Efficient automatic scheduling method for road obstacle clearance rescue workers

Abstract

The invention relates to the field of intelligent traffic management, and discloses an efficient road obstacle clearance rescue worker automatic scheduling method comprising the following steps: (a) deploying at least one message queue to cache rescue related data; (b) storing and querying rescue tasks and resource state data through a distributed database system; (c) collecting data from the rescue vehicle and the accident site in real time, and performing preprocessing; (d) asynchronously processing the collected data through a message queue according to the priority and the task type; (e) dynamically scheduling rescue resources by using a multi-factor decision algorithm; the message queue adopts a Kafka cluster and is configured with at least three partitions and at least two copies so as to ensure that a single server or partition fails. Through combination of a multi-factor decision scheduling algorithm and real-time traffic flow prediction, the system can dynamically adjust the scheduling route of the rescue vehicle, and the rescue vehicle is prevented from entering a congested road section, so that the time for arriving at an accident scene is remarkably shortened, and the overall rescue efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent traffic management, and in particular to a method for automatically dispatching efficient road obstacle removal and rescue personnel. Background Art

[0002] With the acceleration of modern urbanization, road traffic is becoming increasingly complex, especially during peak traffic hours or when accidents occur frequently. The demand for road clearance and rescue tasks is gradually increasing. In order to ensure smooth roads and reduce the impact of traffic accidents, the timeliness and efficiency of road clearance and rescue work have become particularly important. The rapid response of road rescue not only helps to reduce the possibility of secondary accidents, but also reduces traffic congestion caused by accidents, thereby improving the operating efficiency of the overall transportation system.

[0003] In the prior art, road rescue dispatch mostly relies on fixed rules or manual dispatch schemes, and often lacks intelligent dynamic dispatch functions. In traditional dispatch systems, the dispatch path of rescue vehicles is usually a simple shortest path calculation based on the current geographical location and the location of the accident, ignoring the impact of actual road conditions and future traffic changes. In addition, the allocation of rescue resources is often based on static task priority division, which is difficult to achieve dynamic adjustment, resulting in insufficient utilization of some resources or delayed response. As the complexity of traffic conditions increases, existing dispatch systems are prone to scheduling bottlenecks when dealing with large-scale, multi-task concurrent rescue scenarios, resulting in low overall rescue efficiency.

[0004] However, the existing technology has the following problems when facing complex traffic environments: the dispatching system cannot dynamically adjust the rescue route according to the real-time traffic flow, especially in traffic congestion. Rescue vehicles are easily stuck in congested sections, which prolongs the time to reach the accident scene and affects the rescue efficiency. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides an efficient method for automatically dispatching road obstacle clearance rescue personnel, which solves the problem that the dispatching system is unable to dynamically adjust the rescue route according to the real-time traffic flow, resulting in prolonged time for rescue vehicles to arrive at the accident scene, affecting the rescue efficiency.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an efficient road obstacle removal and rescue personnel automatic dispatching method, comprising the following steps:

[0007] (a) deploying at least one message queue to cache rescue-related data;

[0008] (b) store and query rescue mission and resource status data through a distributed database system;

[0009] (c) Collect data from rescue vehicles and accident sites in real time and perform pre-processing;

[0010] (d) asynchronously process the collected data through message queues according to priority and task type;

[0011] (e) Utilize multi-factor decision-making algorithms to dynamically dispatch rescue resources.

[0012] Preferably, the message queue uses a Kafka cluster configured with at least three partitions and at least two replicas to ensure continued availability and processing of data when a single server or partition fails.

[0013] Preferably, the distributed database system adopts Cassandra, and the primary key in its data model design includes a combination of timestamp and geographic location to optimize the deployment of rescue resources and the query efficiency of response time.

[0014] Preferably, the data preprocessing step further comprises using regular expressions to validate and format the format of GPS data collected from the sensor.

