Path prediction and energy supplement scheduling method and system for unmanned electric heavy truck

Abstract

The disclosure provides a path prediction and energy supplement scheduling method and system for unmanned electric heavy trucks, relating to the technical field of intelligent expressways, comprising: constructing a multivariate coupling energy consumption prediction model to predict path energy consumption; introducing path energy consumption into the fast energy supplement process at the end of the fast charging station, performing source-network-load-storage collaborative control, and obtaining flexible scheduling results of the fast charging station group; constructing a parking control model, introducing the flexible scheduling results of the fast charging station group into the parking control model, using a fusion SLAM algorithm based on LiDAR and vision for positioning, generating a parking path, and performing accurate energy charging through charging port identification and attitude feedback; constructing a vehicle-pole-network-cloud collaborative optimization scheduling model, obtaining the actual charging power in the energy charging process, inputting the actual charging power into the vehicle-pole-network-cloud collaborative optimization scheduling model, constructing an objective function, solving the optimal collaborative optimization scheduling result, and combining green electricity utilization rate to calculate carbon emission reduction, thereby realizing the path prediction and energy supplement accurate scheduling process of unmanned electric heavy trucks.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent highway technology, specifically to a method and system for path prediction and energy replenishment scheduling for unmanned electric heavy trucks. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] With the advancement of smart highways and the evolution of trunk logistics towards electrification, intelligence, and automation, unmanned electric heavy-duty trucks, as a new generation of green transportation equipment, are gradually being promoted in highway and long-distance logistics scenarios. However, there are still many technical bottlenecks in the current operation of unmanned electric heavy-duty trucks, specifically in the following aspects: (1) Insufficient accuracy of energy consumption prediction affects route scheduling and range assurance. Unmanned electric heavy-duty trucks face challenges in long-haul transportation, including long travel distances, frequent load changes, and complex and variable road conditions. Energy consumption is highly nonlinear and dynamic, influenced by factors such as road gradient, traffic conditions, weather conditions, motor efficiency, and vehicle load. Existing prediction methods are mostly based on simplified models or single feature variables, making it difficult to adapt to complex real-world scenarios. This results in large prediction errors, leading to a mismatch between route planning and energy replenishment strategies, which in turn affects transportation efficiency and operational safety.

[0004] (2) Fast charging station clusters have strong scheduling rigidity and low green electricity utilization efficiency. Megawatt-level fast charging is a key means to support the high-frequency operation of heavy-duty trucks, but it has a significant impact on the grid load and the peak energy is difficult to reduce. Traditional fast charging stations operate independently and lack group control and source-load coordination strategies. They cannot effectively absorb the flexible resources of distributed photovoltaic, electricity price regulation and energy storage systems, resulting in high energy replenishment costs, system instability and low local consumption rate of green electricity, which restricts the development of green logistics.

[0005] (3) Manual plugging and unplugging is inefficient and cannot support unmanned operation. Currently, most electric heavy-duty trucks still require manual operation for charging, parking, gun docking, and plugging / unplugging. This is inefficient, costly, and poses safety hazards in unmanned vehicle operation scenarios. Especially in 24-hour high-frequency operation modes, manual recharging has become a major bottleneck restricting the large-scale deployment of unmanned heavy-duty trucks, necessitating the development of highly reliable, fully automated parking and charging technologies.

[0006] (4) The energy replenishment process lacks carbon emission control and green electricity certification methods. Despite the increasing integration of renewable energy into charging stations, the lack of transparency regarding the source of electricity during the charging process, the inability to quantify carbon reduction benefits, and the difficulty in establishing and trading green electricity rights severely restrict the development and use of carbon assets. Existing systems lack a reliable authentication mechanism covering the entire process from power source to vehicle to platform, making it difficult to support green electricity traceability, carbon label identification, and low-carbon operation assessment.

