Intelligent transport capacity dispatching system and method based on GPS

Abstract

The invention discloses an intelligent transport capacity dispatching system and method based on a GPS. The system comprises a transport capacity state monitoring subsystem and an intelligent dispatching management subsystem. And by deploying a GPS positioning and multi-parameter sensor, real-time acquisition and analysis of information such as the position, the speed and the energy consumption of the transport capacity unit are realized. According to the system, CNN and Transform structure extraction features are fused, LSTM is combined to predict a transport capacity demand, and a PPO reinforcement learning algorithm is adopted to dynamically plan a path. Abnormality detection (SVM), PID feedback adjustment and cloud edge cooperative calculation are integrated in the scheduling process, and the scheduling precision and the response speed are improved. The system has the capabilities of task filing, online optimization and data security guarantee, and is suitable for complex traffic and logistics scheduling scenes.

Description

Technical Field

[0001] The present invention relates to the field of transportation scheduling and intelligent optimization control, and in particular to an intelligent transportation capacity scheduling method and system that integrates GPS real-time positioning, cloud-edge collaborative computing and intelligent scheduling algorithms. Background Art

[0002] With the rapid development of smart logistics and intelligent transportation systems, the efficient scheduling of transportation resources has become a core issue in transportation management. In particular, in scenarios such as urban distribution, emergency transportation, and shared mobility, accurately grasping the real-time location of transportation units and dynamically allocating tasks has a critical impact on the transportation system's response speed and resource utilization. However, traditional transportation scheduling methods often rely on manual rules or static scheduling models, lacking real-time awareness of transportation status and traffic conditions. These methods are unable to adapt to dynamic and changing task requirements and complex road conditions, often leading to problems such as route conflicts, resource waste, and response delays.

[0003] While existing technologies have incorporated some GPS positioning and route recommendation capabilities, they still lack the ability to jointly model and intelligently optimize multiple factors, including the operational status of transport units, mission urgency, and traffic conditions. Furthermore, dispatch systems, often based on a centralized computing architecture based on central servers, face processing bottlenecks when faced with frequent positioning updates and multiple concurrent tasks, making it difficult to achieve efficient coordination and real-time decision-making for multiple transport units.

[0004] This invention proposes an intelligent transport scheduling system that integrates GPS real-time positioning, edge computing, deep learning scheduling algorithm and task closed-loop feedback mechanism. In response to the complex needs of dynamic management of multiple transport units, task priority allocation and path optimization, it constructs a full-process intelligent scheduling technology path from location perception, path prediction, intelligent allocation to feedback optimization, which significantly improves the system's perception of transportation resources, allocation efficiency and emergency response capabilities, and is suitable for various application scenarios such as urban traffic management, logistics distribution, and online car-hailing scheduling. Summary of the Invention

[0005] The purpose of the present invention is to provide a GPS-based intelligent transport capacity scheduling method and system, which can integrate real-time positioning data with intelligent scheduling algorithms to achieve accurate position tracking, status identification, dynamic task allocation and intelligent path optimization of multiple transport capacity units, thereby improving the utilization efficiency of transportation resources and task responsiveness.

[0006] To achieve the above-mentioned purpose, the present invention provides a GPS-based intelligent transportation capacity scheduling system, comprising: a GPS-based intelligent transportation capacity scheduling system, the system comprising:

[0007] Intelligent capacity status monitoring subsystem and intelligent capacity scheduling and management subsystem;

[0008] The intelligent transport status monitoring subsystem comprises:

[0009] The transport unit body;

[0010] The multi-parameter monitoring network component is deployed inside and outside the transport unit body, comprising a position and status monitoring submodule that collects the GPS position, speed, driving trajectory, fuel consumption and load status parameters of the transport unit in real time;

[0011] The edge computing unit is deployed in the transport unit or adjacent position, receives the data collected by the position and status monitoring submodule, and performs data preprocessing, feature extraction, fusion analysis and status recognition tasks;