[0015] Preferably, the asynchronous processing step further includes using a priority queue algorithm to prioritize the collected data, giving priority to rescue missions with high accident severity and the closest distance.

[0016] Preferably, step (e) further includes monitoring and optimizing resource allocation and task execution efficiency, wherein the multi-factor decision algorithm determines the task priority based on the following formula:

[0017]

[0018] Among them, P(i) represents the priority of task i, U(i) and R(i) represent the urgency, data size and response time of the task respectively, and w1, w2, w3 are weight factors.

[0019] Preferably, the weight factors w1, w2, w3 of the multi-factor decision algorithm are determined based on statistical analysis of historical data, and the optimal value of each factor is determined through machine learning model training to adapt to different traffic and accident situations, wherein the training step includes:

[0020] (a) Collect data on historical rescue missions;

[0021] (b) Evaluate the rescue efficiency under different weight combinations;

[0022] (c) Gradient descent method is used to optimize the weight factors to minimize the rescue response time.

[0023] Preferably, the resource allocation and performance monitoring step further includes implementing a feedback-based dynamic resource allocation system, which adjusts rescue resource allocation according to real-time road conditions and real-time locations of rescue vehicles.

[0024] Preferably, the task priority formula further includes factors that take into account the scope of impact and potential hazards of the accident, which are automatically analyzed through the description of the urgency level and impact on traffic flow in the accident report.

[0025] Preferably, the failover capability of the message queue includes the function of automatically detecting failed nodes and reallocating data streams to healthy nodes, ensuring the continuity of data processing and high availability of the system.

[0026] The present invention provides an efficient method for automatically dispatching road obstacle clearing and rescue personnel. It has the following beneficial effects:

[0027] 1. The present invention combines a multi-factor decision-making scheduling algorithm with real-time traffic flow prediction. The system can dynamically adjust the dispatching route of rescue vehicles to prevent rescue vehicles from entering congested sections, thereby significantly reducing the time to arrive at the accident scene and improving the overall rescue efficiency.

[0028] 2. The system of the present invention can dynamically adjust the allocation of rescue personnel and equipment according to the urgency of the accident, the availability of resources and the forecast of traffic flow, so as to ensure that limited rescue resources can be put into use in the most reasonable way and avoid waste of resources or delay in rescue.

[0029] 3. The present invention adopts a distributed database and message queue architecture to ensure that the system can still operate stably under high concurrency and large-scale rescue tasks, and has good scalability, can adapt to rescue tasks of different scales, and avoid system crashes due to excessive data traffic.

[0030] 4. By introducing a machine learning model, the present invention enables the system to automatically optimize the weight factors in the scheduling algorithm through historical data, continuously improve decision-making accuracy, reduce manual intervention, and make the system more intelligent and adaptive. Through asynchronous writing and message queue technology, the system avoids the bottleneck of traditional synchronous data processing, making the allocation and scheduling of rescue tasks smoother and more efficient, and able to cope with a large number of sudden task requests.

[0031] 5. The present invention combines real-time drone monitoring and traffic flow prediction. The system can accurately obtain detailed information on the accident scene and surrounding traffic conditions, and update dispatch decisions in real time, thereby improving the accuracy and timeliness of rescue dispatch, especially in complex traffic environments. By reasonably allocating rescue resources and optimizing dispatch routes, unnecessary waiting and repeated rescue can be avoided, reducing vehicle fuel consumption and labor costs, and overall reducing operating costs during the rescue process. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Embodiment 1:

[0035] Please refer to the attached Figure 1 The embodiment of the present invention provides an efficient method for automatically dispatching road obstacle removal and rescue personnel, comprising the following steps:

[0036] (a) deploying at least one message queue to cache rescue-related data;

[0037] (b) store and query rescue mission and resource status data through a distributed database system;

[0038] (c) Collect data from rescue vehicles and accident sites in real time and perform pre-processing;

[0039] (d) asynchronously process the collected data through message queues according to priority and task type;

[0040] (e) Utilize multi-factor decision-making algorithms to dynamically dispatch rescue resources.

