34results about How to "Achieving joint optimization" patented technology

The invention provides a drug recommendation method fusing medical record features and risk control, and belongs to the technical field of medical information. Existing drug recommendation depends on external medical knowledge to a great extent, so that a recommendation system is poor in expandability and insufficient in clinical practicability. On the basis, the invention provides a drug recommendation method which does not depend on external medical knowledge and is only based on electronic medical record data. According to the method, an individualized illness state recognition and evolution modeling module is designed, records most representative for the current illness state can be automatically recognized from historical medical records of a patient, the disease course development process is described, and individualized medication preferences and treatment modes are mined. According to the method, semantic features of data are fully utilized, so that joint optimization of recommendation risks and effects is realized under the condition of no external knowledge intervention. Experimental results show that the method is obviously superior to existing mainstream methods in multiple evaluation indexes, and has good accuracy, robustness and clinical application prospects.

This invention discloses a method for predicting energy consumption in the production of prefabricated building components, comprising the following steps: collecting and processing historical time-series data of prefabricated component production, establishing a time-series input matrix with multi-period feature enhancement; constructing a deep time-series quantile attention network, extracting local time-series features and global periodic dependencies in parallel through time-series quantile attention blocks to form a fused time-series representation; inputting the fused time-series representation into a progressive quantile decoder based on a quantile regression mechanism, outputting a component production energy consumption prediction interval with multiple confidence levels; adaptively adjusting the network configuration and training strategy according to data frequency and sample size to achieve model optimization and deployment; acquiring the target prefabricated component production time-series data to be predicted, and outputting the component production energy consumption prediction interval for the corresponding target period. This invention achieves joint optimization of prediction intervals and point predictions, and can output reliable intervals that meet the target confidence level, providing a quantitative basis for production scheduling risk decision-making.

The application discloses an adaptive tuning device and method of an anti-interference satellite communication antenna, and relates to the technical field of satellite communication, wherein the device comprises a multi-mode reconfigurable surface antenna module, an environment sensing module, a prediction module, a cooperative tuning module and a closed-loop feedback correction module; the application constructs a multi-mode radiation system through the multi-mode reconfigurable surface antenna module and the environment sensing module, obtains an electromagnetic environment and a link state, the prediction module predicts interference and signal deviation based on multi-source data, the cooperative tuning module synchronously executes radiation mode switching and array amplitude and phase pre-tuning according to a prediction result, preset nulls are formed before interference arrives, and the closed-loop feedback correction module collects signal quality in real time and corrects a prediction model and a tuning strategy, so that a prediction type, closed-loop self-optimizing active anti-interference tuning mechanism is formed, interference is not arrived, and tuning is performed in advance, and the anti-interference capability and communication stability under a dynamic complex electromagnetic environment are significantly improved.

The application is suitable for the technical field of power market dispatching, and provides a multi-market coordinated dispatching optimization method and system for wind-solar-water storage integration, which comprises the following steps: obtaining wind-solar-water storage system operation parameters, adjustable industrial load response characteristic parameters and power market rule parameters, constructing an adjustable load aggregation model, and establishing a multi-market regulation capacity mapping relationship for electricity energy, frequency modulation and standby auxiliary service markets; based on the mapping relationship and the operation parameters, a multi-market day-ahead joint coordinated dispatching model is constructed, with the optimal system operation comprehensive index as the target, and the base load output of various resources, the frequency modulation and standby capacity reservation, the energy storage charging and discharging and the load power regulation plan as the decision variables, and the coordinated dispatching scheme is obtained by solving the model and executed. The application effectively improves the overall operation economy, regulation flexibility and coordinated dispatching efficiency of the high-proportion renewable energy power system in the multi-market environment.

The invention discloses a bogie fault identification method based on an agent model and significance sorting, and the method comprises the steps: carrying out the preprocessing of the collected original state information in the operation process of a bogie, and obtaining the processed information; performing factor calibration on the processed information to obtain a fault factor vector; establishing a mapping relation between the fault factor and an actual physical fault state; the fault factor vector is used as an input variable, the fault severity quantitative index is used as an output variable, and a fault function meeting monotonicity or segmented monotonicity constraint is constructed; and constructing an improved Kriging model, fitting a nonlinear relationship between the fault factors and the fault function, and quantifying the proportion of the main effect and the interaction effect of each fault factor. According to the method, multi-factor significance sorting and high-precision prediction capability can be considered, and technical support is provided for intelligent fault diagnosis of the railway vehicle bogie.

