing L(s t ,v t ,F) and the cost of the next state J * (s t+1 ,v t+1 ), we can get the optimal control input F for the current state.

[0129] S1.4. Backward Recursion

[0130] Dynamic programming uses a reverse recursive approach, starting from the terminal state and gradually backtracking to the initial state. The specific steps are as follows:

[0131] 1. Initialization: Set the cost of the terminal state to 0.

[0132] 2. Backward traversal: Starting from the terminal state, gradually backtrack to the initial state. At each step, calculate the minimum cost and optimal control input for the current state.

[0133] 3. Update the cost function: In each step, the cost function and the optimal control strategy are updated according to the recursive formula.

[0134] S1.5. Calculate the optimal control strategy

[0135] By recursively working backward, we finally obtain the optimal control policy from the initial state to the terminal state. This policy describes the optimal driving force that the vehicle should adopt in each state to minimize the total fuel consumption.

[0136] The dynamic programming module defines a state transition model, a cost function, and a penalty function, and uses a backward recursive approach to calculate the optimal vehicle control strategy. This strategy effectively minimizes fuel consumption under various operating conditions and provides the foundation for subsequent speed calculations for hill climbing.

[0137] S2. Slope working condition identification

[0138] S2.1. Extracting slope data

[0139] The real-time data acquisition module obtains the slope data of the 2 kilometers ahead of the vehicle and the historical elevation data of the 3 kilometers, and generates a unified road condition segment.

[0140] S2.2. Identify the top of the slope

[0141] According to the stored slope data, the top position and the top elevation value of the slope are identified.

[0142] The slope top in the stored slope data is extracted by extracting the maximum value (such as Figure 2 As shown), and the identification threshold is defined as 50m when identifying the maximum value, that is: if the maximum altitude in the position interval [s(i)-50, s(i)+50] is h(i), then the position of s(i) is considered to be a slope top, and the corresponding slope top altitude is h(i). In particular, for the farthest distance of the front slope data, we judge: if the maximum altitude in the position interval [s(250)-50, s(250)] is h(250), then the position of s(250) is considered to be a slope top, and the corresponding slope top altitude is h(250). Assume that the number of identified slope tops is N, and the slope top position and slope top altitude are s respectively. t (k) and h t (k), where k = 1, 2,…, 250.

[0143] S2.3. Identify the bottom of the slope

[0144] Identify the bottom position and corresponding parameters of the road condition ahead. Here we mainly focus on the uphill process, so for the identified top, we need to identify the bottom position and altitude corresponding to the top position. From the bottom to the top of the slope constitutes a complete uphill condition, such as Figure 3 shown.

[0145] 1) For the 2nd, 3rd, ..., Nth uphill conditions, we obtain the corresponding slope bottom by identifying the minimum altitude between two adjacent slope top positions, that is: for k = 2, 3, ..., N, if the position interval [s t (k-1),s t (k)] is h(i), then the position of s(i) is considered to be the kth slope bottom, that is, the slope bottom position s b (k) = s(i), slope bottom elevation h b (k) = h(i).

[0146] (2) For the first uphill condition, we obtain the first slope bottom by identifying the minimum altitude between the 5km starting point and the first slope top position, that is: if the position interval [s(1),s t (1)] The maximum altitude is h(i), then the position of s(i) is considered to be the kth slope bottom, that is, the slope bottom position s b (1) = s(i), slope bottom elevation h b (1) = h(i).

[0147] For each uphill condition within 5 km, the uphill condition parameters are calculated based on the position and elevation of the top and bottom of the slope, including the horizontal displacement change Δs(k) and the elevation change Δh(k), where k = 1, 2, ..., N, that is,

[0148] Δs(k)=s t (k)-s b (k),Δh(k)=h t (k)-h b (k)

[0149] S2.4. Calculate the slope bottom parameters

[0150] For each uphill condition within 5 km, according to the uphill condition parameters, the uphill condition parameters Δs(k) and Δh(k), the expected speed v set by the driver set , maximum vehicle speed v max , vehicle mass M and vehicle dynamic fuel consumption weight ω v Calculate the speed v of the vehicle at the bottom of the slope b (like Figure 4 shown):

[0151] v b (k) = f Vb (Δs(k),Δh(k),v set ,v max ,M,ω v ),k=1,2,…,N

[0152] The calculation process is:

[0153] vb (k)=v set +max[0.9ω v (v max -v set ),Δv b (k)],k=1,2,…,N

[0154] Slope amount, and vehicle dynamic fuel consumption weight ω v ∈[0,1], ω v The larger it is, the greater the average speed of the vehicle when rushing up the slope, and the stronger the power; conversely, the smaller the average speed of the vehicle when rushing up the slope, the more attention is paid to the vehicle's driving economy.

