CN117519944A - An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration - Google Patents

An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration Download PDF

Info

Publication number
CN117519944A
CN117519944A CN202311649495.6A CN202311649495A CN117519944A CN 117519944 A CN117519944 A CN 117519944A CN 202311649495 A CN202311649495 A CN 202311649495A CN 117519944 A CN117519944 A CN 117519944A
Authority
CN
China
Prior art keywords
module
vehicle
computing
computing power
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202311649495.6A
Other languages
Chinese (zh)
Inventor
吴昊
李哲瀚
马帅
黄锐
张恒鑫
孙翔宇
胡煜敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Luo Jiaxi
Original Assignee
Luo Jiaxi
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Luo Jiaxi filed Critical Luo Jiaxi
Priority to CN202311649495.6A priority Critical patent/CN117519944A/en
Publication of CN117519944A publication Critical patent/CN117519944A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/48—Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806—Task transfer initiation or dispatching
    • G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
    • G06F11/302—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a software system
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5072—Grid computing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Quality & Reliability (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention belongs to the technical field of unmanned, and discloses an unmanned vehicle based on cooperation of computing power perception and edge cloud computing. According to the invention, through ingenious combination of the calculation force sensing unit and the edge cloud computing unit, idle expert knowledge data in the prior art are fully utilized, and the available deep learning model is generated by combining the deep neural network learning technology, so that the train driving gear control sequence is quickly and accurately obtained in real time to realize unmanned control of the vehicle, and when an interference abnormality occurs in the running process of the vehicle, the time-variable force sensing data module, the first-level calculation force sensing data module and the global calculation force sensing data module can be used for optimizing the detected algorithm model again well, so that the running safety of the vehicle is improved.