The invention provides an oil-water separation effect prediction method and device based on a Lonion model. The method comprises the following steps: defining an input independent variable vector f (x) = [x1, x2,..., x8] T; performing feature engineering extension on the input independent variable vector to generate an extended feature vector f (z); the extended feature vector f (z) comprises an original feature vector, a polynomial feature vector and a physical interaction feature vector; inputting the extended feature vector f (z) into a pre-trained prediction model to obtain a predicted value, output by the pre-trained prediction model, of the oil content of the effluent of the air flotation synergistic device; inputting the predicted value into an integrated prediction framework to obtain a final predicted value and uncertainty estimation; displaying the final predicted value and the uncertainty estimation to a user through a display interface; according to the technical scheme, the accuracy, robustness and physical consistency of oil content prediction of the effluent can be improved.

This invention relates to the field of autonomous navigation technology for indoor mobile robots, specifically disclosing a SLAM method for indoor mobile robots, comprising: acquiring environmental image information of the indoor mobile robot; performing ray sampling and attention-based encoding fusion on an RGB-D image sequence to obtain first optimized parameter information; performing pose optimization processing on the initial pose information according to a target total loss function to obtain tracked frames with pose information; determining keyframes for mapping based on the tracked frames; performing ray sampling and attention-based encoding fusion on the keyframes to obtain second optimized parameter information; and performing joint pose and mapping optimization processing on the second optimized parameter information according to the target total loss function to obtain pose and map construction results. The SLAM method for indoor mobile robots provided by this invention can improve positioning accuracy in complex indoor environments.

The application provides a pig behavior recognition method and device based on skeleton topology constraint and an electronic device. The method comprises the following steps: S1, obtaining a pig image to be recognized; S2, inputting the pig image into a trained pig behavior recognition model to obtain a pig behavior recognition result; the pig behavior recognition model comprises a backbone network, a neck network and a detection head; the backbone network is used for multi-scale feature extraction on the input pig image; the neck network is used for feature fusion and enhancement on the multi-scale features extracted by the backbone network to generate a multi-scale enhanced feature map; the detection head comprises a key point prediction branch, a visual classification detection branch and a behavior classifier. The application deeply fuses the skeleton space topology features between key points and visual priori, so that the behavior discrimination can explicitly utilize the geometric dependency relationship between key points, and effectively solves the problem of accurate pig behavior monitoring in a complex pigsty environment.

The invention discloses a nerve-mechanics coupling software multi-mode motion control method based on a pre-training language model. Comprising a mass point-spring-based soft robot body structure module, a rhythm oscillation signal generation and parameter control module modulated based on a pre-training language model, a muscle driving and mechanical calculation module based on nerve-mechanical mapping, and a simulation and debugging module based on physical simulation. The method is oriented to gait planning and intelligent control of the bionic soft robot in an unstructured environment. In order to solve the problem that gait switching depends on artificial parameters and is difficult to quickly adjust along with instructions, a pre-training language model is introduced, and high-level semantic instructions are mapped into neural controller characteristic parameters. The robot can smoothly switch wriggling, steering and rolling behaviors under the working conditions of a hard plane, viscous fluid and the like through a natural language without remodeling, and a control and analysis method for effectively improving the autonomous environment adaptability of a software system is provided.

The invention relates to the technical field of network resource scheduling and artificial intelligence, and provides a computing power routing method based on multi-intelligence collaboration, which comprises a satellite, an unmanned aerial vehicle, a ground base station, a user terminal and other multi-layer heterogeneous nodes. Network nodes are dynamically networked through a wireless link, and node movement, topology change and multi-hop communication are supported. Each node has certain computing power and storage capability, and can cooperatively complete data processing and task distribution. The core routing nodes are deployed in a network backbone in a distributed manner and are responsible for intelligent routing and resource scheduling. The system integrates key technologies such as link quality quantification, distributed election, reinforcement learning of intelligent agents, virtual wireless interface simulation, flow control and route issuing. The method provided by the invention supports large-scale simulation, dynamic routing optimization and computing power resource collaboration, is suitable for scenes such as satellite internet, unmanned aerial vehicle clusters, emergency communication and intelligent transportation, and realizes wide-area, intelligent and high-reliability communication and computing power services.