[0155] In Δv b During the calculation process, Δs(k) and Δh(k) are used as independent variables, and the ramp speed MAP table is calibrated by dynamic programming method for different vehicle masses M, as shown in the following example: Figure 5 As shown. In the process of formulating the MAP table, 9 different uphill conditions are defined, namely:

[0156] (1) 1- Short slope: Δs≤500m, Δh≤5m

[0157] (2) 2- Medium-long small slope: 500m<Δs≤1500m, Δh≤5m

[0158] (3) 3-Long small slope: Δs>1500m, Δh≤5m

[0159] (4) 4-Short medium slope: Δs≤500m, 5m<Δh≤15m

[0160] (5) 5-Medium-long medium slope: 500m<Δs≤1500m, 5m<Δh≤15m

[0161] (6) 6-Long medium slope: Δs>1500m, 5m<Δh≤15m

[0162] (7) 7- Short steep slope: Δs≤500m, Δh>15m

[0163] (8) 8- Medium to long slope: 500m<Δs≤1500m, Δh>15m

[0164] (9) 9-Long and steep slope: Δs>1500m, Δh>15m

[0165] For smaller slopes (1-short slope, 2-medium-long slope), the slope impact Δv b The minimum value is 0.9ω v (v max -v set), for larger slopes (6-long medium slope, 7-short large slope, 8-medium long slope, 9-long long slope); the slope impact Δv b is a quadratic function of the average slope Δs / Δh; for the remaining slopes (3-long small slope, 4-short medium slope, 5-medium long medium slope); the slope amount Δv b is a quadratic function of Δh. The above quadratic functions are all obtained by fitting the DP optimization results.

[0166] S3, vehicle speed optimization calculation

[0167] S3.1. Calculate vehicle speed at the top of the slope

[0168] Calculate the corresponding speed at the top of the slope by calculating the speed at the bottom of the slope, such as Figure 6 As shown. Here, it is assumed that the vehicle accelerates as little as possible during the uphill climb (i.e., it uses as little driving force as possible and tries to reach the top of the slope by sliding). Therefore, it is assumed that the vehicle is only subject to the following three resistances during the uphill climb: (1) gravity, (2) rolling friction, and (3) wind resistance. The speed at the top of the slope is calculated using the kinetic energy theorem:

[0169] During the kth uphill process, the vehicle is subjected to the resultant force F up (k)

[0170]

[0171] Where g is the acceleration due to gravity, f is the wheel rolling resistance coefficient, C D is the air resistance coefficient, A is the truck's frontal area, and ρ is the air density. set The driver sets the vehicle speed v set Instead of actual vehicle speed, the average slope Δs / Δh is used to simplify the calculation of the gravity component and rolling resistance.

[0172] like It means that in the absence of driving force, the vehicle will slow down to below the minimum value at the top of the slope, so the top speed v t (k)=v min ;otherwise, The speed at the top of the slope is

[0173]

[0174] S3.2. Calculate the left boundary speed

[0175] If s b (1)≠s(1), that is, the first slope bottom within 5 km is not the left boundary value, which means that the starting point within 5 km is in a downhill condition. Calculate the vehicle speed at the left boundary of 5 km (that is, the speed at s(1)), as follows: Figure 7 Assume that the vehicle has no driving force during this process, that is, the resultant force is

[0176]

[0177] (1) If Then it means that in the position interval [s(1),s b (1)], in the absence of driving force, the vehicle speed at the vehicle position s(1) should be below the minimum value, then the left boundary speed v is defined as start is the minimum speed, that is, v start =v min .

[0178] (2) If Then it means that in the position interval [s(1),s b (1)], in the absence of driving force, the vehicle speed at the vehicle position s(1) should be above the maximum value, then the left boundary speed v is defined as start is the maximum vehicle speed, that is, v start =v max .

[0179] (3) If Then it means that in the position interval [s(1),s b (1)], in the absence of driving force, the vehicle speed at the vehicle position s(1) is within the speed setting interval [v min ,v max ], then calculate the left boundary speed v start for

[0180]

[0181] In order to facilitate subsequent interpolation calculations, we use the variable F Vstart Indicates the calculation of the left boundary speed, that is: if s b (1)≠s(1), then F Vstart =1; otherwise F Vstart =0.