[0041] In one embodiment, system architecture and deployment

[0042] Step 1: Message Queue Module

[0043] In this embodiment, the message queue module is deployed using a Kafka cluster to efficiently process real-time data reported from rescue vehicles and accident sites. The Kafka cluster is configured with at least three partitions and two replicas to ensure that when any single node fails, the data can still be recovered from other replica nodes. At the same time, Kafka's partitioning mechanism can evenly distribute the data load on different nodes to avoid the problem of overloading a single node.

[0044] During the specific operation, you first need to deploy the Kafka cluster, determine the IP address and port number of each node in the cluster, set parameters such as broker.id and log.dirs in the configuration file to ensure reliable storage and transmission of data, and enable the failover mechanism in the cluster to ensure that when a node fails, the remaining nodes can automatically take over the data flow to achieve high availability.

[0045] The creation of a Kafka message queue is divided into two main steps: topic creation and message producer / consumer configuration. The number of partitions of a topic determines the ability of parallel processing, and the number of replicas ensures data redundancy and recoverability. An appropriate batch sending strategy is set on the producer side to improve transmission efficiency.

[0046] Step 2: Database cluster module

[0047] In this embodiment, the database cluster module uses the Cassandra database. The distributed architecture of Cassandra can ensure high availability and scalability of data storage. In the system, each data record (such as the location of the rescue vehicle, the urgency of the task, etc.) is stored with a timestamp and a geographic location as a composite primary key to facilitate fast query and task response.

[0048] The keyspace design of Cassandra requires that the primary key fields of each data table be clearly defined. The timestamp and geographic location fields need to be combined in order as part of the primary key to achieve query optimization. For example, when storing the real-time location data of rescue vehicles, they are first partitioned according to the geographic location of the vehicle, and then arranged in chronological order to quickly obtain the latest vehicle information.

[0049] In Cassandra, in order to ensure data consistency and recoverability, each piece of data is automatically replicated to multiple nodes to ensure that data will not be lost even if some nodes fail.

[0050] The message queue adopts a Kafka cluster configured with at least three partitions and at least two replicas to ensure the continuous availability and processing of data when a single server or partition fails. The distributed database system adopts Cassandra, and the primary key in its data model design includes a combination of timestamp and geographic location to optimize the deployment of rescue resources and the query efficiency of response time. The data preprocessing step further includes using regular expressions to verify and format the format of GPS data collected from sensors.

[0051] In one embodiment, data collection and preprocessing

[0052] Step 3: Data Collection

[0053] In this embodiment, the real-time data collection of rescue vehicles and accident sites is completed through a variety of sensors. Each rescue vehicle is equipped with GPS, speed sensors and video monitoring equipment to collect the precise location, real-time speed and road conditions of the vehicle. At the same time, surveillance cameras are also deployed at key traffic nodes to collect videos of the accident scene and surrounding traffic flow data.

[0054] In each rescue vehicle, the GPS sensor sends its current location at fixed time intervals, the speed sensor synchronously collects the current vehicle's moving speed, and the video surveillance regularly uploads road image data. All of this data is sent to the message queue through the wireless communication network for processing.

[0055] Step 4: Data Preprocessing

[0056] In this embodiment, the data preprocessing module verifies and cleans the received GPS data. First, the GPS data is formatted using regular expressions to ensure that the data format is correct and to filter out non-compliant data. For noisy data (such as abnormal vehicle position jumps), the system automatically eliminates possible erroneous information by comparing the displacement of the two previous and next time points.

[0057] During the preprocessing process, the system will build a dynamic data model based on the historical data of each rescue vehicle to analyze the normal driving trajectory of the vehicle. If it is found that the uploaded data deviates from the normal trajectory by more than a certain range, the system will mark the data as abnormal and trigger the correction mechanism.