[0007] In summary, existing unmanned electric heavy trucks suffer from problems such as low accuracy in route energy consumption prediction, poor energy replenishment scheduling efficiency, incomplete automatic energy replenishment process, and unquantifiable carbon emission reduction benefits, which cannot support the large-scale, low-carbon, and efficient operation of unmanned heavy trucks in trunk logistics and smart highway scenarios. Summary of the Invention

[0008] To address the aforementioned issues, this disclosure proposes a method and system for route prediction and energy replenishment scheduling for unmanned electric heavy-duty trucks. It constructs a multi-source data-driven route energy consumption prediction model and a multi-source load scheduling model for station clusters, achieving joint optimization of heavy-duty truck transportation, charging, station capacity, and grid load prediction. Furthermore, it introduces a blockchain traceability mechanism to form a traceable and authenticated path for green energy replenishment and a quantitative model for carbon emission reduction benefits, supporting carbon credits, carbon accounts, and green electricity trading scenarios. In response to sudden load changes or equipment failures, it automatically adjusts power supply strategies to ensure the safe and stable operation of the system.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A method for path prediction and energy replenishment scheduling for unmanned electric heavy-duty trucks includes: Acquire multimodal data of electric heavy-duty truck operation and preprocess it; A multivariate coupled energy consumption prediction model is constructed. The preprocessed multimodal data is input into the multivariate coupled energy consumption prediction model, and the predicted path energy consumption value is output. By incorporating the predicted path energy consumption into the rapid energy replenishment process at the fast charging station, and employing reinforcement learning methods for source-grid-load-storage coordinated control, flexible scheduling results of the fast charging station cluster are obtained. A parking control model is constructed, and the flexible scheduling results of the fast charging station group are input into the parking control model. Based on the flexible scheduling results of the fast charging station group, a fusion SLAM algorithm based on LiDAR and vision is used for localization to generate a parking path. Precise charging is achieved through charging port recognition and attitude feedback. A vehicle-charging pile-network-cloud collaborative optimization scheduling model is constructed to obtain the actual charging power during the charging process. The actual charging power is input into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, an objective function is constructed, the optimal collaborative optimization scheduling result is solved, and carbon emission reduction is calculated in combination with green electricity utilization rate to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A route prediction and energy replenishment scheduling system for unmanned electric heavy-duty trucks includes: The data acquisition module is used to acquire and preprocess multimodal data of the electric heavy-duty truck's operation. The energy consumption prediction module is used to build a multivariate coupled energy consumption prediction model. It inputs the preprocessed multimodal data into the multivariate coupled energy consumption prediction model and outputs the path energy consumption prediction value. The station group scheduling module is used to introduce the predicted path energy consumption into the fast charging station's rapid energy replenishment process, and adopts reinforcement learning method to carry out source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group. The parking control module is used to build a parking control model. It inputs the flexible scheduling results of the fast charging station group into the parking control model. Based on the flexible scheduling results of the fast charging station group, it uses a fusion SLAM algorithm based on LiDAR and vision to locate the vehicle, generate a parking path, and perform precise charging through charging port recognition and attitude feedback. The collaborative optimization module is used to build a vehicle-charging pile-network-cloud collaborative optimization scheduling model, obtain the actual charging power during the charging process, input the actual charging power into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, construct the objective function, solve the optimal collaborative optimization scheduling result, and calculate carbon emission reduction by combining green electricity utilization rate, so as to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program that, when executed by a processor, implements the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0012] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0013] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0014] Compared with the prior art, the beneficial effects of this disclosure are as follows: This disclosure presents a route prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks. It integrates multimodal information such as road gradient, traffic flow, weather, load, SOC, and motor efficiency, and is based on a hybrid deep learning network of LSTM and Transformer to achieve high-precision dynamic energy consumption prediction of unmanned electric heavy-duty trucks under different road and load conditions, thereby improving the rationality of transportation route planning and the safety of driving range.

[0015] This disclosure presents a route prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks. Based on the distribution of fast charging stations, power capacity, operating time window, photovoltaic output, electricity price, and energy storage status, it constructs a multi-source load scheduling model for station clusters and develops a flexible optimization algorithm for the peak energy replenishment demand of heavy-duty trucks. This improves the coordination of fast charging, the local utilization rate of green electricity, and the response capability to peak and valley electricity prices, while reducing the operating cost of the energy replenishment system and the impact on the power grid.