The invention discloses a three-dimensional human body posture estimation method and device and a medium, and belongs to the technical field of computer vision. The method comprises the following steps: extracting two-dimensional human body posture key points based on an acquired video picture to obtain a two-dimensional human body posture key point sequence; projecting the two-dimensional key point sequence to a feature space through nonlinear high-dimensional mapping to obtain a high-dimensional feature space matrix; inputting a three-dimensional human body posture estimation model based on the high-dimensional feature matrix to obtain a three-dimensional human body posture key point sequence; based on the three-dimensional human body posture key point sequence, a three-dimensional human body posture estimation result is obtained through the three-dimensional coordinate point positions. According to the method, through the three-dimensional human body posture estimation model, the anti-interference capability of feature extraction is enhanced, and the robustness of three-dimensional posture estimation in a complex dynamic scene is remarkably improved.

This invention discloses a method for predicting the unsteady acoustic quality of electric drives based on intelligent sensors. Specifically, it includes: acquiring noise signals and extracting sound pressure level, loudness, sharpness, speech intelligibility, and priority speech interference level to form a feature vector sequence and dataset; constructing a neural network to determine nodes and form network parameter vectors; establishing an information entropy budget allocation model, mapping each feature to an information budget variable vector, setting constraints to construct an objective function containing budget constraint terms and prediction error terms, and embedding the budget variables into the feature representation; constructing a particle swarm, uniformly encoding the information budget vector and network parameter vector into particle state vectors and then initializing them; introducing an information budget modulation factor during iteration to weight the velocity update component, achieving co-evolution of the two; updating the optimal particle based on fitness, outputting the optimal information budget and network parameters to construct a prediction model; and inputting the feature vector sequence to output the prediction result. This invention assigns independent budgets to each acoustic feature, dynamically co-optimizes, and achieves high accuracy and fast response.

The invention discloses a semantic communication optimization method and system based on sparse reward re-optimization, and belongs to the technical field of wireless communication, and the method comprises the steps: constructing a semantic communication optimization model; establishing a reward prediction model, and predicting a reward value at a future moment in reinforcement learning; reconstructing a reward sequence based on a reward value prediction result; and performing reinforcement learning on the semantic communication channel based on the reconstructed reward sequence to realize transmission strategy updating. According to the method, the problems of sparse rewards and unstable training in semantic communication reinforcement learning can be effectively relieved under the conditions of complex channels and limited feedback, joint optimization of a semantic layer and a channel layer is realized, and an intelligent communication system is promoted to develop towards an efficient, self-adaptive and interpretable semantic transmission direction while low delay and high reliability are ensured.

This invention discloses a method for collaborative search and capture of unmanned aerial vehicle (UAV) swarms in a denied environment, belonging to the field of UAV swarm technology. The method includes the following steps: The method comprises two phases: a search phase and a capture phase. In the search phase, the UAV swarm, based on jointly solved stimulus values, covers the target area as efficiently as possible, exploring both dynamic and static targets. In the capture phase, the alpha UAV leads the arrived UAVs to attack the target until it is eliminated, completing the mission. This invention introduces a UAV collaborative localization method based on positional confidence, which is more realistic, allows for more accurate target localization, and facilitates decision-making applicable to real-world scenarios.

The application discloses a kind of financial special network optimization method and system for AI big model, it is related to communication network technical field.The method is by collecting AI business demand and network state, using the SID programming ability of SRv6, generates exclusive path strategy for AI training, inference and sensitive data flow;Real-time monitoring of computing power node load, link computing power state and network scheduling, realize the collaborative optimization of "network-computing power";Introduce the hop-by-hop path authentication mechanism based on dynamic authentication segment identifier Authen-SID, enhance the security of data transmission path;End-to-end visualization operation and maintenance and closed-loop feedback of service flow are realized by using stream detection technology.The application solves the problems of traffic scheduling mismatch, network and computing power collaboration deficiency, path security guarantee deficiency and weak operation and maintenance visualization capability in the prior art, realizes the deep adaptation of network resources and AI business demand, improves the computing power collaborative efficiency, path security and operation and maintenance intelligent level.