[0012] The data fusion mechanism is integrated in the edge computing unit, comprising a convolutional neural network (CNN) that extracts local features of transport operation, and a Transformer architecture that performs correlation modeling of multi-modal data to realize fusion analysis of different monitoring parameter data in the time domain and the spatial domain;

[0013] The intelligent scheduling module comprises a long short-term memory network (LSTM)-based transport demand and status time series analysis submodule that learns the historical transport demand and vehicle operation status change pattern to predict future short-term demand trends, and a dynamic path planning unit that dynamically adjusts the vehicle driving path based on real-time traffic information, vehicle status and demand prediction results to realize fine scheduling of transport tasks, and the control strategy is constructed based on reinforcement learning algorithm and real-time optimization algorithm;

[0014] The transport intelligent scheduling and management subsystem comprises a hierarchical management architecture, comprising:

[0015] The perception layer uses fixed roadside sensors and vehicle-mounted sensors to collect road environment parameters and real-time status data of the transport unit;

[0016] The data transmission layer transmits the data collected by the perception layer to the upper layer decision analysis module, the transmission protocol includes Modbus, I2C, UDP, CAN, and local area network or 5G network is used for data transmission;

[0017] The decision-making layer uses edge computing units and cloud platforms to perform real-time analysis and anomaly diagnosis of sensor data, supports local and global data fusion, and conducts capacity scheduling assessment and predictive analysis based on deep learning and a knowledge base of traffic management experts. The decision-making layer includes anomaly detection and early warning models, including: identifying anomalies in vehicle trajectory and driving behavior based on Bayesian networks, and diagnosing vehicle faults or abnormal status trends based on support vector machines (SVMs);

[0018] Closed-loop task scheduling and execution process, including:

[0019] The task work order generation and scheduling submodule automatically generates scheduling tasks based on real-time monitoring data and abnormality diagnosis results, and allocates tasks based on the urgency of the transportation task and the availability of transportation units;

[0020] The task execution submodule guides the transport unit to the designated location to perform the transport task based on GIS positioning and dispatching task information, and uses the vehicle terminal and system to perform real-time data update and feedback;

[0021] The feedback optimization submodule uses real-time task execution data and capacity status data to perform feedback optimization and iterative upgrades on the path planning algorithm and anomaly detection model.

[0022] A GPS-based intelligent transportation capacity scheduling method includes the following steps:

[0023] S01: Use the GPS positioning system to obtain real-time geographic location data of multiple transport resources, and collect vehicle status data such as speed, battery power, and fuel level;

[0024] S02: Based on real-time location data, order urgency, vehicle status, and road condition information, the cloud-based dispatching platform calculates the optimal transport scheduling plan and performs task allocation.

[0025] S03: Use a convolutional neural network (CNN) to analyze and predict the vehicle's task priority. The convolutional neural network (CNN) processes road conditions, vehicle status, and task information into images, extracts image features, and predicts task priority. The CNN calculation process includes:

[0026] S04: Based on the calculated scheduling plan, use the Proximal Policy Optimization (PPO) algorithm to perform path planning;

[0027] S05: Use PID feedback control algorithm to dynamically adjust the vehicle speed during task execution;

[0028] S06: During the mission execution, a support vector machine (SVM) is used to monitor the vehicle status and provide fault warnings. The SVM classifies characteristic data such as the vehicle's current position, path deviation, and fuel level to identify whether there is a fault or mission deviation.

[0029] S07: After the task is completed, the task scheduling strategy is optimized through the PPO algorithm and the CNN feedback mechanism. The optimization process includes learning from historical scheduling tasks and feedback results, and automatically adjusting the scheduling strategy to improve the efficiency and accuracy of the system.

[0030] The basic form of the PID feedback regulation algorithm is:

[0031] Where, e(t) = T c (t)-T set is the temperature deviation, K p , K i , K d are the proportional, integral and derivative gains respectively.

[0032] Convolutional neural network (CNN) specifically includes the following steps: S01: Use convolution operation to extract local features of the image; S02: Convolution kernel Convolve the input image to get the feature map: Where I is the input image and F is the feature map;

[0033] S03: Use activation function to increase nonlinear expression ability; S04: Predict output y.