Description

Unmanned vehicle based on cooperation of computing power perception and edge cloud computing and cooperation method thereof
Technical Field
The invention relates to the technical field of unmanned aerial vehicles, in particular to an unmanned aerial vehicle based on cooperation of computing power perception and edge cloud computing and a cooperation method thereof.
Background
Along with the automatic driving entering rational period mainly based on single car intelligence and the great investment of new national infrastructure, the cooperation of the roads in China starts entering the hot period;
the unmanned vehicle is cooperated with the global positioning system by means of artificial intelligence, visual computing, radar, a monitoring device, so that the computer can automatically and safely operate the motor vehicle without any active operation of human beings; existing unmanned vehicles and unmanned vehicle communication systems have limitations in terms of real-time, reliability and efficiency.
Disclosure of Invention
In order to solve the problems, the invention provides an unmanned vehicle based on cooperation of computing power perception and edge cloud computing, which comprises a central cloud computing center, a vehicle-mounted system, a computing power perception unit, an edge cloud computing unit and an unmanned system;
the central cloud computing center comprises a database module, a deep learning module, an evaluation module and a central communication module;
the vehicle-mounted system comprises a vehicle flow detection system, a road side system and an intelligent traffic system for realizing vehicle and road information interaction and sharing through wireless communication equipment; the traffic flow detection system and the road side system respectively collect traffic flow information and road side information into the intelligent traffic system;
the computing force sensing unit comprises a time-varying force sensing data module, a first-level computing force sensing data module and a global computing force sensing data module;
the edge cloud computing unit comprises a driving mode control module, a deep learning model module, an intelligent algorithm module and an edge communication module;
the unmanned system comprises a cluster control system, an unmanned aerial vehicle/vehicle cluster system and a satellite positioning system;
the central cloud computing center: for storing, analyzing and transmitting signals of data;
the vehicle-mounted system comprises: the device is used for detecting the traffic flow, the speed, the distance between vehicles and the navigation of vehicles;
the computing power sensing unit: the power computing node is used for monitoring power computing nodes in the power computing network, and changing power computing resource time-varying data in the power computing sensing data module during updating;
the edge cloud computing unit: the system is used for controlling the driving module according to the data obtained from the road information, analyzing the road information and transmitting the analyzed information to the vehicle-mounted system, so that the unmanned vehicle can be driven intelligently.
Preferably, the traffic flow detection system is used for detecting the traffic flow condition of the road, the road side system is used for detecting the condition of the road, the wireless communication equipment transmits the traffic flow condition detected by the traffic flow detection system and the road side system to the intelligent traffic system, and the intelligent traffic system stores and analyzes the signals and then transmits the signals to the central cloud computing center.
Preferably, the time-varying data monitoring module is used for monitoring the detected road information nodes, updating the time-varying data of the computing power resources in the time-varying power perception data module and transmitting the updated data to the central cloud computing center in time;
the first-level computing power updating module is used for updating the first-level computing power perception attribute data in the first-level computing power perception data module based on the computing power resource time-varying data;
and the global computing power updating module is used for generating the computing power label of the computing power node in the global computing power perception data module based on the primary computing power perception attribute data.
Preferably, the driving mode control module can process the received state data, identify whether the running environment state of the corresponding sub-line is normal, if the running state of the line is abnormal, the unmanned vehicle stops running, the deep learning model module is triggered to work according to the identification result of information acquisition,
the deep learning model module can process and analyze received line and vehicle basic data by using an available deep learning model, store analysis results in a data storage module in the central cloud computing center, and simultaneously transmit the processing and analysis results to a driving gear control sequence and transmit driving gear control sequence data to the edge communication module;
the intelligent algorithm module can timely analyze and process the route of the vehicle when the vehicle runs by utilizing a bionic intelligent algorithm, and transmits the analyzed and processed driving curve data to the edge communication module.
Preferably, the cluster control system of the cluster control system can be deployed at a cloud end to control the unmanned aerial vehicle/vehicle cluster, is responsible for task allocation of the whole unmanned aerial vehicle/vehicle cluster, integrates feedback information of the unmanned aerial vehicle/vehicle cluster, thereby realizing functions of system decision, cluster task adjustment and the like, and can also finish tasks with higher calculation power resource requirements, and further issue calculation results or execution tasks to unmanned aerial vehicle/vehicle onboard equipment;
the control unmanned aerial vehicle/vehicle cluster acquires environmental information for an unmanned system through a sensor, an embedded unit of an airborne computing device performs data analysis, performs data fusion calculation of multiple sensors based on data types, completes the processing of sensing signals, executes instructions such as control and the like, can complete execution tasks with larger computing resource requirements such as collaborative path planning, collaborative obstacle avoidance, collaborative patrol and the like, and comprises a communication module, a control module, a sensing module, a positioning module and the like;
the communication module is responsible for local communication between unmanned aerial vehicles/vehicles and remote communication between the unmanned aerial vehicles/vehicles and the cluster control system, and has the main functions of receiving task instructions of the cluster control system, sending data to the cluster control system and acquiring other unmanned aerial vehicle/vehicle information, sending the task instructions or other unmanned aerial vehicle/vehicle information to the control module after the communication module acquires the task instructions or other unmanned aerial vehicle/vehicle information, further processing the task instructions or other unmanned aerial vehicle/vehicle information by an onboard computer, executing or providing information by a driving algorithm and the like; the communication module is favorable for carrying out external connection on 4G/5G operator networks between the edge unmanned aerial vehicle/unmanned aerial vehicle and the cloud platform, and mainly transmits one-dimensional data collected by the edge sensor, such as altitude, GPS longitude and latitude, motor running speed, IMU gyroscope motion state and the like; two-dimensional data acquired by infrared, depth cameras and laser radars; the method comprises the steps that position parameters of different edge devices are received through a cloud platform between a plurality of unmanned vehicles and the unmanned vehicles, and respective states and position information are interacted;
the control module comprises equipment such as an onboard computer, a flight control device and the like, and is used for completing specified tasks and ensuring normal flight and running of the unmanned aerial vehicle/vehicle;
the sensing module mainly comprises an airborne sensor, a camera can acquire image information, data is provided for an image processing algorithm, the image processing algorithm is a main source of environment information, a laser radar obtains 3D point cloud, distance information of objects in the environment can be acquired, and data is provided for 3D modeling, path planning and the like;
the positioning module receives positioning information of a satellite system, obtains position coordinates of the unmanned aerial vehicle/vehicle after calculation, and the like, and is a precondition for completing a cluster task;
the satellite positioning system: positioning information is provided for the drone/vehicle.
A cooperation method of an unmanned vehicle based on cooperation of computing power perception and edge cloud computing comprises the following steps:
s1, firstly, registering available unmanned aerial vehicles and available computing resources on unmanned aerial vehicles into a computing pool, and recording related information such as processing capacity, communication characteristics and the like; this may provide basic data for subsequent scheduling and allocation;
(1) Computing power resource registration
When the unmanned aerial vehicle or the unmanned aerial vehicle enters a cooperative working environment, the unmanned aerial vehicle or the unmanned aerial vehicle can register available computing resource information to a computing pool; registration includes, but is not limited to, processor type, processor core number, memory capacity, storage capacity, etc.;
(2) Computing power resource management
The management of the computing power resources involves distributing available computing power resources to unmanned aerial vehicles or unmanned aerial vehicles needing to execute tasks, and can use a priority scheduling algorithm to conduct resource management, distribute different priorities for different types of tasks, and distribute the tasks to corresponding unmanned aerial vehicles or unmanned aerial vehicles according to the emergency degree and importance of the tasks;
s2, task demand acquisition can be completed through communication with a task initiator, the task initiator can provide information such as description, priority, time limit and the like of tasks for a collaborative system and send the information to a computing pool, the task demand acquisition can also be realized through sensor data collection and processing, and an unmanned plane and an unmanned vehicle can sense environments and extract task demands by using various sensors (such as cameras, radars, laser scanners and the like);
based on a task scheduling algorithm of priority, different priorities are allocated to different types of tasks, and the execution sequence of the tasks is determined according to the emergency degree and importance of the tasks; prioritization may be implemented using a priority queue or scheduling rule;