[0016] This disclosure presents a path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks. It constructs a parking control model, integrates parking path planning and attitude recognition algorithms based on LiDAR and visual perception, and designs a charging nozzle system with self-interlocking and error compensation capabilities. This enables a closed-loop control process for unmanned electric heavy-duty trucks, encompassing "autonomous parking—automatic plugging / unplugging—intelligent charging—automatic departure," meeting the requirements for all-weather unmanned operation. This disclosed method for route prediction and energy replenishment scheduling of unmanned electric heavy-duty trucks constructs a data fusion and scheduling collaboration architecture centered on a cloud platform, achieving joint optimization of heavy-duty truck transportation, charging, station capacity, and grid load prediction. It introduces a blockchain traceability mechanism to form a traceable and authenticated path for green energy replenishment and a quantitative model for carbon emission reduction benefits, supporting carbon credits, carbon accounts, and green electricity trading scenarios, and promoting the development of green logistics. It automatically adjusts power supply strategies in response to sudden load changes or equipment failures, ensuring the safe and stable operation of the system. Attached Figure Description

[0017] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0018] Figure 1 This is an architecture diagram of a path prediction and energy replenishment scheduling method for unmanned electric heavy trucks according to an embodiment of this disclosure. Detailed Implementation

[0019] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0020] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0021] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0022] Example 1 One embodiment of this disclosure provides a path prediction and energy replenishment scheduling method for unmanned electric heavy trucks, the steps of which include: Step 1: Acquire and preprocess multimodal data of electric heavy-duty truck operation; Step 2: Construct a multivariate coupled energy consumption prediction model. Input the preprocessed multimodal data into the multivariate coupled energy consumption prediction model and output the predicted path energy consumption value. Step 3: Introduce the predicted path energy consumption into the fast charging station's rapid energy replenishment process, and use reinforcement learning to carry out source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group; Step 4: Construct a parking control model. Input the flexible scheduling results of the fast charging station group into the parking control model. Based on the flexible scheduling results of the fast charging station group, use the fusion SLAM algorithm based on LiDAR and vision to locate the vehicle and generate a parking path. Accurate charging is achieved through charging port recognition and attitude feedback. Step 5: Construct a vehicle-charging pile-network-cloud collaborative optimization scheduling model, obtain the actual charging power during the charging process, input the actual charging power into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, construct the objective function, solve for the optimal collaborative optimization scheduling result, and calculate the carbon emission reduction by combining the green electricity utilization rate, so as to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

[0023] As one embodiment, this disclosure presents a route prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks, which solves the problems of low accuracy in route energy consumption prediction, poor energy replenishment scheduling efficiency, incomplete automatic energy replenishment process, and unquantifiable carbon emission reduction benefits in the prior art. This method supports the large-scale, low-carbon, and efficient operation of unmanned heavy-duty trucks in trunk logistics and smart highway scenarios. The specific implementation process is as follows: Step 1: Acquire multimodal data of electric heavy-duty truck operation and preprocess it; Specifically, multimodal data of electric heavy truck operation is acquired, including road slope, vehicle speed, total vehicle load, state of charge of on-board power battery, ambient temperature, air humidity, motor speed and motor temperature.

[0024] Preprocessing operations include: normalizing the multimodal data and converting it into an input feature vector; Specifically, the input feature vector is: at time step t (Sampling interval is Δ) t At each second, the vehicle state and environmental parameters together constitute the input feature vector: x t =[ G t , V t , W t , T t , H t , SOC t , n t , T m ( t )] in, G t This represents the road slope; a positive value indicates an uphill slope, and a negative value indicates a downhill slope. V t (m / s) represents the vehicle's speed; W t (kg) represents the vehicle's current total mass (including its own weight and cargo weight). T t (°C) represents the ambient temperature; H t (%) represents air humidity, reflecting the impact of meteorological conditions on air resistance and air conditioning load; SOC t (%) represents the state of charge of the vehicle's power battery. n t (rpm) represents the motor speed; T m ( t (°C) represents the motor temperature.

[0025] Furthermore, a model of motor efficiency and instantaneous power is constructed, namely motor efficiency. or motor With rotational speed n t and temperature T m ( t The relationship can be represented by a two-dimensional fitting function:

[0026] in,P out ( t )( kW ) represents the mechanical power output of the motor; P in ( t )( kW ) represents the input electrical power of the motor.