[0034] Predict the output y using the loss function The loss function is cross entropy loss: in, is the probability value predicted by the CNN model.

[0035] The SVM diagnosis steps include: S01: find an optimal hyperplane to maximize the interval; S02: through the constraint condition: y i (w T x i +b)≥1,i=1,…,N; the optimization goal is to maximize the interval That is, minimize the objective function: S03: Introducing the Lagrange multiplier α i , the obtained support vector solution is used for classification prediction: where α i It is the weight coefficient of the support vector, and the prediction result f(x) determines whether the device has a negative correlation problem.

[0036] PPO algorithm, the processing steps are S01: initializing environment and task setting; S02: calculating value function (V(s)); S03: policy gradient (PPO) optimization; S04: evaluating action-state value function (Q(s,a)); S05: training and optimization of PPO.

[0037] Further, the system further enhances the identification ability of the transport capacity scheduling system for the running state of the transport capacity unit by introducing additional transport capacity state parameters such as fuel consumption rate, load change and driving trajectory stability.

[0038] Further, the integrated intelligent transport capacity scheduling system of the present application generates an optimal scheduling strategy according to real-time task demand, task urgency, transport capacity unit load state and path accessibility and other factors during the transportation process through the collaborative scheduling of multiple types of transport capacity units. The system supports multiple transport capacity forms, including city distribution vehicles, online car hailing, autonomous mobile logistics vehicles, etc. Through the collaborative operation and local decision synchronization mechanism of each transport capacity unit, real-time exchange of task state is realized, enabling efficient and intelligent execution of scheduling tasks, adapting to variable and high-frequency city and logistics distribution scenarios.

[0039] Further, the present application is suitable for long-term stable scheduling operation in various complex transportation environments such as high temperature, high cold, high humidity and long distance driving. Therefore, the system uses high-strength vehicle embedded devices that meet the intelligent transportation equipment protection standards, the device shell is made of impact-resistant composite materials and weather-resistant coating, and combined with temperature control isolation structure and multi-layer EMI shielding technology, to ensure the stable operation of the edge computing module in complex climate and electromagnetic interference environment, and to ensure the robustness and reliability of the system.

[0040] Further, the present application optimizes the collection accuracy and response timeliness of scheduling parameters in large-scale transport capacity scheduling systems, including path length, traffic density, fuel consumption cost, vehicle grade, etc., supports heterogeneous signal fusion from GPS, OBD, fuel consumption meter and other devices, and realizes spatial state collaborative analysis between multiple transport capacity units combined with graph convolutional neural network algorithm. In terms of fleet layout and state perception strategy, combined with static base station information and dynamic vehicle-mounted perception unit, a flexible and adjustable transport capacity information collection network is constructed to improve the perceptibility and scheduling coverage capability of the system in multi-region and distributed transportation scenarios.

[0041] Preferably, the present application further integrates a multi-modal transport capacity state collaborative perception module, including a high-speed driving stability evaluation component, a low-power fuel consumption monitoring module, an IMU inertial unit, a high-precision GPS and a route deviation monitor, etc., through CAN bus and I 2 ​The parallel communication method of the C protocol synchronizes various status information to the edge control core. The system builds a time series state modeling layer based on long short-term memory (LSTM) networks and gated recurrent units (GRU). Combining a CNN and Transformer fusion architecture, it extracts operational trends and potential anomalies, dynamically adjusting scheduling cycles, route generation strategies, and capacity matching parameters. This significantly improves the system's adaptability and control capabilities to complex transportation demands and dynamic traffic conditions.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] First, the present invention solves the problem of scattered and weakly correlated status data in the capacity scheduling process. Through multimodal data fusion technology, it effectively integrates multi-source information from GPS, vehicle status monitoring, traffic environment perception, etc., and uses graph convolutional networks (GCN) to perform deep modeling in spatial and temporal dimensions to construct a global scheduling situation map. This mechanism realizes real-time holistic modeling of the operating status of capacity units and the traffic environment, significantly improving the system's comprehensive perception and judgment capabilities for complex transportation tasks, and is particularly suitable for multi-capacity, multi-task, and high-concurrency scheduling scenarios.