s3, the core of the power network resource scheduling system is a resource scheduling strategy. The system provides 4 computing power resource scheduling strategies according to different optimization indexes: cost-aware scheduling policies, load-aware scheduling policies, energy-efficiency-aware scheduling policies, and Service Layer Agreement (SLA) -aware scheduling policies. The 4 scheduling strategies can be used in a modularized combination mode, so that multi-objective optimization is realized. The application to be deployed is marked as S0, the application type label is marked as L, and the applications which are deployed and have the same label are marked as a set S= { S1, S2, …, … Sn }, in the order from near to far in time;
an initialization module: the method mainly uses a service application label matching method;
an application demand matching module: performing similarity analysis of computing power resource requirements and network resource requirements;
history policy matching module: if the similar application with S0 is not found after the label matching is carried out on all the applications in S, a history strategy matching module is used for scheduling; (the task to be deployed is scheduled with a historically identical or similar task alignment)
A resource scheduling and deploying module: after the micro-service matching screening process is finished, if the micro-service with higher similarity is found, deploying according to a resource scheduling strategy used by the matched micro-service, otherwise, randomly selecting one resource scheduling strategy to deploy;
and S4, selecting computing power resources with proper conditions from the computing power pool to execute the task according to the scheduling strategy. The method comprises the steps of carrying out resource allocation by considering the factors such as availability of resources, performance requirements, task priority and the like;
s5, monitoring the use condition of resources and the progress of the task in real time in the task execution process; the resource allocation or the task reallocation is timely adjusted by monitoring indexes such as the load of the resource and the execution time of the task, so that the overall execution efficiency is optimized;
(1) State evaluation and feedback control
Evaluating the status of the task performer based on the sensor data and providing appropriate feedback control signals; the state evaluation can be performed based on a filter or other state estimation method to obtain accurate task executor state information;
(2) Anomaly detection and handling
Abnormal conditions in the task execution process, such as sensor faults, communication interruption, resource exhaustion and the like, are monitored in real time; an anomaly detection algorithm may be used to identify and handle anomalies;
(3) Completion evaluation
According to the task description or the standard, the completion degree of the task is evaluated, and corresponding indexes or reports are generated to monitor the execution condition of the task; the completion evaluation can be defined according to task requirements and conditions, such as time stamp, position detection, target achievement, and the like;
s6, if abnormal conditions such as resource faults, communication interruption or task failure occur, corresponding abnormal positions and fault recovery are needed; this may include reallocating resources, replacing failed components, or rescheduling tasks;
and S7, dynamically expanding or contracting resources in the computing pool according to the number of tasks and the change of complexity. When the task amount is increased, more unmanned aerial vehicles or unmanned vehicles can be added to increase the computing power resource; and when the tasks are reduced, resources can be reduced to save energy costs.
In summary, the present invention includes at least one of the following beneficial technical effects:
(1) The advantages of edge computing and cloud computing are combined, an efficient and multi-device interconnection communication networking mode is provided, and end-to-end task coordination is achieved; and the unmanned vehicles, the unmanned aerial vehicle and the cloud platform are interconnected and intercommunicated, and the simultaneous execution of multiple tasks is realized.
(2) By means of the computing power resource scheduling algorithm, intelligent allocation and optimization of tasks are achieved, perception of computing power and data is achieved, and overall performance and response speed of the system are improved.
(3) The edge cloud collaborative architecture is utilized to realize training of cloud deep learning tasks, and generation and reasoning of an edge lightweight model, so that edge sinking and rapid deployment of target detection and recognition tasks are realized.
Drawings
FIG. 1 is a schematic diagram of a power network regulation system of the present invention;
FIG. 2 is a schematic diagram of the dynamic resource expansion and contraction system of the present invention.
Detailed Description
The embodiment of the invention discloses an unmanned vehicle based on cooperation of computing power perception and edge cloud computing, which comprises a center cloud computing center, a vehicle-mounted system, a computing power perception unit, an edge cloud computing unit and an unmanned system;
the central cloud computing center comprises a database module, a deep learning module, an evaluation module and a central communication module;
the vehicle-mounted system comprises a vehicle flow detection system, a road side system and an intelligent traffic system for realizing vehicle and road information interaction and sharing through wireless communication equipment; the traffic flow detection system and the road side system respectively collect traffic flow information and road side information into the intelligent traffic system;
the computing force sensing unit comprises a time-varying force sensing data module, a first-level computing force sensing data module and a global computing force sensing data module;
the edge cloud computing unit comprises a driving mode control module, a deep learning model module, an intelligent algorithm module and an edge communication module;
1) Cloud deep learning pre-training model generation
The cloud end distributes corresponding pre-training models, such as YOLO, mobileNet, efficientNet and other networks, to the edges of different unmanned aerial vehicles and unmanned vehicles according to different scenes deployed by the unmanned aerial vehicles and different tasks, such as image segmentation, suspicious object recognition, image classification, visual navigation and other tasks, calculation power (tips) of different edge nodes, network quality states (Mbps, ping) and other modes of model weight (such as network pruning, knowledge distillation and model compression), the cloud end is connected with a data lake construction, an initialized pre-training model is constructed through non-real-time data and long-period data, the data lake recovers non-high privacy transmitted by the edge end, positive feedback exists on model accuracy, edge data with high image quality are obtained, finally, model updating and weight reduction are carried out on the basis of the pre-training model according to the deployment scene of the edge end and the fed back local data, and the accuracy after model optimization is detected,
2) Edge model reasoning and data collaboration
The edge end simply classifies the locally collected image data according to night, daytime, near view and distant view; firstly, for a real-time task, identifying through detection precision (accuracy), and feeding back an identification result; meanwhile, data are saved locally, detection accuracy is inverted, images lower than a threshold value of 70 minutes are primarily screened according to an image quality evaluation algorithm, a small number of high-privacy pictures are trained locally according to the privacy class of the images, and the rest pictures are uploaded to a cloud when network resources are idle (PRB utilization rate is less than 20 percent); in addition, when the cloud end completes updating and light weight of the pre-training model, the edge feeds back according to the local computing force tips, the size (MLOPS) of the updated model, the model reasoning speed (FPS > 20) and the accuracy (mAP) until the computing force, the model size, the reasoning speed and the accuracy meet the requirements of the edge end, and the model is finally issued to the edge end to complete model updating.
The unmanned system comprises a cluster control system, an unmanned aerial vehicle/vehicle cluster system and a satellite positioning system;
the central cloud computing center: for storing, analyzing and transmitting signals of data;
the vehicle-mounted system comprises: the device is used for detecting the traffic flow, the speed, the distance between vehicles and the navigation of vehicles;
the computing power sensing unit: the power computing node is used for monitoring power computing nodes in the power computing network, and changing power computing resource time-varying data in the power computing sensing data module during updating;
the edge cloud computing unit: the system is used for controlling the driving module according to the data obtained from the road information, analyzing the road information and transmitting the analyzed information to the vehicle-mounted system, so that the unmanned vehicle can be driven intelligently.
Further, the traffic flow detection system is used for detecting the traffic flow condition of the road, the road side system is used for detecting the condition of the road, the wireless communication equipment transmits the traffic flow condition detected by the traffic flow detection system and the road side system to the intelligent traffic system, and the intelligent traffic system stores and analyzes the signals and then transmits the signals to the central cloud computing center.
Further, the time-varying data monitoring module is used for monitoring the detected road information nodes, updating the time-varying data of the computing power resources in the time-varying power perception data module and timely transmitting the updated data to the central cloud computing center;
the first-level computing power updating module is used for updating the first-level computing power perception attribute data in the first-level computing power perception data module based on the computing power resource time-varying data;
and the global computing power updating module is used for generating the computing power label of the computing power node in the global computing power perception data module based on the primary computing power perception attribute data.
Furthermore, the driving mode control module can process the received state data, identify whether the running environment state of the corresponding sub-line is normal, if the running state of the line is abnormal, the unmanned vehicle stops running, the deep learning model module is triggered to work according to the identification result of information acquisition,
the deep learning model module can process and analyze received line and vehicle basic data by using an available deep learning model, store analysis results in a data storage module in the central cloud computing center, and simultaneously transmit the processing and analysis results to a driving gear control sequence and transmit driving gear control sequence data to the edge communication module;