The invention discloses a large-scale MIMO (Multiple-Input Multiple-Output) communication-sensing integrated joint precoding and power distribution method and system, relates to the field of wireless communication, and aims to solve the problem that the existing method cannot give consideration to both improvement of communication rate and high-precision estimation of a target. The method is technically characterized by comprising the following steps of: 1, establishing a large-scale MIMO (Multiple Input Multiple Output) general sensing integrated base station downlink joint beam forming pre-coding system model; step 2, establishing minimization interference of joint power distribution and a Cramer-Rao bound ISAC downlink precoding optimization model; 3, initializing quantum positions of particles, and constructing a fitness function according to the established optimization model; 4, selecting an optimal quantum position; 5, judging the decay type of the particles, and generating new particles; 6, mapping the updated quantum positions of the particles into positions, calculating fitness function values of all the updated particles, and updating the global optimal quantum position; and step 7, performing iterative optimization to obtain an optimal ISAC joint precoding and power distribution scheme.

The invention provides a dual unmanned aerial vehicle (UAV) trajectory optimization method and system applied to a multi-eavesdropper and no-fly zone coexistence scene. The method comprises the following steps: constructing a communication system model; constructing an optimized mathematical model with user scheduling, unmanned aerial vehicle beamforming and unmanned aerial vehicle flight path as constraints by taking maximization of the user safety rate as a target according to the communication system model; a block coordinate descent method is adopted to decouple the optimized mathematical model into sub-problems only about user scheduling, unmanned aerial vehicle beam forming and flight tracks of the unmanned aerial vehicle U and the unmanned aerial vehicle J; for each non-convex sub-problem, a convex approximation method is adopted to convert the non-convex sub-problem into a convex optimization form; an iterative algorithm is adopted to sequentially solve user scheduling, unmanned aerial vehicle wave beam occurrence and convex subproblems of flight paths of the user scheduling and the unmanned aerial vehicle wave beam occurrence, and finally successive approximation is carried out and a global optimal solution of an original overall optimization problem is obtained; according to the method, the waveform and the track are jointly optimized, and the eavesdropping capability of an eavesdropper is effectively interfered by using a radar signal, so that safe communication between the unmanned aerial vehicle and a ground user is ensured, and the method has obvious innovativeness and uniqueness in system design and application scenes.

The application provides a time-frequency-space domain resource coordination optimization method in a multi-network fusion communication scene, relates to the field of space-air-ground integration, and comprises the following steps: a multi-network fusion communication model is constructed, and a multi-network fusion communication process is divided into an access stage and a backhaul stage; in the access stage, a channel model of a ground terminal to a ground access point and a channel model of the ground terminal to a relay unmanned aerial vehicle are respectively established, and in the backhaul stage, channel models between a relay unmanned aerial vehicle and a satellite and between a ground access point and a cloud server are respectively established; an access stage optimization problem and a backhaul stage optimization problem are respectively constructed; a feasible trajectory of the unmanned aerial vehicle is generated based on K-means clustering, and a linear programming sub-problem under a fixed trajectory that can be solved is obtained; and the backhaul stage optimization problem is split into a backhaul sub-problem of the unmanned aerial vehicle to the ground access point and a short-wave backhaul sub-problem. The application globally improves system throughput, resource utilization and task reliability under the multiple constraints of unmanned aerial vehicle movement, terminal rate, minimum traffic volume and the like.

The invention relates to an image generation method and device, and the method comprises the steps: carrying out the training of a VAE (variational auto-encoder) model, and enabling the VAE model to be used for encoding an input image into a potential vector in a potential space, and decoding and reconstructing the potential vector into an image; in the training process of the VAE model, constructing and maintaining a group of learnable cluster centers; in each training iteration, according to the distance from the current potential vector to each cluster center, distributing the potential vector to the corresponding cluster, and updating the position of the cluster center based on the obtained distribution result; a total loss function used for updating VAE model parameters is calculated, and after training is completed, a VAE model with a structured potential space is obtained; and sampling potential vectors near the cluster center corresponding to the target category, and decoding by using a decoder of the VAE model to generate an image belonging to the target category. Therefore, the problem that the category of the image generated by the traditional VAE is uncontrollable can be fundamentally solved.