[0058] The asynchronous processing step further includes using a priority queue algorithm to prioritize the collected data, giving priority to rescue tasks with high accident severity and closest distance, and the step (e) also includes monitoring and optimizing resource allocation and task execution efficiency, wherein the multi-factor decision algorithm determines the task priority based on the following formula:

[0059]

[0060] Among them, P(i) represents the priority of task i, U(i) and R(i) represent the urgency, data size and response time of the task respectively, and w1, w2, w3 are weight factors.

[0061] The weight factors w1, w2, w3 of the multi-factor decision-making algorithm are determined based on statistical analysis of historical data, and the optimal value of each factor is determined through machine learning model training to adapt to different traffic and accident situations, wherein the training step includes:

[0062] (a) Collect data on historical rescue missions;

[0063] (b) Evaluate the rescue efficiency under different weight combinations;

[0064] (c) Gradient descent method is used to optimize the weight factors to minimize the rescue response time.

[0065] The resource allocation and performance monitoring step further includes implementing a feedback-based dynamic resource allocation system, which adjusts the rescue resource allocation according to the real-time road conditions and the real-time location of the rescue vehicle; the task priority formula further includes factors that consider the scope of the accident and potential hazards, which are automatically analyzed through the urgency level in the accident report and the description of the impact on traffic flow; the failover capability of the message queue includes the function of automatically detecting faulty nodes and reallocating data flows to healthy nodes, ensuring the continuity of data processing and the high availability of the system.

[0066] In one embodiment, step 5: asynchronous writing mechanism

[0067] In this embodiment, the asynchronous write mechanism realizes non-blocking transmission of data streams through Kafka. After the data is collected, it is first sent to the Kafka message queue through the producer interface. The message queue is responsible for caching this data in different partitions so that subsequent consumers can obtain it on demand.

[0068] When the system load is high, Kafka's asynchronous writing mechanism can effectively distribute a large amount of data streams to different nodes smoothly, avoiding the performance bottleneck when the database directly handles large batch requests. During the asynchronous writing process, the system also detects the health status of each message queue node through the heartbeat mechanism, and automatically switches to other nodes when a fault is detected.

[0069] Step 6: Priority Classification:

[0070] In this embodiment, after the data enters the message queue, it will be prioritized according to factors such as the urgency of the task and the severity of the accident. The priority classification algorithm will be based on the above formula Assign a weight to each task. Tasks with higher data priority will be scheduled to the next processing stage. In the scheduling system, tasks are first evaluated for urgency by the algorithm and sorted from high to low according to urgency. Accidents with high urgency and close locations will be processed first, while data with relatively low urgency will enter the secondary priority queue to ensure the rational use of resources.

[0071] Multi-factor decision-making scheduling algorithm

[0072] Step 7: Scheduling priority calculation

[0073] In this embodiment, scheduling priority calculation is the core part of the scheduling system. The scheduling algorithm calculates the priority of each task based on the three main parameters of urgency U(i), data volume D(i) and response time R(i). The weight factors w1, w2, and w3 are obtained through historical data and machine learning model training.

[0074] In order to improve the flexibility and accuracy of scheduling, this embodiment trains a machine learning model, uses the rescue time and success rate of historical tasks as training data, and optimizes each weight factor. The specific training steps include:

[0075] Collect historical mission data, including the response time of each mission and the use of rescue resources;

[0076] Compare different weight combinations and analyze their impact on response time;

[0077] The weight factor is optimized by the gradient descent algorithm to converge to the optimal solution.

[0078] The system regularly updates the weight factors to ensure that the dispatch strategy is always self-optimizing based on current road and accident conditions.

[0079] Monitoring and resource optimization

[0080] Step 8: Real-time monitoring and feedback adjustment

[0081] In this embodiment, the system tracks the location of the rescue vehicle and the task execution status in real time through the monitoring module. The monitoring module detects and adjusts resource allocation in a timely manner by comparing the actual task execution progress with the planned progress. During the rescue mission, if the progress of a rescue vehicle is delayed, the system will recalculate the optimal path or dispatch other idle rescue resources to take over the task.