[0044] Secondly, the present invention uses a long-short-term memory (LSTM) network to construct a time-series forecasting model for capacity demand and operating status, combining vehicle operating efficiency and historical driving routes to form a closed-loop scheduling and control mechanism. The LSTM model can learn the fluctuation trends of task demand in different regions and predict future peak and valley periods. This assists the system in optimizing capacity allocation and routing strategies in advance, thereby avoiding resource redundancy or delivery delays and enhancing the stability and adaptability of the overall transportation system.

[0045] Thirdly, this invention addresses the challenges of traditional transport scheduling systems, such as rigid task allocation, frequent path conflicts, and low scheduling efficiency. By integrating reinforcement learning with game reasoning mechanisms, the system can optimize the dynamic allocation of transport resources and path planning strategies in real time based on multiple factors, including task urgency, vehicle operating status, road congestion, and task coverage. This scheduling system not only intelligently generates optimal dispatch plans but also fine-tunes strategies based on real-time execution, significantly improving scheduling response speed and resource utilization, enabling adaptive collaboration across multiple transport systems.

[0046] Finally, this invention significantly reduces the communication and computing pressure on the central system by offloading data processing and intelligent decision-making tasks to edge computing nodes, allowing large amounts of real-time positioning and state perception data to be processed on the vehicle side or near the vehicle. Leveraging edge intelligent decision-making capabilities, the system can complete local path adjustments and task reception responses within milliseconds, improving the dispatch system's processing power and response speed under high-frequency updates and high-concurrency commands, ensuring continuous and stable operation in complex urban traffic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] To more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or exemplary embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the technical solutions shown in these drawings without inventive effort.

[0048] Figure 1 This is the flow chart of the GPS-based intelligent transportation dispatching system;

[0049] Figure 2 This is the timing diagram of the GPS-based intelligent transportation scheduling system;

[0050] Figure 3 It is the intelligent capacity dispatch trend chart of the GPS-based intelligent capacity dispatch system;

[0051] Figure 4 This is the C4 diagram of the GPS-based intelligent transportation dispatching system; DETAILED DESCRIPTION

[0052] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0053] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.

[0054] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0055] See Figure 1 The following is a flowchart of a GPS-based intelligent transport dispatch system. The dispatch process begins with data collection. The onboard GPS module and status sensors synchronously collect operational information such as the transport unit's location, speed, trajectory, load status, and fuel consumption. This collected data is uploaded in real time to an edge computing unit deployed on the vehicle or a nearby location for preliminary preprocessing, including key feature extraction and time series alignment, to improve the efficiency and accuracy of subsequent dispatch analysis.

[0056] Preprocessed data is fed into the anomaly detection module, where the system uses built-in models to identify unusual patterns in operational behavior. If an anomaly is detected, such as unusual parking, overloading, or route deviation, the system immediately triggers a risk warning and leverages relevant control mechanisms (such as task reallocation and route adjustment instructions) for emergency intervention. If the status is normal, the data is stored in the capacity database for subsequent scheduling model training and system optimization. This process, through a closed-loop control strategy, enables dynamic perception and real-time response to capacity operational status. Multi-source data from transport units and the road environment is processed in real time via an edge gateway. After extracting key spatiotemporal features, only high-value scheduling information is reported to the cloud platform for in-depth predictive analysis and policy optimization. At the system design level, the synergy between edge computing and fusion analysis relies on efficient spatiotemporal models and lightweight algorithmic frameworks. For example, the system utilizes a spatiotemporal data fusion mechanism that combines unified timestamp alignment and spatial interpolation techniques to address the challenges of coordinating multi-source sensor data with inconsistent sampling frequencies and heterogeneous formats, thereby ensuring high real-time performance and decision-making accuracy in a multi-node, dynamic operating environment. The mathematical expression of the scheduling method can be expressed as: in, is the fused data, w i (t) is the weight coefficient, x i (t) is the raw data of the i-th sensor. By dynamically adjusting the weight coefficient, the algorithm can adapt to the data characteristics in different scenarios and ensure the accuracy of the fusion results.