the intelligent algorithm module can timely analyze and process the route of the vehicle when the vehicle runs by utilizing a bionic intelligent algorithm, and transmits the analyzed and processed driving curve data to the edge communication module.
Furthermore, the cluster control system can be deployed at a cloud end to control the unmanned aerial vehicle/vehicle cluster, is responsible for task allocation of the whole unmanned aerial vehicle/vehicle cluster, integrates feedback information of the unmanned aerial vehicle/vehicle cluster, thereby realizing the functions of system decision, cluster task adjustment and the like, and can also finish tasks with higher calculation force resource requirements, and further issue calculation results or execution tasks to unmanned aerial vehicle/vehicle onboard equipment;
the control unmanned aerial vehicle/vehicle cluster acquires environmental information for an unmanned system through a sensor, an embedded unit of an airborne computing device performs data analysis, performs data fusion calculation of multiple sensors based on data types, completes the processing of sensing signals, executes instructions such as control and the like, can complete execution tasks with larger computing resource requirements such as collaborative path planning, collaborative obstacle avoidance, collaborative patrol and the like, and comprises a communication module, a control module, a sensing module, a positioning module and the like;
the communication module is responsible for local communication between unmanned aerial vehicles/vehicles and remote communication between the unmanned aerial vehicles/vehicles and the cluster control system, and has the main functions of receiving task instructions of the cluster control system, sending data to the cluster control system and acquiring other unmanned aerial vehicle/vehicle information, sending the task instructions or other unmanned aerial vehicle/vehicle information to the control module after the communication module acquires the task instructions or other unmanned aerial vehicle/vehicle information, further processing the task instructions or other unmanned aerial vehicle/vehicle information by an onboard computer, executing or providing information by a driving algorithm and the like; the communication module is favorable for carrying out external connection on 4G/5G operator networks between the edge unmanned aerial vehicle/unmanned aerial vehicle and the cloud platform, and mainly transmits one-dimensional data collected by the edge sensor, such as altitude, GPS longitude and latitude, motor running speed, IMU gyroscope motion state and the like; two-dimensional data acquired by infrared, depth cameras and laser radars; the method comprises the steps that position parameters of different edge devices are received through a cloud platform between a plurality of unmanned vehicles and the unmanned vehicles, and respective states and position information are interacted;
the control module comprises equipment such as an onboard computer, a flight control device and the like, and is used for completing specified tasks and ensuring normal flight and running of the unmanned aerial vehicle/vehicle;
the sensing module mainly comprises an airborne sensor, a camera can acquire image information, data is provided for an image processing algorithm, the image processing algorithm is a main source of environment information, a laser radar obtains 3D point cloud, distance information of objects in the environment can be acquired, and data is provided for 3D modeling, path planning and the like;
the positioning module receives positioning information of a satellite system, obtains position coordinates of the unmanned aerial vehicle/vehicle after calculation, and the like, and is a precondition for completing a cluster task;
the satellite positioning system: positioning information is provided for the drone/vehicle.
A cooperation method of an unmanned vehicle based on cooperation of computing power perception and edge cloud computing comprises the following steps:
s1, firstly, registering available unmanned aerial vehicles and available computing resources on unmanned aerial vehicles into a computing pool, and recording related information such as processing capacity, communication characteristics and the like; this may provide basic data for subsequent scheduling and allocation;
(1) Computing power resource registration
When the unmanned aerial vehicle or the unmanned aerial vehicle enters a cooperative working environment, the unmanned aerial vehicle or the unmanned aerial vehicle can register available computing resource information to a computing pool; registration includes, but is not limited to, processor type, processor core number, memory capacity, storage capacity, etc.;
(2) Computing power resource management
The management of the computing power resources involves distributing available computing power resources to unmanned aerial vehicles or unmanned aerial vehicles needing to execute tasks, and can use a priority scheduling algorithm to conduct resource management, distribute different priorities for different types of tasks, and distribute the tasks to corresponding unmanned aerial vehicles or unmanned aerial vehicles according to the emergency degree and importance of the tasks;
s2, task demand acquisition can be completed through communication with a task initiator, the task initiator can provide information such as description, priority, time limit and the like of tasks for a collaborative system and send the information to a computing pool, the task demand acquisition can also be realized through sensor data collection and processing, and an unmanned plane and an unmanned vehicle can sense environments and extract task demands by using various sensors (such as cameras, radars, laser scanners and the like);
based on a task scheduling algorithm of priority, different priorities are allocated to different types of tasks, and the execution sequence of the tasks is determined according to the emergency degree and importance of the tasks; prioritization may be implemented using a priority queue or scheduling rule;
s3, the core of the power network resource scheduling system is a resource scheduling strategy. The system provides 4 computing power resource scheduling strategies according to different optimization indexes: cost-aware scheduling policies, load-aware scheduling policies, energy-efficiency-aware scheduling policies, and Service Layer Agreement (SLA) -aware scheduling policies. The 4 scheduling strategies can be used in a modularized combination mode, so that multi-objective optimization is realized. The application to be deployed is marked as S0, the application type label is marked as L, and the applications which are deployed and have the same label are marked as a set S= { S1, S2, …, … Sn }, in the order from near to far in time;
an initialization module: the method mainly uses a service application label matching method;
an application demand matching module: performing similarity analysis of computing power resource requirements and network resource requirements;
history policy matching module: if the similar application with S0 is not found after the label matching is carried out on all the applications in S, a history strategy matching module is used for scheduling; (the task to be deployed is scheduled with a historically identical or similar task alignment)
A resource scheduling and deploying module: after the micro-service matching screening process is finished, if the micro-service with higher similarity is found, deploying according to a resource scheduling strategy used by the matched micro-service, otherwise, randomly selecting one resource scheduling strategy to deploy;
and S4, selecting computing power resources with proper conditions from the computing power pool to execute the task according to the scheduling strategy. The method comprises the steps of carrying out resource allocation by considering the factors such as availability of resources, performance requirements, task priority and the like;
s5, monitoring the use condition of resources and the progress of the task in real time in the task execution process; the resource allocation or the task reallocation is timely adjusted by monitoring indexes such as the load of the resource and the execution time of the task, so that the overall execution efficiency is optimized;
(1) State evaluation and feedback control
Based on the sensor data, the status of the task performer is assessed and appropriate feedback control signals are provided. The state evaluation can be performed based on a filter or other state estimation method to obtain accurate task executor state information;
(2) Anomaly detection and handling
Abnormal conditions in the task execution process, such as sensor faults, communication interruption, resource exhaustion and the like, are monitored in real time; an anomaly detection algorithm may be used to identify and handle anomalies;
(3) Completion evaluation
According to the task description or the standard, the completion degree of the task is evaluated, and corresponding indexes or reports are generated to monitor the execution condition of the task; the completion evaluation can be defined according to task requirements and conditions, such as time stamp, position detection, target achievement, and the like;
s6, if abnormal conditions such as resource faults, communication interruption or task failure occur, corresponding abnormal positions and fault recovery are needed; this may include reallocating resources, replacing failed components, or rescheduling tasks;
and S7, dynamically expanding or contracting resources in the computing pool according to the number of tasks and the change of complexity. When the task amount is increased, more unmanned aerial vehicles or unmanned vehicles can be added to increase the computing power resource; and when the tasks are reduced, resources can be reduced to save energy costs.
The unmanned vehicle and the unmanned vehicle communication system have the following advantages:
(1) Real-time performance: the edge cloud computing is utilized to transfer the processing task to the cloud, so that image recognition and data transmission can be performed at a higher speed;
(2) Reliability: adopting a high-speed low-delay communication protocol and algorithm to ensure stable transmission and accurate analysis of data;
(3) Efficiency is that: and the available computing power resources are dynamically allocated and managed through a computing power resource scheduling algorithm, so that the system operation efficiency and the resource utilization rate are improved to the maximum extent.
The intelligent traffic and unmanned aerial vehicle application development can be promoted in the fields of logistics, aviation, agriculture and the like. Through integrating the edge cloud computing, image recognition and computing power resource scheduling technology, the performance of the unmanned vehicle and unmanned vehicle communication system can be improved, and more innovation and business opportunities are brought to related industries.
The above embodiments are not intended to limit the scope of the present invention, so: all equivalent changes in structure, shape and principle of the invention should be covered in the scope of protection of the invention.