[0082] The monitoring data is regularly reported to the dispatch center, which adjusts the task priority and resource allocation strategy based on the feedback information. For example, if the traffic conditions at an accident site are more complicated than expected, the system will automatically increase its priority and dispatch more resources to handle it.

[0083] Embodiment 2:

[0084] 1. Introduction of traffic flow prediction model

[0085] On the basis of the present invention, a traffic flow prediction module is added, which uses historical traffic data and real-time collected traffic status data to predict the traffic flow in the next 10 to 30 minutes through a long short-term memory network (LSTM) model.

[0086] 1.1 Construction and training of LSTM model

[0087] Input Data: The system obtains traffic data from multiple sources, including:

[0088] Real-time traffic flow information (number of vehicles, speed, etc.) provided by road sensors.

[0089] Real-time images from ground surveillance cameras are combined with image processing algorithms to analyze traffic density.

[0090] External data such as weather and time serve as supplementary parameters of the prediction model.

[0091] The time series of these input data are represented as:

[0092] x={x1,x 2, ...,x n}

[0093] Among them, x i Represents the traffic flow data at the i-th time point.

[0094] Output of LSTM model: LSTM model predicts the traffic flow at m future time points by learning the time correlation of input data. The prediction result of the model can be expressed as:

[0095] Y={y1,y 2, ...,y m}

[0096] Among them, y j is the predicted flow value at the jth time point in the future

[0097] 1.2 Application of forecast data

[0098] The prediction results will provide predictive data support for rescue dispatch. For example, in sections of roads where traffic congestion is expected to occur, the system will dynamically adjust the dispatch routes of rescue vehicles to avoid entering high-congestion areas and reduce response time.

[0099] 2. Combination of traffic flow prediction and scheduling algorithm

[0100] Combined with the aforementioned multi-factor decision-making dispatch algorithm, traffic flow forecast data will become a new variable in the dispatch system and participate in the calculation of decision priorities. The system dynamically adjusts the dispatch of rescue resources by adding weight factors related to traffic flow.

[0101] 2.1 Update of multi-factor scheduling algorithm

[0102] After combining traffic flow forecast data, the dispatch priority formula is improved as follows:

[0103]

[0104] Among them, T(i) is a dynamic parameter based on the traffic flow prediction results. This value is calculated by the future traffic flow of the road section predicted by the LSTM model. If the traffic prediction results show that the path taken by the rescue vehicle will be highly congested in the future, the T(i) value will increase, thereby reducing the priority of this path.

[0105] 2.2 Calculation of congestion coefficient

[0106] Traffic flow prediction result Y = {y1,y 2, ...,y m} for each y j represents the traffic flow at time j in the future. The system evaluates the congestion coefficient based on the traffic level of each path. Assuming that the standard vehicle throughput of a path is C, the congestion coefficient T(i) of each path can be calculated by the following formula:

[0107]

[0108] in:

[0109] yj represents the predicted traffic flow at time point j.

[0110] C represents the standard throughput of the path, that is, the number of vehicles that can pass per unit time in the absence of congestion.

[0111] T(i) represents the congestion coefficient of the path. The larger the value, the more congested the path is.

[0112] The system dynamically adjusts the scheduling priority based on this congestion coefficient. If T(i) is higher than a certain threshold, it indicates that the path is extremely congested, and the scheduling system will select other paths or resources for rescue.

[0113] 3. Real-time adjustment of dispatching paths

[0114] When a route is predicted to experience severe traffic congestion while a rescue vehicle is traveling, the system adjusts the vehicle's travel path based on the current dispatch priority and new forecast data.

[0115] The system calls the API of the map service (such as the Google Maps API) to recalculate the optimal route from the current rescue vehicle location to the accident site, taking into account the current traffic conditions and predicting future traffic changes.

[0116] If congestion is detected on the original route, the system immediately calculates an alternative route and updates the route guidance in real time through the rescue vehicle's navigation system.