[0027] Furthermore, instantaneous mechanical power is calculated from traction force and speed:

[0028] Total traction force:

[0029] Furthermore, rolling resistance:

[0030] in, m t = W t / g For the total mass of the vehicle. g =9.81m / s 2 Represents gravitational acceleration; C r This represents the rolling resistance coefficient.

[0031] Furthermore, the air resistance is:

[0032] in, r air ( T t , H t () represents air density. C d This is the drag coefficient. A f ( m ²) represents the vehicle's frontal area.

[0033] Furthermore, the slope resistance is:

[0034] Furthermore, the acceleration resistance is:

[0035] in, a t =( V t Vt 1) / Δ t To accelerate the vehicle.

[0036] Furthermore, the input electrical power of the motor:

[0037] Instantaneous energy consumption and total energy consumption prediction:

[0038] in, E t ( kWh (Time step) t Energy consumption.

[0039] Step 2: Construct a multivariate coupled energy consumption prediction model. Input the preprocessed multimodal data into the multivariate coupled energy consumption prediction model and output the predicted path energy consumption value. Specifically, the multivariate coupled energy consumption prediction model is a hybrid deep learning model based on LSTM and Transformer. To account for multi-factor coupling and uncertainties in actual operation, a hybrid deep learning model is adopted. f θ right x t k:t Energy consumption prediction for sequences:

[0040] in, X t This is a sequence of multimodal input vectors from the previous several steps. f θ It is an LSTM-Transformer hybrid deep learning model with the following parameters: i , k This is the length of the history window.

[0041] The predicted total energy consumption is:

[0042] The multivariate coupled energy consumption prediction model disclosed herein is connected to the interface of the subsequent sub-module system. The following data output by the multivariate coupled energy consumption prediction model disclosed herein is used by the next module for flexible scheduling of fast charging station groups, including energy consumption curves, total energy consumption prediction values ​​and charging demand values.

[0043] Furthermore, the charging demand value is:

[0044] in, Einit (kWh) represents the remaining battery power of the vehicle before the mission begins.

[0045] Furthermore, the energy consumption curve is: { } t =1 T .

[0046] Step 3: Introduce the path energy consumption prediction results into the fast charging station's rapid energy replenishment process, and use reinforcement learning methods to carry out source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group; Specifically, the modules of this system are used to achieve overall load balancing and maximize green energy consumption of the station group through source-grid-load-storage coordinated control during the rapid energy replenishment process at the station end.

[0047] The energy consumption curve, total energy consumption prediction value, and charging demand value output by the multivariate coupled energy consumption prediction model are introduced into the fast charging station's rapid energy replenishment process. Through source-grid-load-storage coordinated control, with the goal of maximizing the load leveling and green energy consumption of the entire station group, an optimization objective function for the station group is constructed. The constraints include power balance, energy storage capacity limit, and station capacity limit. Reinforcement learning or mixed integer linear programming is used to achieve real-time optimization and solve the problem, obtaining the flexible scheduling result of the fast charging station group. The flexible scheduling result of the fast charging station group is used to generate a charging reservation task order, which is then passed to the next controller, automatic parking control.

[0048] First, construct the objective function for optimizing the website cluster:

[0049] in, P i c ( t ) is the first i Each charging station time t The charging power; P i pv ( t ) is the first i Solar power output at the site; P i grid ( t () represents the power output of the station from the power grid; P i es ( t () represents the charging / discharging power of the energy storage device; positive indicates discharging, and negative indicates charging. C i ( t This refers to the site's load capacity limit; π ( t ) is the electricity price function.

[0050] Furthermore, the constraints are as follows: (1) Power balance:

[0051] (2) Energy storage capacity limitations:

[0052] (3) Station capacity limitations:

[0053] The scheduling optimization function disclosed herein is solved using reinforcement learning or mixed integer linear programming (MILP) to obtain flexible scheduling results for fast charging station clusters, enabling real-time optimization. Based on the flexible scheduling results of the fast charging station clusters, a charging reservation task order is generated and passed to the next automatic parking subsystem. The specific solution process is as follows: (1) Linearization and auxiliary variables The objective quadratic term can be treated as a quadratic optimization (QP) or a linearized approximation: Bundle( P i c P i pv P i es ) 2 Replace with auxiliary variables z i,t And approximate it using tangent / piecewise linearization, or directly use the L1 form. Linearization substitution (introducing positive and negative auxiliary variables).