Claims (6)

1.一种基于算力感知和边缘云计算协同的无人车，其特征在于，所述无人车包括中心云计算中心、车载系统、算力感知单元、边缘云计算单元和无人系统；1. An unmanned vehicle based on computing power sensing and edge cloud computing collaboration, characterized in that the unmanned vehicle includes a central cloud computing center, a vehicle-mounted system, a computing power sensing unit, an edge cloud computing unit and an unmanned system; 所述中心云计算中心包括数据库模块、深度学习模块、评估模块和中心通信模块；The central cloud computing center includes a database module, a deep learning module, an evaluation module and a central communication module; 所述车载系统包括车流量检测系统、路侧系统和通过无线通讯设备实现车、路信息交互和共享的智能交通系统；所述车流量检测系统和所述路侧系统分别将车流量信息和路侧信息汇总至所述智能交通系统中；The vehicle-mounted system includes a traffic flow detection system, a roadside system and an intelligent transportation system that realizes vehicle and road information interaction and sharing through wireless communication equipment; the traffic flow detection system and the roadside system respectively combine traffic flow information and roadside information. The side information is aggregated into the intelligent transportation system; 所述算力感知单元包括时变算力感知数据模块、一级算力感知数据模块和全局算力感知数据模块；The computing power sensing unit includes a time-varying computing power sensing data module, a first-level computing power sensing data module and a global computing power sensing data module; 所述边缘云计算单元包括驾驶模式控制模块、深度学习模型模块、智能算法模块和边缘通信模块；The edge cloud computing unit includes a driving mode control module, a deep learning model module, an intelligent algorithm module and an edge communication module; 所述无人系统包括集群控制系统、无人机/车集群系统及卫星定位系统；The unmanned system includes a cluster control system, a drone/vehicle cluster system and a satellite positioning system; 所述中心云计算中心：用于对数据进行存储、分析和信号的传输；The central cloud computing center: used for data storage, analysis and signal transmission; 所述车载系统：用于车流量检测、车速检测、车距检测和车辆导航；The vehicle-mounted system: used for traffic flow detection, vehicle speed detection, vehicle distance detection and vehicle navigation; 所述算力感知单元：用于对算力网络中的算力节点进行监控，更新时变算力感知数据模块中的算力资源时变数据；The computing power sensing unit is used to monitor the computing power nodes in the computing power network and update the time-varying data of computing power resources in the time-varying computing power sensing data module; 所述边缘云计算单元：用于对道路信息得到的数据进而进行驾驶模块的控制，对道路信息的分析以及对分析好的信息传输到车载系统中，进而使得无人车智能驾驶。The edge cloud computing unit is used to control the driving module based on the data obtained from the road information, analyze the road information, and transmit the analyzed information to the vehicle system, thereby enabling the autonomous vehicle to drive intelligently. 2.根据权利要求1所述的一种基于算力感知和边缘云计算协同的无人车，其特征在于，所述车流量检测系统用于检测道路车流量情况，所述路侧系统用于对道路的情况进行检测，而无线通讯设备将车流量检测系统检测到的车流量情况和路侧系统对道路检测的情况进行传输到智能交通系统中，智能交通系统将信号进行存储和分析，然后传输到中心云计算中心中。2. An unmanned vehicle based on computing power perception and edge cloud computing collaboration according to claim 1, characterized in that the traffic flow detection system is used to detect road traffic flow conditions, and the roadside system is used to detect road traffic flow conditions. The conditions of the road are detected, and the wireless communication equipment transmits the traffic flow conditions detected by the traffic flow detection system and the road detection conditions by the roadside system to the intelligent transportation system. The intelligent transportation system stores and analyzes the signals, and then transmitted to the central cloud computing center. 3.根据权利要求1所述的一种基于算力感知和边缘云计算协同的无人车，其特征在于，所述时变数据监控模块，用于通过对检测到的道路信息节点进行监控，对时变算力感知数据模块中的算力资源时变数据进行更新，及时将更新的数据传输到中心云计算中心；3. An unmanned vehicle based on computing power perception and edge cloud computing collaboration according to claim 1, characterized in that the time-varying data monitoring module is used to monitor the detected road information nodes, Update the time-varying data of computing power resources in the time-varying computing power sensing data module, and transmit the updated data to the central cloud computing center in a timely manner; 一级算力更新模块，用于基于所述算力资源时变数据，更新一级算力感知数据模块中的一级算力感知属性数据；A first-level computing power update module, configured to update the first-level computing power sensing attribute data in the first-level computing power sensing data module based on the time-varying data of the computing power resources; 全局算力更新模块，用于基于所述一级算力感知属性数据，在全局算力感知数据模块中生成所述算力节点的算力标签。The global computing power update module is configured to generate the computing power label of the computing power node in the global computing power sensing data module based on the first-level computing power sensing attribute data. 4.根据权利要求1所述的一种基于算力感知和边缘云计算协同的无人车，其特征在于，所述驾驶模式控制模块能对收到的状态数据进行处理，并识别对应子线路的运行环境状态是否正常，如果线路运行状态不正常，就会使得无人车停止运行，根据信息采集的识别结果触发深度学习模型模块工作，4. An unmanned vehicle based on computing power perception and edge cloud computing collaboration according to claim 1, characterized in that the driving mode control module can process the received status data and identify the corresponding sub-line Whether the operating environment status is normal. If the line operating status is abnormal, the unmanned vehicle will stop running, and the deep learning model module will be triggered based on the identification results of the information collection. 所述深度学习模型模块能利用可用深度学习模型对收到的线路和车辆基础数据进行处理分析，并且将分析的结果在中心云计算中心内的数据存储模块中进行存储，与此同时，从处理分析的结果传输到驾驶档位操纵序列，并将驾驶档位操纵序列数据传输给边缘通信模块；The deep learning model module can use the available deep learning model to process and analyze the received line and vehicle basic data, and store the analysis results in the data storage module in the central cloud computing center. At the same time, from the processing The analysis results are transmitted to the driving gear manipulation sequence, and the driving gear manipulation sequence data is transmitted to the edge communication module; 所述智能算法模块能利用仿生智能算法对车辆行驶时的路线进行及时的分析和处理，并将分析和处理后驾驶曲线数据传输到边缘通信模块。