[0182] S3.3. Calculate right boundary speed

[0183] If s t (N)≠s(250), that is, the last slope top within 5km is not the right boundary value, which means that the end point within 5km is in a downhill condition. Calculate the right boundary speed of 5km (that is, the speed at s(250)), as follows: Figure 8 Assume that the vehicle has no driving force during this downhill section, that is, the resultant force is

[0184]

[0185] (1) If It means that in the position interval [s t(N),s(250)], in the absence of driving force, the vehicle reaches the terminal position [s t (N),s(250)], the vehicle speed will be below the minimum value, so the right boundary speed v is defined end is the minimum speed, that is, v end =v min .

[0186] (2) If It means that in the position interval [s t (N),s(250)], in the absence of driving force, the vehicle reaches the terminal position [s t (N),s(250)], the vehicle speed will be above the maximum value, so the right boundary speed v is defined end is the maximum vehicle speed, that is, v end =v max .

[0187] (3) If It means that in the position interval

[0188] [s t (N),s(250)], in the absence of driving force, the vehicle reaches the terminal position [s t (N),s(250)], the vehicle speed will be in the vehicle speed setting range [v min ,v max ], then calculate the right boundary speed v end for

[0189]

[0190] In order to facilitate subsequent interpolation calculations, we use the variable F Vend Indicates the calculation of the right boundary speed, that is: if s t (N)≠s(250), then F Vend =1; otherwise F Vend =0.

[0191] S3.4. Obtain the reference speed corresponding to the road conditions ahead by smooth interpolation

[0192] According to the speed of the bottom, top, left and right boundaries of the slope, a feature point position sequence S for speed interpolation is established. interp With the speed sequence V interp , the number of feature points is 2N+F Vstart +F Vend .

[0193] (1) If F Vstart =1, F Vend =1, then the feature point position sequence S interp With the speed sequence V interpfor

[0194] S interp =[s(1);s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N);s(250)] (2N+2)×1

[0195] V interp =[v start ;v b (1);v t (1);v b (2);v t (2);…;v b (N); v t (N); v end ] (2N+2)×1

[0196] (2) If F Vstart =1, F Vend = 0, then the feature point position sequence S interp With the speed sequence V interp for

[0197] S interp =[s(1);s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N)] (2N+1)×1

[0198] V interp =[v start ;v b (1);v t (1);v b (2);v t (2);…;v b (N); v t (N)] (2N+1)×1

[0199] (3) If F Vstart =0, F Vend =1, then the feature point position sequence S interp With the speed sequence V interp for

[0200] S interp =[s b (1);st (1);s b (2);s t (2);…;s b (N);s t (N);s(250)] (2N+1)×1

[0201] V interp =[v b (1);v t (1);v b (2);v t (2);…;v b (N);v t (N);v end ] (2N+1)×1

[0202] (4) If F Vstart =0, F Vend = 0, then the feature point position sequence S interp With the speed sequence V interp for

[0203] S interp =[s b (1);s t (1);s b (2);s t (2);…;s b (N);s t (N)] 2N×1

[0204] V interp =[v b (1);v t (1);v b (2);v t (2);…;v b (N); v t (N)] 2N×1

[0205] Through the feature point position sequence S interp With the speed sequence V interp , interpolate the displacement data s(i) stored within 5km, i=1,2,…,250 corresponding to the reference vehicle speed data v ref (i), i=1,2,…,250 (such as Figure 9 ), which is used in the subsequent vehicle speed tracking module.

[0206] To verify the energy-saving effect of the reference speed module and its impact on vehicle dynamics, this report analyzes multiple working conditions based on the Simulink simulation platform. The simulation benchmark uses fixed cruise speed control, which does not dynamically adjust with road conditions; after adding the reference speed module, the reference speed is dynamically adjusted using the slope information in front of the vehicle, thereby optimizing fuel economy and power output.

[0207] The simulation results show that under different vehicle speeds and weights, the reference speed module significantly reduces fuel consumption. For a fixed vehicle weight of 49t, at a constant speed of 80kph, fuel consumption is reduced by 0.2%, at a constant speed of 70kph and 60kph, fuel consumption is reduced by 0.9% and 2.1% respectively; for a fixed speed of 80kph, under different vehicle weights, fuel consumption is reduced by 0.6%(15t), 2.3%(30t) and 0.2%(49t) respectively. The reference speed module optimizes the power demand in the slope condition by dynamically adjusting the top-of-slope speed and the bottom-of-slope speed, while having minimal impact on the average vehicle speed.

[0208] Typical working condition analysis shows that the reference speed module actively reduces the top-of-slope speed in uphill conditions to avoid power deficiency and reduce fuel consumption; in downhill conditions, it optimizes the utilization efficiency of gravitational potential energy by adjusting the bottom-of-slope speed in advance, further reducing energy consumption. In summary, the reference speed module balances energy consumption optimization and power output, significantly improving fuel economy.