[0057] See Figure 2 , which is the timing diagram of the GPS-based intelligent transportation scheduling system;.

[0058] When using a long short-term memory (LSTM) network for temperature prediction, a series of preset values ​​are required for the model. The LSTM network parameter settings have a significant impact on prediction accuracy. The following is a table of LSTM model parameter presets, see Table 1.

[0059] Table 1

[0060]

[0061] In this context, the forget gate, input gate, and output gate mechanism introduced into the long short-term memory (LSTM) network structure enable the model to retain key historical features in multi-step prediction tasks, effectively avoiding the gradient vanishing problem common in traditional recurrent neural networks, thereby improving the model's performance in long-term dependency modeling.

[0062] In capacity demand and route planning forecasting, the core responsibility of the LSTM model is to leverage multi-dimensional time series data, including historical transport mission records, GPS trajectory data, road conditions, and vehicle loads, to infer capacity demand trends and scheduling pressure distribution within a specific time window. Given a continuous sequence of vehicle operating states and external environmental parameters, the model dynamically learns the underlying temporal dependencies and outputs forecasts of scheduling demand at several future moments. This assists the system in pre-routing routes and pre-deploying capacity resources, effectively improving the system's overall operational efficiency and emergency response capabilities.

[0063] Given time series data (x1, x2, ..., x t )(where x t The model training process continuously updates network weights to minimize prediction error and optimize prediction results. The LSTM structure allows it to extract long-term dependencies from sequences, which is particularly important for data types that are highly dependent on historical state. Specifically, the LSTM performs forward propagation through the following process:

[0064] ht=σ(Wh·[xt,ht-1]+bh) is one of the core formulas of LSTM, where: t is the input data at the current moment; h t-1 is the state at the previous moment; W h is the weight matrix; b h The bias σ is the activation function, and the sigmoid function is usually used to control the degree of information passing.

[0065] Specifically:

[0066] S01 Forget gate calculation: determines which past memories the network forgets t =σ(W f x t +U f h t-1 +b f ); S02 input gate calculation: control the update of current input information i t =σ(W i x t +U i h t-1 +b i ); S03 candidate cell state calculation: generate candidate memory at the current moment, S04 Cell state update: Combine the results of the forget gate and the input gate to update the cell state. S05 output gate calculation: determines the content of the output at the next moment, o t =σ(W o x t +Uo h t-1 +b o ); S06 hidden state update: update the current hidden state through the control of the output gate, h t =o t tanh(c t );

[0067] Once model training is complete, the system can output a forecast of transport demand for future periods, which can be compared with the regional dispatch load threshold. If the predicted transport demand exceeds the preset carrying capacity or dispatch load threshold, the system will automatically trigger a transport resource allocation mechanism, such as increasing available vehicles, adjusting task priorities, or optimizing driving routes. This linkage control mechanism achieves dynamic optimization and precise scheduling of transport resources by connecting the time series forecast results with the dispatch execution system in real time. This not only enhances the system's ability to respond to sudden demands, but also significantly improves the operating efficiency and service reliability of the overall transport system.

[0068] By introducing the PID control algorithm, the transport capacity system can be precisely adjusted. c When (t) exceeds the set value, the basic form of the PID controller is: Where, e(t) = T c (t)-T set is the temperature deviation, K p , K i , K d are the proportional, integral, and differential gains, respectively. This formula ensures that the system can keep the vehicle within the forward-related operating range by dynamically adjusting the system's operating state.