The invention discloses a joint optimization method for DNN partition and resource allocation, which realizes collaborative reasoning and resource optimization of a deep neural network in an environment of a single edge server and a plurality of mobile devices, and comprises the following steps of: determining DNN hierarchy segmentation points of each device and the number of computing resource blocks and wireless resource blocks purchased from the edge server; decoupling modeling is performed on a DNN structure and an edge environment through a heterogeneous graph attention network, an optimal partition point and a resource deployment strategy are determined through Actor-Critic under the constraint of an M / M / 1 queuing theory model, and an edge-end cooperative reasoning scheme with price-demand-time delay closed-loop optimization is formed.

The invention belongs to the technical field of video quality evaluation, and particularly relates to a neural typing method based on a multi-modal brain map and semi-supervised deep clustering. Comprising the following steps: constructing a multi-modal brain network according to multi-modal neural image data; the brain network of each mode comprises a normal brain network and an abnormal brain network; performing pre-training on the MBVAE model by adopting a multi-modal brain network to obtain a pre-trained MBVAE model; performing fine tuning training on the pre-trained MBVAE model in combination with a clustering module to obtain a trained deep embedded clustering model; obtaining multi-modal neural image data of a user, constructing a multi-modal brain network, and inputting the multi-modal brain network into the trained deep embedded clustering model for processing to obtain a neural typing result; according to the method, the accuracy and reliability of ASD neural typing results are improved, the problems of gender imbalance and multi-site heterogeneity existing in ASD data are effectively solved, and the robustness and generalization ability of the model are enhanced.

The application relates to the field of air traffic control and speech recognition technology, and particularly relates to an air traffic control instruction end-to-end speech recognition method under high noise conditions, which comprises the following steps: pre-processing to-be-recognized speech to extract original speech frequency features; adopting an adaptive attention noise reduction module to perform noise reduction intensity control, frequency band gain control and pitch tracking processing on the original speech frequency features to obtain noise reduction speech; pre-processing the noise reduction speech to extract noise reduction speech frequency features; adopting a speech recognition module to perform shared encoder coding, connection time sequence classification decoder decoding and attention-based encoder-decoder decoding on the noise reduction speech frequency features to obtain output speech recognition transcription text; and the application can improve the speech recognition accuracy of air traffic control instructions under high noise conditions.

The invention relates to a neural network model training method and device, a neural network model calculation method and device, electronic equipment and a computer program product. According to the training method of the neural network model, the neural network model comprises a neural network layer, and the method comprises the following steps: dividing a target tensor in the neural network layer into at least two data blocks; for each data block, determining a corresponding public index according to the amplitude information of the elements in the data block; normalizing all elements in the data block relative to the common index to obtain a corresponding mantissa; for each element, only corresponding mantissa and sign bit are stored; the public indexes are stored according to data blocks, layers or channels; when the neural network model is trained, the public index, the bit width of the mantissa and the weight parameter of the neural network layer are updated through back propagation. According to the technical scheme, the requirements for the storage space and the memory bandwidth of the neural network model can be remarkably reduced on the premise that the precision is controllable.

The invention relates to the technical field of intelligent detection of road engineering, in particular to an intelligent detection method and system for pavement paving temperature uniformity based on an infrared thermal image.In the paving construction process, an infrared thermal imager behind a paver is used for collecting original infrared thermal image data, abnormal area segmentation and overall temperature uniformity grade marking are completed, and the pavement paving temperature uniformity is obtained. Constructing a training sample set; carrying out normalization and de-noising processing on the collected data; self-adaptive region segmentation is realized based on the temperature gradient, multi-dimensional statistical features are extracted, and a region feature map is generated; constructing a multi-source feature neural network, performing feature fusion and spatial weighting on the regional feature map, the de-noised data and the masks, and fusing prior knowledge to realize abnormal region pixel-level segmentation and temperature uniformity level evaluation; and designing a multi-task collaborative loss function to complete model training and parameter optimization. According to the invention, real-time, accurate and intelligent detection of pavement paving temperature distribution can be realized, and the automation level of temperature uniformity evaluation and defect identification is effectively improved.