[0054] Introduce the peak term of max(·) into an auxiliary variable. q t :

[0055] Write the peak penalty term as k ∑ t q t .

[0056] Energy storage cannot be charged and discharged simultaneously: Introducing a binary variable b i,t :

[0057] (2) Rolling time domain solution Each solution HFor each MILP time interval (H≤T, e.g., H=12), the action of the first time interval is executed, the time is slid forward and the solution is repeated.

[0058] The optimizer recommends using Gurobi / CPLEX, setting a time limit (e.g., 30s-120s) to ensure real-time performance, and using warm-start (from the previous solution) to accelerate the process.

[0059] (3) Output and fault tolerance If MILP fails to converge within the time limit, return the current feasible solution (heuristic / truncation solution) and trigger the next round.

[0060] Anomalies (connection failure / vehicle delay) trigger local re-optimization (only for the affected sites / vehicles) to reduce the solution size and response time.

[0061] Step 4: Construct a parking control model. Input the flexible scheduling results of the fast charging station group into the parking control model. Based on the flexible scheduling results of the fast charging station group, use the fusion SLAM algorithm based on LiDAR and vision to locate the vehicle and generate a parking path. Accurate charging is achieved through charging port recognition and attitude feedback. Specifically, the parking control model is a full-process control model of "path planning - precise parking - automatic plugging and unplugging - intelligent recharging - automatic departure", and the specific control process of the parking control model is as follows: (1) Define the parking state as s =[ x , y , i ] indicates the vehicle's position and heading angle.

[0062] (2) Localization is performed using a fusion SLAM algorithm based on LiDAR and vision, and a path is generated using Bézier curves:

[0063] in, c ={ c 1, c 2,…, c n} represents the candidate trajectory points for vehicle planning, each c i Two-dimensional coordinates ( x i , y i ); c i ref For reference, the coordinates of the target point on the parking path; || c i ci ref ||2: The squared Euclidean distance between the trajectory point and the reference point, reflecting the path deviation; β ObstacleCost is a dimensionless obstacle cost weighting coefficient used to adjust the importance of obstacle avoidance constraints. c ): Distance cost function between the path and obstacles (the closer the distance, the greater the cost); n is the number of discrete points on the path.

[0064] Then, the execution end uses a closed-loop control algorithm:

[0065] in, d For steering control, x des The desired vehicle position (lateral target position); x The vehicle's current location; i des The desired vehicle steering angle; i The vehicle's current steering angle; k 1, k 2 is the control gain parameter, used to adjust the control ratio of position error and angle error.

[0066] After obtaining the steering control input, the parking charging process is controlled. A charging gun control structure with six degrees of freedom servo adjustment is designed. Combined with port position recognition and attitude feedback, millimeter-level insertion and removal accuracy is achieved for precise charging. Redundant safety protection devices are used to improve stability. The specific control process is as follows: First, determine the input, including: P dock =[ x d , y d , z d , d , i d , ψ d ], which is the target pose of the parking / charging station in the charging gun reference frame ( m , rad ).

[0067] And to acquire real-time sensor data, including: visual camera images, LiDAR point clouds, force / torque sensor readings, F meas =[ F x , F y ,F z , M x , M y , M z Gun end encoder / joint angle q .

[0068] And station / vehicle communication commands: plug-in permission flag, maximum permissible contact force threshold. F max Allowable insertion depth d max .

[0069] Then, (1) perform initial static calibration and safety checks (initialization): 1) Verify system health: Camera / LiDAR / force sensor / encoder online self-test passed.

[0070] 2) Obtain the reference positions of the parking space and the pile entrance. P dock (Source: Cloud Dispatch / Parking Location).

[0071] 3) Station confirmation of "Connection Allowed": If the station or vehicle does not allow it, exit and report.

