CN117425137A - Wireless edge federal learning transmission scheduling method and device - Google Patents
Wireless edge federal learning transmission scheduling method and device Download PDFInfo
- Publication number
- CN117425137A CN117425137A CN202311402807.3A CN202311402807A CN117425137A CN 117425137 A CN117425137 A CN 117425137A CN 202311402807 A CN202311402807 A CN 202311402807A CN 117425137 A CN117425137 A CN 117425137A
- Authority
- CN
- China
- Prior art keywords
- training task
- receiving
- message
- vehicle
- federal learning
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 81
- 230000005540 biological transmission Effects 0.000 title claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 113
- 238000012790 confirmation Methods 0.000 claims abstract description 24
- 238000004891 communication Methods 0.000 claims abstract description 18
- 230000002776 aggregation Effects 0.000 claims description 51
- 238000004220 aggregation Methods 0.000 claims description 51
- 230000006870 function Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 230000004044 response Effects 0.000 claims description 5
- 230000007246 mechanism Effects 0.000 abstract description 5
- 230000008569 process Effects 0.000 description 20
- 238000010586 diagram Methods 0.000 description 8
- 238000010295 mobile communication Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 3
- 101100172132 Mus musculus Eif3a gene Proteins 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 101001077478 Homo sapiens RNA guanine-N7 methyltransferase activating subunit Proteins 0.000 description 1
- 102100025054 RNA guanine-N7 methyltransferase activating subunit Human genes 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/098—Distributed learning, e.g. federated learning
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/04—Wireless resource allocation
- H04W72/044—Wireless resource allocation based on the type of the allocated resource
- H04W72/0446—Resources in time domain, e.g. slots or frames
Abstract
The application discloses a wireless edge federal learning transmission scheduling method, which is used for roadside units and comprises the following steps: receiving a federal learning training task from the cloud; broadcasting the training task and model parameters on a service channel; receiving a request message from a vehicle-mounted terminal device on a control channel, wherein the request message comprises a training task request; responding to the training task request, sending a training task confirmation instruction to the vehicle-mounted terminal, and allocating a special time slot of a service channel; and receiving a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters. The method also comprises a step for the vehicle-mounted terminal equipment, and the device for realizing the method. The method solves the problems of limited federal learning communication, low efficiency and the like in complex and changeable wireless environments, and particularly provides an edge internet of vehicles transmission scheduling mechanism aiming at the internet of vehicles environment.
Description
Technical Field
The application relates to the technical field of wireless communication and Internet of vehicles, in particular to a wireless edge federal learning transmission scheduling method and equipment.
Background
Because training data is usually generated and stored in different user devices in a distributed manner in a 5G network, especially in a dense terminal scenario oriented to intelligent networking automobiles, internet of things and the like, it becomes particularly important how to implement distributed training and reasoning with lower communication overhead, better convergence, security and privacy protection. Federal learning is used as a distributed machine learning framework, which allows users to train models by using local data sets, and in the training process, the data cannot leave the local of the users, and only the parameter change amount of the models is shared, so that the safety of the data is ensured, and the prediction effect is always better than that of the models independently trained by each federal user.
However, in a wireless network, due to factors such as mobility and channel performance of terminal nodes, bandwidth limitation, etc., the transmission scheduling of nodes participating in federal learning becomes very important to the performance impact of final model training. The existing mechanism considers the aspects of node selection and coordination, model compression and the like, but does not consider the scheduling mechanism and scheme of a channel access layer.
Disclosure of Invention
The invention provides a wireless edge federal learning transmission scheduling method and equipment, and aims to solve the problems of federal learning communication limitation, low efficiency and the like in a complex and changeable wireless environment, and particularly provides an edge internet of vehicles transmission scheduling mechanism aiming at an internet of vehicles environment.
In a first aspect, an embodiment of the present application provides a wireless edge federal learning transmission scheduling method, which is used for a roadside unit, and includes the following steps:
receiving a federal learning training task from the cloud;
broadcasting the training task and model parameters on a service channel;
receiving a request message from a vehicle-mounted terminal device on a control channel, wherein the request message comprises a training task request;
responding to the training task request, sending a training task confirmation instruction to the vehicle-mounted terminal, and allocating a special time slot of a service channel;
and receiving a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters.
In one embodiment of the present application, the method further comprises the steps of: and according to the updated model parameters, parameter aggregation is carried out, and the ith round of model aggregation parameters are broadcasted in a service channel.