The invention discloses a severe convective weather identification method and system based on a multi-stream deep neural network, and the method comprises the steps: obtaining meteorological observation data and reanalysis data of a target region, and extracting the physical quantity characteristics of an atmospheric environment; dividing the physical quantity characteristics of the atmospheric environment into a plurality of different characteristic categories according to the physical attributes of the physical quantity characteristics, wherein each characteristic category forms an independent input stream; constructing a multi-stream deep neural network; inputting each divided input stream feature into a sub-network branch corresponding to the multi-stream deep neural network, and independently performing high-level feature extraction on each sub-network branch to obtain a feature representation of each branch; integrating the feature representation extracted by each sub-network branch in a fusion layer to obtain a fusion feature; and based on the fusion features, discriminating the severe convective weather occurrence probability through an output layer of the multi-stream deep neural network. According to the invention, a technical path with both accuracy and transparency is provided for severe convection weather forecast, and accurate discrimination of short-time heavy rainfall is realized.

The present disclosure relates to the technical field of large-scale mobile robot scheduling, and provides a method and device for intelligent scheduling of large-scale mobile robots assisted by unmanned aerial vehicles, comprising: constructing a collaborative scheduling architecture comprising unmanned aerial vehicles, multiple mobile robots and edge servers; in the collaborative scheduling architecture, based on the communication delay, jitter and control error of each mobile robot, a joint optimization problem is constructed with the objective of minimizing real-time tracking error; the joint optimization problem is modeled as a Markov decision process; the control and communication joint action of the Markov decision process is determined by using an intelligent scheduling algorithm; and the mobile robot control instruction, the unmanned aerial vehicle flight path, the scheduling and phase configuration of the reconfigurable intelligent metasurface of the unmanned aerial vehicle, and the transmission power of the edge server in the collaborative scheduling architecture are adjusted in real time according to the control and communication joint action. The embodiment reduces the communication delay, jitter and trajectory tracking error of the mobile robot, and improves the stability and reliability of the system scheduling.

The application provides a low-orbit satellite uplink beam scheduling and power control method and device based on user distribution, the method comprising: constructing a total communication capacity optimization problem model of a low-orbit satellite uplink system according to distance distribution data between the low-orbit satellite and each user in the coverage area during over-the-top, and a Laplace transform expression corresponding to the low-orbit satellite uplink interference for representing user random distribution, channel fading gain and distance distribution; splitting the total communication capacity optimization problem model into a user beam association matrix optimization sub-problem and a low-orbit satellite uplink power optimization sub-problem and solving them respectively to obtain a target beam association matrix and target transmission power of each user. The application can more accurately describe the uplink coverage performance of the low-orbit satellite during the entire over-the-top period, can realize the joint optimization of the low-orbit satellite uplink beam scheduling and power control, and can reduce the complexity of algorithm calculation.

This application relates to the field of electric official vehicle travel demand management technology, and discloses a method and system for predicting electric official vehicle travel demand. The method includes collecting data on the mileage, travel time, charging records, battery capacity, unit energy efficiency, meteorological information, and traffic flow of electric official vehicles to obtain raw data aligned with a unified timestamp; performing missing value imputation, outlier removal, and time series resampling based on the raw data to form a first dataset. By using empirical mode decomposition to decompose multi-source non-stationary time-series signals, employing arithmetic optimization algorithms to dynamically optimize the parameters of the long short-term memory network, and performing multi-period travel and energy consumption prediction based on the optimized long short-term memory network model, the method addresses the problems of unstable prediction accuracy, insufficient parameter adaptability, and poor model convergence in existing technologies under multi-source nonlinear data environments.

This invention discloses a multi-layer federated reinforcement learning collaborative scheduling method for automated yards, relating to the fields of intelligent scheduling and crane technology. It constructs a multi-yard crane collaborative scheduling optimization method that balances scheduling performance and privacy protection by having multiple yard crane clients collaboratively participate in federated reinforcement learning model training, within a hierarchical architecture of "yard crane local client—yard edge node—scheduling control center". This method reduces the frequency of model parameter transmission between yard cranes and yard edge nodes through client-side local training and a hierarchical federated aggregation mechanism, thereby reducing communication overhead. It enhances the consistency and generalization ability of scheduling strategies between different container areas by utilizing yard edge node model aggregation, while introducing a dynamic weight allocation mechanism to balance the contribution of each yard crane in the global model update. Furthermore, it improves the stability and robustness of the scheduling strategy under dynamic disturbance environments by constructing a reward function, thus achieving efficient collaborative scheduling of multi-yard crane operations in automated yards.

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.