[0072] 4) Set control parameters and thresholds. K p pos (Position ratio) K d pos (Damping gain, used for impedance control position term). K p ori (Posture proportions) K d ori (Damping gain, used for attitude impedance). F thres (Maximum permissible contact force), M thres (Permissible torque threshold). v max (Maximum safe execution speed) pos (Location success determination threshold) ori （ Angle success threshold ( d ins (Current insertion depth) d max (Maximum allowed insertion depth).

[0073] (2) Coarse alignment (Cartesian / joint space low-speed movement): Objective: To move the gun tip to the approach position at a safe speed. P dock pre-approach point P pre The pre-approach distance is generally d pre =0.05~0.2m.

[0074] The control employs a speed / position closed-loop (low-gain PID) system, limiting the speed to... v max,coarse The formula is:

[0075] in, , e p This represents the error vector. If || is detected during the pre-approach process... F meas ||> F thres If a sudden malfunction occurs, stop immediately and initiate abnormal handling.

[0076] (3) Visual-LiDAR fusion fine registration (output error) Δx ) High-precision relative pose with the pile entrance is obtained by combining camera marker detection / depth map with LiDAR local point cloud registration (ICP or NDT):

[0077] If the visual / point cloud confidence level is low (below the threshold), increase the number of observations or switch to force-guided mode.

[0078] (4) Impedance / force control to execute insertion (core control law) Smooth insertion is achieved using six-degree-of-freedom impedance control (Cartesian impedance control), and the control law can be expressed as (illustrated):

[0079] Desirable force / torque [F] cmd M cmd] Convert to joint force / position commands.

[0080] Furthermore, an impedance control principle is adopted: when encountering external force (such as a slight touch), the system absorbs the deviation with a set compliance and continues to make minor adjustments to avoid hard collisions.

[0081] Insertion velocity planning: Based on the current residual pose error and contact force, a gradual deceleration strategy is adopted.

[0082] Force / torque safety stop determination: If at any moment any component satisfies | F i |> F thres or | M i |> M thres Immediately switch to a safe stop and return to the previous safe pose, and record this as a failed retry.

[0083] (5) Successful connection determination A connection is successful if and only if the following conditions are met simultaneously:

[0084] Once the conditions are met, the electrical handshake protocol is triggered (CAN / PLC handshake with the station / vehicle BMS). If the handshake is successful, the automatic plug-in mechanism executes charging and collects data such as the actual charging power to transmit to the next module vehicle-pile-network-cloud submodule. After the parking and charging process is completed, the automatic plug-in mechanism executes charging and collects data such as the actual charging power to transmit to the next module vehicle-pile-network-cloud submodule.

[0085] Step 5: Construct a vehicle-charging pile-network-cloud collaborative optimization scheduling model, obtain the actual charging power during the charging process, input the actual charging power into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, construct the objective function, solve for the optimal collaborative optimization scheduling result, and calculate the carbon emission reduction in combination with the green electricity utilization rate to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks; Specifically, this sub-module of the present disclosure realizes the coordinated operation optimization of unmanned electric heavy trucks (vehicles), intelligent charging piles (piles), charging stations and power distribution systems (networks) and dispatch cloud systems (clouds) through a cloud platform, and calculates carbon emission reductions in conjunction with green electricity utilization rates.

[0086] First, a collaborative optimization scheduling model is constructed, with the objective function being the weighted minimization of the overall transportation task delay cost, charging station load fluctuation cost, and energy consumption economic cost:

[0087] in, N v This refers to the number of unmanned heavy trucks; N s The number of charging stations; C delay,i For the first i Costs of transportation delays caused by vehicle dispatching or charging; C load,j For the firstj The cost of load fluctuation penalties for each charging station; C energy,i For the first i The cost of electricity for vehicle refueling; α 1, α 2, α 3 is the weighting coefficient, which is set according to operational priority.

[0088] Furthermore, the constraints include: (1) Vehicle energy constraints

[0089] Ensure that the vehicle's State of Charge (SOC) is not lower than the safe value after the mission is completed.

[0090] (2) Charging station capacity constraints

[0091] Every moment t The total power of the charging station shall not exceed its maximum capacity.

[0092] (3) Task time window constraints

[0093] Ensure that the transportation task is completed on time.