[0209] The above describes the preferred embodiments of the present application in detail. It should be understood that those skilled in the art can make many modifications and changes without creative labor based on the concept of the present application. Therefore, any technical solution obtained by logical analysis, reasoning or limited experiment based on the existing technology according to the concept of the present application shall be within the protection scope determined by the claims.

Claims

1. A method for generating vehicle reference speed based on dynamic programming, characterized in that: The method comprises the following steps: S1. Construct a state transition model, cost function, and price function, obtain the terminal state, traverse backward from the terminal state to the initial state, and calculate the minimum cost and optimal control input of all states as the DP optimization result during the traversal process; S2. Obtaining an actual road condition segment, extracting slope data from the actual road condition segment, identifying the top and bottom of the slope based on the slope data, and calculating the slope bottom impact parameter based on the DP optimization result; S3. Calculate the top speed, left boundary speed, and right boundary speed based on the slope bottom overshoot parameters, and perform smooth interpolation to obtain a reference speed corresponding to the road condition ahead; The cost function represents the fuel consumption rate of the vehicle in each state; the price function represents the cumulative cost from the initial state to the terminal state; The identification of the top and bottom of the slope is specifically as follows: The top of the slope in the slope data is extracted by extracting the maximum value, and the identification threshold is defined as 50m. The maximum altitude within , then it is believed that The location is a slope top , the corresponding slope top elevation is , let the number of identified slope tops be , the top position and top altitude are and ,in i Indicates the index of the location point, , ; For the For an uphill working condition, the corresponding slope bottom is obtained by identifying the minimum altitude between two adjacent slope top positions, that is: , if the position interval The minimum internal altitude is , then it is believed that The location is The bottom of the slope, that is, the position of the bottom of the slope , slope bottom elevation , For the first uphill condition, the first slope bottom is obtained by identifying the minimum altitude between the 5km starting point and the first slope top position, that is: if the position interval The minimum internal altitude is , then it is believed that The location is The bottom of the slope, that is, the position of the bottom of the slope , slope bottom elevation The 5km includes the previous 2km and the previous 3km. The slope bottom parameters calculated based on the DP optimization results are as follows: The uphill working condition parameters are calculated based on the top position, top elevation, bottom position and bottom elevation of the slope. The uphill working condition parameters include the horizontal displacement change. and elevation change ,in , then calculate the slope bottom rushing parameters based on the uphill working condition parameters, the slope bottom rushing parameters include the rushing speed of the vehicle under different uphill working conditions and slope , the slope amount Obtained by fitting the DP optimization results; The left boundary speed is calculated as follows: If the first slope bottom is not the left boundary value, the left boundary speed calculation ; Otherwise, the left boundary speed calculation ; The right boundary speed is calculated as follows: If the last slope top is not the right boundary value, the right boundary speed calculation situation ; Otherwise, the right boundary speed calculation ; The reference speed is obtained by interpolation based on the feature point position sequence and the speed sequence. The feature point position sequence and the speed sequence are: like , , then the feature point position sequence Speed ​​sequence for: in, For the end point, is the left boundary speed, is the right boundary speed; like , , then the feature point position sequence Speed ​​sequence for: like , , then the feature point position sequence Speed ​​sequence for: like , , then the feature point position sequence Speed ​​sequence for ； in Indicates the speed at the top of the slope. .

2. The method for generating a vehicle reference speed based on dynamic programming according to claim 1, characterized in that: The state transition model is: in, is the vehicle position at time t+1, is the vehicle position at time t, is the vehicle speed at time t, is the time variation, is the vehicle acceleration at time t; The cost function is: in, Indicates driving force, represents the cost function, represents the cost function.

3. The method for generating a vehicle reference speed based on dynamic programming according to claim 1, characterized in that: The speed of the vehicle when climbing the slope for: The expected speed set by the driver , maximum vehicle speed , vehicle dynamic fuel consumption weight , slope .

4. The method for generating a vehicle reference speed based on dynamic programming according to claim 3, characterized in that: The speed at the top of the slope is: Among them, the speed of the vehicle going uphill , horizontal displacement change , No. The total force on the vehicle during the uphill process , vehicle quality .

5. The method for generating a vehicle reference speed based on dynamic programming according to claim 4, characterized in that: No. The total force on the vehicle during the uphill process for: in, is the acceleration due to gravity, is the wheel rolling resistance coefficient, is the air resistance coefficient, is the frontal area of ​​the vehicle, is the air density, the change in horizontal displacement , elevation change , the desired speed set by the driver .

Images