The intelligent algorithm module can use the bionic intelligent algorithm to conduct timely analysis and processing of the vehicle's driving route, and transmit the analyzed and processed driving curve data to the edge communication module. 5.根据权利要求1所述的一种基于算力感知和边缘云计算协同的无人车，其特征在于，所述集群控制系统集群控制系统可部署在云端，控制无人机/车集群，负责整个无人机/车集群的任务分配，综合无人机/车集群的反馈信息，从而实现系统决策、调整集群任务等功能，集群控制系统还可完成算力资源需求较高的任务，进而将计算结果或执行任务下发至无人机/车机载设备；5. An unmanned vehicle based on computing power perception and edge cloud computing collaboration according to claim 1, characterized in that the cluster control system can be deployed in the cloud to control the drone/vehicle cluster, Responsible for the task allocation of the entire UAV/vehicle cluster, and integrating the feedback information of the UAV/vehicle cluster to achieve system decision-making, adjustment of cluster tasks and other functions. The cluster control system can also complete tasks with high computing resource requirements, and then Send calculation results or execution tasks to drone/vehicle onboard equipment; 所述控制无人机/车集群通过传感器为无人系统获取环境信息，机载计算设备嵌入式单元，进行数据分析，基于数据类型进行多传感器的数据融合计算，完成对于感知信号的处理以及执行控制等指令，可完成如协同路径规划、协同避障、协同巡逻等计算资源需求较大的执行任务，重要模块组成有通信模块、控制模块、感知模块及定位模块等；The control drone/vehicle cluster obtains environmental information for the unmanned system through sensors, and the embedded unit of the airborne computing device performs data analysis, performs multi-sensor data fusion calculations based on data types, and completes the processing and execution of the sensing signals. Control and other instructions can complete execution tasks that require large computing resources, such as collaborative path planning, collaborative obstacle avoidance, collaborative patrol, etc. The important modules include communication module, control module, perception module and positioning module, etc.; 通信模块负责无人机/车之间的本地通信及无人机/车与集群控制系统之间的远程通信，主要功能为接收集群控制系统的任务指令，向集群控制系统发送数据及获取其他无人机/车信息，通信模块获得任务指令或其他无人机/车信息后，发送到控制模块，由机载电脑进一步处理，驱动算法执行或提供信息等；通信模块在边缘无人机/无人车和云平台之间利于4G/5G的运营商网络进行域外连接，主要传输边缘传感器的采集一维数据，例如海拔高度、GPS经纬度、电机运转速率，IMU陀螺仪运动状态等数据；红外、深度相机、激光雷达采集的二维数据；多个无人车与无人机之间，通过云平台接收不同边缘设备的位置参数,交互各自状态以及位置信息；The communication module is responsible for local communication between UAVs/vehicles and remote communication between UAVs/vehicles and the cluster control system. Its main functions are to receive task instructions from the cluster control system, send data to the cluster control system and obtain other unmanned aerial vehicles. Human-machine/vehicle information, after the communication module obtains mission instructions or other drone/vehicle information, it is sent to the control module, which is further processed by the onboard computer to drive algorithm execution or provide information, etc.; the communication module is at the edge of the drone/unmanned vehicle The connection between people and vehicles and the cloud platform is conducive to 4G/5G operator networks for out-of-area connections. It mainly transmits one-dimensional data collected by edge sensors, such as altitude, GPS longitude and latitude, motor operating speed, IMU gyroscope motion status and other data; infrared, Two-dimensional data collected by depth cameras and lidar; multiple unmanned vehicles and drones receive location parameters of different edge devices through the cloud platform, and interact with each other's status and location information; 控制模块包含机载电脑、飞控等设备，完成指定任务以及保证无人机/车的正常飞行及行驶；The control module includes on-board computers, flight control and other equipment to complete designated tasks and ensure the normal flight and driving of the drone/car; 感知模块主要由机载传感器组成，摄像头可获取图像信息，为图像处理算法提供数据，是环境信息的主要来源，激光雷达获得3D点云，可获取环境中物体的距离信息，为3D建模、路径规划等提供数据；The perception module is mainly composed of airborne sensors. The camera can obtain image information and provide data for the image processing algorithm. It is the main source of environmental information. The lidar obtains 3D point clouds and can obtain distance information of objects in the environment, which is used for 3D modeling and Provide data for path planning, etc.; 定位模块接收卫星系统的定位信息，解算后得到无人机/车的位置坐标等，是完成集群任务的前提；The positioning module receives the positioning information of the satellite system and obtains the position coordinates of the drone/vehicle after solving the problem, which is the prerequisite for completing the cluster task; 所述卫星定位系统：为无人机/车提供定位信息。The satellite positioning system: provides positioning information for drones/vehicles. 6.一种基于算力感知和边缘云计算协同的无人车的协同方法，其特征在于，包括以下步骤：6. A collaborative method for unmanned vehicles based on computing power awareness and edge cloud computing collaboration, which is characterized by including the following steps: S1、首先，将可用的无人机和无人车上的算力资源注册到算力池中，并记录其相关信息，如处理能力、通信特性等；这样可以为后续的调度和分配提供基础数据；S1. First, register the available computing power resources of drones and unmanned vehicles into the computing power pool, and record their relevant information, such as processing capabilities, communication characteristics, etc.; this can provide a basis for subsequent scheduling and allocation. data; (1)算力资源注册(1)Computing resource registration 当无人机或无人车进入协同工作环境时，它们可以向算力池注册其可提供的算力资源信息；注册包括但不限于处理器类型、处理器核心数、内存容量、存储容量等；When drones or unmanned vehicles enter a collaborative work environment, they can register the computing resource information they can provide with the computing power pool; registration includes but is not limited to processor type, number of processor cores, memory capacity, storage capacity, etc. ; (2)算力资源管理(2)Computing resource management 算力资源管理涉及将可用的算力资源分配给需要执行任务的无人机或无人车，可以使用优先级调度算法来进行资源管理，为不同类型的任务分配不同的优先级，根据任务的紧急程度和重要性，将任务分配给相应的无人机或无人车；Computing resource management involves allocating available computing resources to drones or unmanned vehicles that need to perform tasks. You can use priority scheduling algorithms for resource management to assign different priorities to different types of tasks. According to the task Based on the urgency and importance, assign the task to the corresponding drone or