[0069] SVM diagnoses insulation degradation trends:

[0070] Support Vector Machine (SVM) is a binary classification model that separates data of different categories by constructing a hyperplane. In insulation degradation diagnosis, assuming that the input data is x∈R n Represents the various monitoring features of the equipment, and the output y∈{-1,1} represents the insulation status (-1 means normal, 1 means deteriorated). The goal of SVM is to find an optimal hyperplane that maximizes the interval: w T In x+b=0, w is the normal vector and b is the bias. By the constraint condition: y i (w T x i +b)≥1,i=1,…,N; the optimization goal is to maximize the interval That is, minimize the objective function:

[0071] Through the Lagrange multiplier method, the Lagrange multiplier α is introduced i To solve the constrained optimization problem. Finally, the support vector solution can be used for classification prediction:

[0072] The diagnostic system, built by combining convolutional neural networks (CNN) and support vector machines (SVM), can perform a multi-dimensional health analysis of the operating status of transport units. CNN is used to analyze image data from on-board cameras or sensor terminals to identify appearance issues such as vehicle structural anomalies, body damage, and tire wear; SVM assesses the deterioration trend of vehicle performance based on operating parameters collected by sensors (such as engine temperature, fuel consumption rate, vibration characteristics, etc.). The collaborative work of the two significantly improves the state perception and fault identification capabilities of the transport system under multi-modal input, thereby achieving more accurate fault warnings and maintenance decisions, and reducing scheduling risks caused by delayed manual inspections or inaccurate experience judgments.

[0073] If the system needs to be further integrated into edge computing nodes, real-time local diagnosis and fault level assessment can be achieved, and vehicles with abnormal status can be prioritized for elimination or replacement during task scheduling, thereby improving the safety and robustness of overall transport scheduling.

[0074] See Figure 3 、 Figure 4 , Figure 3 Intelligent capacity dispatching trend chart for GPS-based intelligent capacity dispatching system, Figure 4 This is the C4 diagram of the GPS-based intelligent transportation dispatching system;

[0075] In a GPS-based intelligent transportation dispatch system, the system dynamically allocates transportation resources through the collaboration of multi-source data streams and a model-driven mechanism. The overall process begins with the real-time collection of vehicle location data and order information. The system simultaneously obtains each vehicle's GPS coordinates, cargo status, and estimated arrival time. It also accesses information such as the total number of orders, weight, volume, and route requirements for the target region. After collection, the data is fed into an edge gateway for preliminary processing, primarily involving feature extraction and data alignment, thereby improving the efficiency and stability of subsequent computing modules.

[0076] The system then dynamically calculates the unit price for freight at the destination by combining key metrics such as transport distance, number of orders, and vehicle availability. This calculation model utilizes a data-driven pricing function, taking into account market supply and demand structures and comparing it with historical trends to determine whether to trigger a dispatch warning. If the unit price in a particular region falls below a safety threshold, or if vehicle supply is far insufficient to cover existing orders, the system automatically sends a warning to dispatchers via an app or text message.

[0077] At the intelligent scheduling decision-making level, the system further analyzes the forecast results and current capacity distribution, optimizing vehicle routing and task allocation strategies, and automatically directing resources from high-density areas to areas of supply and demand imbalance. To further improve scheduling accuracy, the system incorporates a multimodal diagnostic subsystem that combines a convolutional neural network (CNN) and a support vector machine (SVM). The CNN assesses vehicle status (such as vehicle aging) from image data, while the SVM detects task execution efficiency and equipment risk trends using sensor data, enabling more robust capacity health assessments.

[0078] The entire system operates on a closed-loop management architecture and edge-cloud collaboration. The edge rapidly responds to environmental changes on the front lines and filters low-value information, while the cloud performs in-depth analysis and long-term model updates. This structure effectively supports the stable operation of the capacity scheduling system in dynamic and complex scenarios, significantly improving resource utilization, scheduling response speed, and transportation safety.

[0079] Instructions for use:

[0080] The intelligent transport dispatch system leverages multimodal sensing and edge computing technologies to enable dynamic route planning and efficient task scheduling for transport tasks. Upon system startup, multi-parameter sensors deployed inside and outside the transport unit simultaneously collect key data such as vehicle GPS location, speed, trajectory, load status, and fuel consumption. This real-time data serves as the basis for subsequent route planning and dispatch decisions.

[0081] The collected data is preprocessed by the edge computing unit, including feature extraction, time alignment, and data fusion. The system embeds a Transformer-based spatiotemporal data fusion model and a convolutional neural network (CNN) to improve the accuracy of vehicle operating status recognition and response efficiency. Subsequently, the preprocessed key features are input into the long short-term memory (LSTM) module for time series modeling, predicting near-term transportation demand and scheduling bottlenecks.