[0094] Furthermore, the utilization rate of green electricity is quantified in relation to carbon emission reduction, including: Calculation of actual green electricity utilization rate:

[0095] in, E j green For the first j The green electricity consumed by each charging station (from photovoltaic, wind power, etc.), E j total For the first j Total charging capacity of each charging station.

[0096] Formula for calculating carbon emission reduction:

[0097] in, E total Total charging capacity for all charging stations; EF grid Carbon emission factor per unit of electricity generated by the power grid (kgCO2 / kWh).

[0098] Carbon emission reductions can be treated as quantifiable operational results, and companies can apply for third-party carbon verification and certification to trade in the voluntary carbon market (sale of carbon credits). This can also serve as direct data support for companies to fulfill their targets, reducing the cost of purchasing carbon allowances.

[0099] Example 2 One embodiment of this disclosure provides a path prediction and energy replenishment scheduling system for unmanned electric heavy trucks, including: The data acquisition module is used to acquire and preprocess multimodal data of the electric heavy-duty truck's operation. The energy consumption prediction module is used to build a multivariate coupled energy consumption prediction model. It inputs the preprocessed multimodal data into the multivariate coupled energy consumption prediction model and outputs the path energy consumption prediction value. The station group scheduling module is used to introduce the predicted path energy consumption into the fast charging station's rapid energy replenishment process, and adopts reinforcement learning method to carry out source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group. The parking control module is used to build a parking control model. It inputs the flexible scheduling results of the fast charging station group into the parking control model. Based on the flexible scheduling results of the fast charging station group, it uses a fusion SLAM algorithm based on LiDAR and vision to locate the vehicle, generate a parking path, and perform precise charging through charging port recognition and attitude feedback. The collaborative optimization module is used to build a vehicle-charging pile-network-cloud collaborative optimization scheduling model, obtain the actual charging power during the charging process, input the actual charging power into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, construct the objective function, solve the optimal collaborative optimization scheduling result, and calculate carbon emission reduction by combining green electricity utilization rate, so as to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

[0100] Example 3 One embodiment of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0101] Example 4 One embodiment of this disclosure provides a non-transitory computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, they implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0102] Example 5 One embodiment of this disclosure provides an electronic device, including a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks.

[0103] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure one One or more processes and / or boxes Figure one A device that provides the functions specified in one or more boxes.

[0104] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure one One or more processes and / or boxes Figure one The steps of the function specified in one or more boxes.

[0105] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.

Claims

1. A method for path prediction and energy replenishment scheduling for unmanned electric heavy-duty trucks, characterized in that, include: Acquire multimodal data of electric heavy-duty truck operation and preprocess it; A multivariate coupled energy consumption prediction model is constructed. The preprocessed multimodal data is input into the multivariate coupled energy consumption prediction model, and the path energy consumption prediction results are output. The path energy consumption prediction results are introduced into the fast charging station's rapid energy replenishment process, and reinforcement learning methods are used for source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group. A parking control model is constructed, and the flexible scheduling results of the fast charging station group are input into the parking control model. Based on the flexible scheduling results of the fast charging station group, a fusion SLAM algorithm based on LiDAR and vision is used for localization to generate a parking path. Precise charging is achieved through charging port recognition and attitude feedback. A vehicle-charging pile-network-cloud collaborative optimization scheduling model is constructed to obtain the actual charging power during the charging process. The actual charging power is input into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, an objective function is constructed, the optimal collaborative optimization scheduling result is solved, and carbon emission reduction is calculated in combination with green electricity utilization rate to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

2. The path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks as described in claim 1, characterized in that, Acquire multimodal data of unmanned electric heavy truck operation, including road slope, vehicle speed, total vehicle load, on-board power battery state of charge, ambient temperature, air humidity, motor speed and motor temperature.

3. The path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks as described in claim 1, characterized in that, The step of inputting the preprocessed multimodal data into a multivariate coupled energy consumption prediction model and outputting path energy consumption prediction values ​​includes: constructing an input feature vector from the preprocessed multimodal data, inputting the feature vector into the multivariate coupled energy consumption prediction model, and using the path energy consumption prediction results at multiple future time steps using a hybrid deep learning model. in, X t This is a sequence of multimodal input vectors from the previous several steps. f θ It is an LSTM-Transformer hybrid deep learning model with the following parameters: θ .