[0117] 4. Feedback and model optimization

[0118] To ensure that the scheduling algorithm maintains the best effect over a long period of time, the system will regularly update and optimize the value of the weight factor w4 through historical data. The specific steps are as follows:

[0119] Collect traffic flow prediction data from historical scheduling results and compare them with actual traffic flow to evaluate the accuracy of the prediction model.

[0120] Through the feedback mechanism in the machine learning model, the parameters of the LSTM model are updated to optimize the accuracy of traffic flow prediction.

[0121] At the same time, according to the actual response time and scheduling efficiency of the rescue mission, the weight factors w1, w2, and w3 in the priority formula are adjusted to adapt to the latest scheduling requirements.

[0122] 5. Combined system workflow

[0123] The system workflow after combining traffic flow prediction is as follows:

[0124] Real-time data collection: Collect road traffic data through sensors and monitoring equipment, and input it into the LSTM model for prediction.

[0125] Traffic flow prediction: The model outputs the traffic flow in the next 10-30 minutes and generates the congestion coefficient T

[0126] (i).

[0127] Priority calculation: Based on the traffic flow prediction results, combined with the accident urgency, task data volume and response time, the scheduling priority of the task is calculated.

[0128] Path optimization: If the prediction results show that the rescue path will be congested, the system will update the rescue vehicle's route in real time.

[0129] Feedback optimization: The system updates model parameters based on the actual execution of tasks to ensure that the scheduling system remains in the optimal state.

[0130] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An efficient method for automatically dispatching road obstacle removal and rescue personnel, characterized in that: The following steps are involved: (a) deploying at least one message queue to cache rescue-related data; (b) store and query rescue mission and resource status data through a distributed database system; (c) Collect data from rescue vehicles and accident sites in real time and perform pre-processing; (d) asynchronously process the collected data through a message queue according to priority and task type; (e) Utilize multi-factor decision-making algorithms to dynamically dispatch rescue resources.

2. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The message queue uses a Kafka cluster configured with at least three partitions and at least two replicas to ensure the continued availability and processing of data is not affected when a single server or partition fails.

3. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The distributed database system adopts Cassandra, and the primary key in its data model design includes a combination of timestamp and geographic location to optimize the deployment of rescue resources and the query efficiency of response time.

4. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The data preprocessing step further includes using regular expressions to validate and format the GPS data format collected from the sensor.

5. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The asynchronous processing step further includes using a priority queue algorithm to prioritize the collected data, giving priority to rescue missions with high accident severity and the closest distance.

6. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The step (e) further includes monitoring and optimizing resource allocation and task execution efficiency, wherein the multi-factor decision algorithm determines the task priority based on the following formula: Among them, P(i) represents the priority of task i, U(i) and R(i) represent the urgency, data size and response time of the task respectively, and w1, w2, w3 are weight factors.

7. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The weight factors w1, w2, and w3 of the multi-factor decision-making algorithm are determined based on statistical analysis of historical data, and the optimal value of each factor is determined through machine learning model training to adapt to different traffic and accident situations. The training steps include: (a) Collect data on historical rescue missions; (b) Evaluate the rescue efficiency under different weight combinations; (c) Gradient descent method is used to optimize the weight factors to minimize the rescue response time.

8. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The resource allocation and performance monitoring step further includes implementing a feedback-based dynamic resource allocation system, which adjusts rescue resource allocation according to real-time road conditions and real-time locations of rescue vehicles.

9. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The task priority formula further includes factors that consider the scope of the accident and potential hazards, which are automatically analyzed through the urgency level and description of the impact on traffic flow in the accident report.

10. The method for automatically dispatching efficient road obstacle removal and rescue personnel according to claim 1, characterized in that: The failover capability of the message queue includes the function of automatically detecting failed nodes and redistributing data streams to healthy nodes, ensuring the continuity of data processing and high availability of the system.

Images