unmanned vehicle; S2、任务需求获取可以通过与任务发起者进行通信来完成，任务发起者可以向协同系统提供任务的描述、优先级、期限等信息，并将其发送给算力池，也可以通过传感器数据收集和处理实现，无人机和无人车可以使用各种传感器(如摄像头、雷达、激光扫描仪等)来感知环境并提取任务需求；S2. Obtaining task requirements can be completed by communicating with the task initiator. The task initiator can provide the description, priority, deadline and other information of the task to the collaborative system and send it to the computing power pool, or it can also collect sensor data And processing implementation, drones and unmanned vehicles can use various sensors (such as cameras, radars, laser scanners, etc.) to perceive the environment and extract mission requirements; 基于优先级的任务调度算法，为不同类型的任务分配不同的优先级，根据任务的紧急程度和重要性来确定任务的执行顺序；可以使用优先队列或调度规则来实现优先级排序；Priority-based task scheduling algorithm assigns different priorities to different types of tasks, and determines the execution order of tasks according to the urgency and importance of the tasks; priority queues or scheduling rules can be used to achieve priority sorting; S3、算力网络资源调度系统的核心是资源调度策略。系统根据不同的优化指标提供4种算力资源调度策略：成本感知调度策略、负载感知调度策略、能效感知调度策略以及服务层协议(SLA)感知调度策略。可以对4种调度策略进行模块化组合使用，实现多目标优化。将待部署的应用记为S0、应用类型标签记为L，按时间由近到远的顺序将已部署过且具有相同标签的应用其记为集合S＝{S1,S2,…,…Sn}；S3. The core of the computing power network resource scheduling system is the resource scheduling strategy. The system provides four computing resource scheduling strategies based on different optimization indicators: cost-aware scheduling strategy, load-aware scheduling strategy, energy efficiency-aware scheduling strategy, and service layer agreement (SLA)-aware scheduling strategy. Four scheduling strategies can be used in modular combination to achieve multi-objective optimization. The application to be deployed is recorded as S0, the application type label is recorded as L, and the applications that have been deployed and have the same label are recorded in order from recent to distant time as a set S={S1, S2,...,...Sn} ; 初始化模块：主要使用业务应用标签匹配的方法；Initialization module: mainly uses the business application tag matching method; 应用需求匹配模块：进行算力资源需求和网络资源需求的相似度分析；Application demand matching module: performs similarity analysis on computing resource requirements and network resource requirements; 历史策略匹配模块：若对S中的所有应用进行标签匹配后，均未发现与S0的相似应用，则使用历史策略匹配模块进行调度；(待部署任务与历史上相同或相似的任务比对进行调度)Historical policy matching module: If no similar application to S0 is found after label matching for all applications in S, the historical policy matching module is used for scheduling; (the task to be deployed is compared with the same or similar tasks in history Scheduling) 资源调度部署模块：微服务匹配筛选过程完成后，若找到了相似度较高的微服务，则按照匹配微服务使用的资源调度策略进行部署，否则随机选择一种资源调度策略部署；Resource scheduling and deployment module: After the microservice matching and screening process is completed, if a microservice with a high degree of similarity is found, it will be deployed according to the resource scheduling strategy used by the matching microservice, otherwise a resource scheduling strategy will be randomly selected for deployment; S4、根据调度策略，从算力池中选择适合条件的算力资源来执行任务。考虑资源的可用性、性能需求和任务优先级等因素，进行资源分配；S4. According to the scheduling policy, select computing power resources suitable for the conditions from the computing power pool to execute the task. Allocate resources considering factors such as resource availability, performance requirements, and task priority; S5、在任务执行过程中，实时监控资源的使用情况和任务进展；通过监测资源的负载、任务的执行时间等指标，及时调整资源分配或重新分配任务，以优化整体的执行效率；S5. During task execution, monitor resource usage and task progress in real time; by monitoring resource load, task execution time and other indicators, timely adjust resource allocation or reallocate tasks to optimize overall execution efficiency; (1)状态评估和反馈控制(1) Status evaluation and feedback control 根据传感器数据，对任务执行者的状态进行评估，并提供适当的反馈控制信号。状态评估可以基于滤波器或其他状态估计方法进行，以获得准确的任务执行者状态信息；Based on the sensor data, the status of the task performer is evaluated and appropriate feedback control signals are provided. State assessment can be performed based on filters or other state estimation methods to obtain accurate task performer state information; (2)异常检测和处理(2) Anomaly detection and processing 实时监控任务执行过程中的异常情况，如传感器故障、通信中断、资源耗尽等；可以使用异常检测算法来识别和处理异常情况；Real-time monitoring of abnormal situations during task execution, such as sensor failure, communication interruption, resource exhaustion, etc.; anomaly detection algorithms can be used to identify and handle abnormal situations; (3)完成度评估(3)Completion evaluation 根据任务描述或标准，评估任务的完成度，并生成相应的指标或报告来监控任务执行情况；完成度评估可以根据任务需求和条件进行定义，例如时间戳、位置检测、目标达成等；Evaluate task completion based on task descriptions or standards, and generate corresponding indicators or reports to monitor task execution; completion evaluation can be defined based on task requirements and conditions, such as timestamps, location detection, goal achievement, etc.; S6、如果发生资源故障、通信中断或任务失败等异常情况，需要进行相应的异常处和故障恢复；这可能包括重新分配资源、更换故障组件或重新规划任务等措施；S6. If abnormal situations such as resource failure, communication interruption or task failure occur, corresponding exception handling and fault recovery need to be carried out; this may include measures such as reallocating resources, replacing faulty components or re-planning tasks; S7、根据任务数量和复杂度的变化，动态地扩展或收缩算力池中的资源。当任务量增加时，可以添加更多无人机或无人车来增加算力资源；而当任务减少时，可以减少资源以节约能源成本。S7. Dynamically expand or shrink the resources in the computing power pool according to changes in the number and complexity of tasks. When the number of tasks increases, more drones or unmanned vehicles can be added to increase computing resources; when the tasks decrease, resources can be reduced to save energy costs.
CN202311649495.6A 2023-12-05 2023-12-05 An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration Withdrawn CN117519944A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311649495.6A CN117519944A (en) 2023-12-05 2023-12-05 An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311649495.6A CN117519944A (en) 2023-12-05 2023-12-05 An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration

Publications (1)

Publication Number Publication Date
CN117519944A true CN117519944A (en) 2024-02-06

Family

ID=89754942

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311649495.6A Withdrawn CN117519944A (en) 2023-12-05 2023-12-05 An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration

Country Status (1)

Country Link
CN (1) CN117519944A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118101657A (en) * 2024-04-23 2024-05-28 长视科技股份有限公司 A method and device for optimizing collaborative computing power among multiple devices for water resources
CN118300990A (en) * 2024-05-30 2024-07-05 亚信科技（中国）有限公司 Power-calculation network energy-saving control method, device, equipment and storage medium
CN119323885A (en) * 2024-10-24 2025-01-17 北京邮电大学 Vehicle-mounted real-time road information acquisition system based on Internet of things and edge calculation
CN119322682A (en) * 2024-12-19 2025-01-17 环球数科股份有限公司 Self-adaptive calculation power scheduling system for large model training
CN119473631A (en) * 2025-01-10 2025-02-18 深圳市欧阳麦乐科技有限公司 Computer and cloud computing power optimization method based on multi-task collaboration
CN120528956A (en) * 2025-07-23 2025-08-22 四川鑫兴自动化及仪表工程有限责任公司 Real-time industrial process optimization control system based on edge computing
CN120980708A (en) * 2025-08-12 2025-11-18 芜湖辛巴网络科技有限公司 A method and system for ultra-low latency cooperative communication
CN121078049A (en) * 2025-11-06 2025-12-05 成都通广网联科技有限公司 Intelligent driving edge cooperative computing system applied to expressway

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118101657A (en) * 2024-04-23 2024-05-28 长视科技股份有限公司 A method and device for optimizing collaborative computing power among multiple devices for water resources
CN118300990A (en) * 2024-05-30 2024-07-05 亚信科技（中国）有限公司 Power-calculation network energy-saving control method, device, equipment and storage medium
CN119323885A (en) * 2024-10-24 2025-01-17 北京邮电大学 Vehicle-mounted real-time road information acquisition system based on Internet of things and edge calculation
CN119322682A (en) * 2024-12-19 2025-01-17 环球数科股份有限公司 Self-adaptive calculation power scheduling system for large model training
CN119473631A (en) * 2025-01-10 2025-02-18 深圳市欧阳麦乐科技有限公司 Computer and cloud computing power optimization method based on multi-task collaboration
CN120528956A (en) * 2025-07-23 2025-08-22 四川鑫兴自动化及仪表工程有限责任公司 Real-time industrial process optimization control system based on edge computing
CN120980708A (en) * 2025-08-12 2025-11-18 芜湖辛巴网络科技有限公司 A method and system for ultra-low latency cooperative communication
CN121078049A (en) * 2025-11-06 2025-12-05 成都通广网联科技有限公司 Intelligent driving edge cooperative computing system applied to expressway

Similar Documents

Publication Publication Date Title
CN117519944A (en) An unmanned vehicle and its collaboration method based on computing power awareness and edge cloud computing collaboration
US11209816B2 (en) Autonomous long range aerial vehicles and fleet management system
Barmpounakis et al. Unmanned Aerial Aircraft Systems for transportation engineering: Current practice and future challenges
US11355021B1 (en) Nodal network infrastructure for unmanned aerial vehicle operations at scale
CN110703802A (en) Bridge automatic detection method and system based on multi-UAV cooperative operation
CN120472719A (en) Low-altitude flight safety management method based on multi-source data monitoring
Hossain et al. A UAV-based traffic monitoring system for smart cities
CN118968758B (en) A method for low-altitude traffic rescue dispatching under open hard shoulder conditions on highways
CN117572891A (en) Unmanned aerial vehicle cluster system and working method thereof
CN116880555A (en) Unmanned aerial vehicle crowd dispatch system based on edge calculation
Politi et al. The future of safe BVLOS drone operations with respect to system and service engineering
Clark et al. CCAST: A framework and practical deployment of heterogeneous unmanned system swarms
CN118714259A (en) Air-ground integrated intelligent inspection system
CN118981903A (en) Digital twin system for port autonomous driving control and management
CN117576888A (en) A highway accident handling system and method based on traffic control drones
CN120198038A (en) A campus express delivery air-ground integrated unmanned delivery system and usage method
CN120848558A (en) Adaptive control allocation method for multiple UAVs based on mobile drone nests
CN119339585A (en) Low-altitude UAV flight safety assurance system based on intelligent meteorological and airspace monitoring
CN120540347A (en) All-weather autonomous inspection method and system based on cluster task dynamic load balancing
Amelin et al. Task allocation algorithm for the cooperating group of light autonomous unmanned aerial vehicles
CN119862243A (en) Low-altitude three-dimensional traffic network division method and system
CN119762275A (en) A grid-based dispatching method for air-ground integrated unmanned aerial vehicles
CN119250327A (en) UAV flight management method, device, system and computer equipment
Telikani Adversarial Learning in Unmanned Aerial Vehicles-aided Intelligent Transportation Systems
Shmelova et al. Collaborative Decision-Making During the Rescue Operation by the Unmanned Aerial Vehicle Group

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20240206

WW01 Invention patent application withdrawn after publication