[0082] For route planning and task allocation, the system integrates reinforcement learning algorithms with real-time optimization strategies. By simulating future state changes and road conditions, it calculates the optimal driving route and task assignment plan. Each scheduling decision not only considers the current distribution of transportation resources, but also incorporates historical transportation patterns and task urgency, ensuring high adaptability and resource utilization.

[0083] The system uses a unified communication protocol (such as UDP, CAN, or 5G networks) to achieve efficient data transmission, ensuring state synchronization and coordinated execution among multiple transport units. During mission execution, vehicles transmit their mission status, location trajectory, and exception information back to the dispatch center in real time. The dispatch center uses a feedback mechanism to dynamically adjust strategies, optimize routes, or reallocate tasks, ensuring timely responsiveness and accurate scheduling in multi-tasking scenarios.

[0084] To ensure data transmission security and the reliability of task scheduling, the system has designed an end-to-end digital signature and authentication mechanism. Each piece of transmitted data is signed by the vehicle using its private key, and the dispatch system uses its public key to verify its origin and integrity, preventing data tampering and unauthorized access.

[0085] The system also integrates a knowledge base and online learning mechanisms. All task execution logs, fault records, and scheduling feedback are automatically archived in a database and synchronized to the cloud platform. This archived information is meticulously labeled, providing a rich training sample for diagnostic model iteration. Furthermore, the system utilizes incremental learning to continuously optimize fault identification models and scheduling algorithms, enabling them to continuously adapt to new operating modes and abnormal situations.

[0086] When an anomaly or failure is detected during task execution, the system automatically generates a maintenance work order and, using GIS positioning, marks the fault location. The system then intelligently assigns maintenance personnel with relevant skills to perform the task based on task attributes and personnel permissions. Once the repair is complete, the system records the repair process data and performs post-verification to ensure the issue is fully resolved and the equipment is restored to normal operation.

[0087] Through closed-loop control mechanism, edge-cloud collaborative computing architecture and deep intelligent model, this system has built an intelligent capacity management platform with prediction, adaptive scheduling and automatic optimization capabilities, greatly improving the efficiency, safety and sustainability of transportation scheduling.

Claims

1. The GPS-based intelligent transport capacity dispatching system is characterized by: The system includes: Intelligent capacity status monitoring subsystem and intelligent capacity scheduling and management subsystem; The intelligent transport capacity status monitoring subsystem includes: Capacity unit body; A multi-parameter monitoring network component is deployed inside and outside the transport unit, including: a position and status monitoring submodule, which collects the GPS position, speed, driving trajectory, fuel consumption and load status parameters of the transport unit in real time; The edge computing unit is deployed at the transport unit or a nearby location, receives the data collected by the location and status monitoring submodule, and performs data preprocessing, feature extraction, fusion analysis, and status recognition tasks; A data fusion mechanism, integrated into the edge computing unit, includes: a convolutional neural network (CNN), which extracts local features of transport operations; a Transformer architecture, which performs correlation modeling of multimodal data and enables fusion analysis of different monitoring parameter data in the time and space domains; The intelligent scheduling module includes: a transportation demand and status time series analysis submodule based on a long short-term memory network (LSTM), which learns historical transportation demand and vehicle operating status change patterns to predict future short-term demand trends; a dynamic path planning unit, which dynamically adjusts vehicle driving paths based on real-time road conditions, vehicle status, and demand forecast results to achieve refined scheduling of transportation tasks. The control strategy is based on a reinforcement learning algorithm and a real-time optimization algorithm. The intelligent transportation dispatching and management subsystem includes a hierarchical management architecture, including: A perception layer, which uses fixed roadside sensors and vehicle-mounted sensors to collect real-time status data of road environment parameters and transport units; The data transmission layer transmits the data collected by the perception layer to the upper-layer decision analysis module. The transmission protocols include Modbus, I2C, UDP, and CAN, and use local area networks or 5G networks for data transmission; The decision-making layer uses edge computing units and cloud platforms to perform real-time analysis and anomaly diagnosis of sensor data, supports local and global data fusion, and conducts capacity scheduling assessment and predictive analysis based on deep learning and a knowledge base of traffic management experts. The decision-making layer includes anomaly detection and early warning models, including: identifying anomalies in vehicle trajectory and driving behavior based on Bayesian networks, and diagnosing vehicle faults or abnormal status trends based on support vector machines (SVMs); Closed-loop task scheduling and execution process, including: The task work order generation and scheduling submodule automatically generates scheduling tasks based on real-time monitoring data and abnormality diagnosis results, and allocates tasks based on the urgency of the transportation task and the availability of transportation units; The task execution submodule guides the transport unit to the designated location to perform the transport task based on GIS positioning and dispatching task information, and uses the vehicle terminal and system to perform real-time data update and feedback; The feedback optimization submodule uses real-time task execution data and capacity status data to perform feedback optimization and iterative upgrades on the path planning algorithm and anomaly detection model.