4. The path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks as described in claim 1, characterized in that, The process of introducing the path energy consumption prediction results into the fast charging station's rapid energy replenishment process and using reinforcement learning for source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group includes: introducing the energy consumption curve, total energy consumption prediction value, and charging demand value output by the multivariate coupled energy consumption prediction model into the fast charging station's rapid energy replenishment process; achieving overall load balancing and maximizing green energy consumption of the station group through source-grid-load-storage coordinated control; firstly, constructing the station group optimization objective function with constraints including power balance, energy storage capacity limitations, and station capacity limitations; and using reinforcement learning or mixed integer linear programming to achieve real-time optimization and solution to obtain the flexible scheduling results of the fast charging station group; and using the flexible scheduling results of the fast charging station group to generate a charging reservation task order, which is then passed to the next controller, automatic parking control.

5. The path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks as described in claim 1, characterized in that, The process involves using a fusion SLAM algorithm based on LiDAR and vision for localization, generating a parking path, and performing precise charging through charging port recognition and attitude feedback. This includes: defining the parking state, representing the vehicle's position and orientation angle; using a fusion SLAM algorithm based on LiDAR and vision for localization; generating a parking path using Bézier curves; using a closed-loop control algorithm to obtain steering control inputs; designing a charging gun control structure with six degrees of freedom servo adjustment; and combining port position recognition and attitude feedback to achieve millimeter-level automatic plug-in / plug-out charging.

6. The path prediction and energy replenishment scheduling method for unmanned electric heavy-duty trucks as described in claim 1, characterized in that, The process involves constructing a vehicle-charging pile-network-cloud collaborative optimization scheduling model to obtain the actual charging power during the charging process. This actual charging power is then input into the model to construct an objective function and solve for the optimal collaborative optimization scheduling result. This includes optimizing the collaborative operation of unmanned electric heavy trucks, intelligent charging piles, charging stations, and the power distribution network with the scheduling cloud system through a cloud platform. The objective function is the weighted minimum of the comprehensive transportation task delay cost, charging station load fluctuation cost, and energy consumption economic cost. Constraints include vehicle energy constraints, charging station capacity constraints, and task time window constraints. The actual green electricity utilization rate and carbon emission reduction are calculated, with carbon emission reduction used as a quantifiable operational result to achieve path prediction and energy replenishment scheduling for unmanned electric heavy trucks.

7. A route prediction and energy replenishment scheduling system for unmanned electric heavy-duty trucks, characterized in that it includes: The data acquisition module is used to acquire and preprocess multimodal data of the electric heavy-duty truck's operation. The energy consumption prediction module is used to build a multivariate coupled energy consumption prediction model. It inputs the preprocessed multimodal data into the multivariate coupled energy consumption prediction model and outputs the path energy consumption prediction value. The station group scheduling module is used to introduce the predicted path energy consumption into the fast charging station's rapid energy replenishment process, and adopts reinforcement learning method to carry out source-grid-load-storage coordinated control to obtain the flexible scheduling results of the fast charging station group. The parking control module is used to build a parking control model. It inputs the flexible scheduling results of the fast charging station group into the parking control model. Based on the flexible scheduling results of the fast charging station group, it uses a fusion SLAM algorithm based on LiDAR and vision to locate the vehicle, generate a parking path, and perform precise charging through charging port recognition and attitude feedback. The collaborative optimization module is used to build a vehicle-charging pile-network-cloud collaborative optimization scheduling model, obtain the actual charging power during the charging process, input the actual charging power into the vehicle-charging pile-network-cloud collaborative optimization scheduling model, construct the objective function, solve the optimal collaborative optimization scheduling result, and calculate carbon emission reduction by combining green electricity utilization rate, so as to realize the path prediction and energy replenishment scheduling process of unmanned electric heavy trucks.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks as described in any one of claims 1-6.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks as described in any one of claims 1-6.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the path prediction and energy replenishment scheduling method for unmanned electric heavy trucks as described in any one of claims 1-6.

Images