2. The GPS-based intelligent transportation capacity scheduling method is characterized by: The method comprises the following steps: S01: Use the GPS positioning system to obtain real-time geographic location data of multiple transport resources, and collect vehicle status data such as speed, battery power, and fuel level; S02: Based on real-time location data, order urgency, vehicle status, and road condition information, the cloud-based dispatching platform calculates the optimal transport scheduling plan and performs task allocation. S03: Use a convolutional neural network (CNN) to analyze and predict the vehicle's task priority. The convolutional neural network (CNN) processes road conditions, vehicle status, and task information into images, extracts image features, and predicts task priority. The CNN calculation process includes: S04: Based on the calculated scheduling plan, use the Proximal Policy Optimization (PPO) algorithm to perform path planning; S05: Use PID feedback control algorithm to dynamically adjust the vehicle speed during task execution; S06: During the mission execution, a support vector machine (SVM) is used to monitor the vehicle status and provide fault warnings. The SVM classifies characteristic data such as the vehicle's current position, path deviation, and fuel level to identify whether there is a fault or mission deviation. S07: After the task is completed, the task scheduling strategy is optimized through the PPO algorithm and the CNN feedback mechanism. The optimization process includes learning from historical scheduling tasks and feedback results, and automatically adjusting the scheduling strategy to improve the efficiency and accuracy of the system.

3. The GPS-based intelligent transportation capacity scheduling method according to claim 2, characterized in that: The basic form of the PID feedback regulation algorithm is: Where, e(t) = T c (t)-T set is the temperature deviation, K p , K i , K d are the proportional, integral and derivative gains respectively.

4. The GPS-based intelligent transportation capacity scheduling method according to claim 2, characterized in that: The convolutional neural network (CNN) specifically includes the following steps: S01: using convolution operation to extract local features of the image; S02: convolution kernel Convolve the input image to get the feature map: Where I is the input image and F is the feature map; S03: Use activation function to increase nonlinear expression ability; S04: Predict output y.

5. The GPS-based intelligent transportation capacity scheduling method according to claim 4, characterized in that: The predicted output y uses the loss function The loss function is cross entropy loss: in, is the probability value predicted by the CNN model.

6. The GPS-based intelligent transportation capacity scheduling method according to claim 2, characterized in that: The SVM diagnosis steps include: S01: finding an optimal hyperplane to maximize the interval; S02: through the constraint condition: y i (w T x i +b)≥1,i=1,…,N; the optimization goal is to maximize the interval That is, minimize the objective function: S03: Introducing the Lagrange multiplier α i , the obtained support vector solution is used for classification prediction: where α i It is the weight coefficient of the support vector, and the prediction result f(x) determines whether the device has a negative correlation problem.

7. The GPS-based intelligent transportation capacity scheduling method according to claim 2 is characterized in that The PPO algorithm specifically includes: S01: initialization of environment and task settings; S02: calculation of value function (V(s)); S03: strategy gradient ( Optimization; S04: Evaluation of the action-state value function (Q(s,a)); S05: Training and optimization of PPO.

Images