In one embodiment of the present application, the method further comprises the steps of: receiving, in the dedicated time slot, a message from the terminal device of at least one of: the ith round of model aggregation parameters receive successful confirmation information; the ith round of model aggregation parameter receiving failure message; the i+1th update parameter transmits an incomplete message; training incomplete messages in the (i+1) th round; learning termination message.
In one embodiment of the present application, the method further comprises the steps of: sending the ith round of model aggregation parameters to the cloud; and receiving an aggregation result from the cloud and broadcasting the aggregation result to the vehicle-mounted terminal equipment in the coverage area.
In one embodiment of the present application, the method further comprises the steps of: and responding to the learning termination message from the special time slot of the service channel, releasing the special time slot, and calculating the contribution degree of the terminal equipment.
In one embodiment of the present application, the method further comprises the steps of: and releasing the special time slot to report vehicle contribution to the cloud in response to the fact that the training task report message is not received by m continuous frames.
In a second aspect, an embodiment of the present application further provides a wireless edge federal learning transmission scheduling method, which is used for a vehicle-mounted terminal device, and includes the following steps:
acquiring broadcast messages of training tasks and model parameters;
transmitting a request message to a roadside unit on a control channel, wherein the request message comprises a training task request;
receiving training task confirmation indication and channel resource occupation state information in a control channel;
determining a special time slot of a service channel according to the channel resource occupation state information;
performing model training;
and sending a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters.
In one embodiment of the present application, the method further comprises the steps of: at the dedicated time slot, sending at least one of the following messages to the roadside unit device: the ith round of model aggregation parameters receive successful confirmation information; the ith round of model aggregation parameter receiving failure message; the i+1th update parameter transmits an incomplete message; training incomplete messages in the (i+1) th round; learning termination message.
In an embodiment of any one of the first or second aspects of the present application, preferably, the request message further includes a service channel slot acquisition request.
In an embodiment of any one of the first or second aspects of the present application, the dedicated time slots are preferably communicated in an encrypted manner.
In a third aspect, embodiments of the present application further provide a roadside unit device, configured to implement the method according to any one of the embodiments of the first aspect of the present application, where at least one module in the roadside unit device is configured to at least one of the following functions: receiving a federal learning training task; broadcasting the training task and model parameters; receiving a request message from a vehicle-mounted terminal; transmitting a training task confirmation instruction; transmitting channel resource occupation state information; receiving a training task report message; performing parameter aggregation, and determining and transmitting parameters of an ith round of model set; receiving an aggregation result from a cloud; determining to release the dedicated time slot; determining the contribution degree of the terminal equipment; a contribution bonus message is sent.
In a fourth aspect, an embodiment of the present application further proposes a vehicle-mounted terminal device, configured to implement a method according to any one of the embodiments of the second aspect of the present application, where at least one module in the vehicle-mounted terminal device is configured to implement at least one of the following functions: receiving the broadcast message; sending the request message; receiving the training task confirmation indication; receiving the channel resource occupation state information; determining a dedicated time slot of a service channel; determining updated model parameters; sending the training task report message; a contribution rewards message is received.
In a fifth aspect, embodiments of the present application further provide a communication device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, which when executed by the processor performs the steps of the method according to any one of the embodiments of the first and second aspects of the present application.
In a sixth aspect, embodiments of the present application further provide a computer readable medium having a computer program stored thereon, the computer program implementing the steps of the method according to any one of the embodiments of the first and second aspects of the present application when executed by a processor.
In a seventh aspect, embodiments of the present application further provide a system for wireless edge federal learning transmission scheduling, including at least 1 roadside unit device according to embodiments of the third aspect of the present application and/or at least 1 vehicle-mounted terminal device according to embodiments of the fourth aspect of the present application.
Preferably, the system for wireless edge federal learning transmission scheduling further comprises a cloud server, wherein the cloud server is used for at least one of the following functions: issuing a training task to the roadside unit equipment; receiving an ith round of model aggregation parameters from the roadside unit device; and performing cloud side aggregation, and sending an aggregation result to the roadside unit equipment.
The above-mentioned at least one technical scheme that this application embodiment adopted can reach following beneficial effect:
the utilization efficiency of communication resources in the federal learning of the wireless edge is improved, so that the efficiency of model training in the federal learning process is improved;
the flexibility and the expandability of the wireless edge federal learning system under the condition of participation of a plurality of heterogeneous mobile terminals are improved, and the enthusiasm of terminal participation can be improved by introducing an incentive mode.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute an undue limitation to the application. In the drawings:
FIG. 1 is a system embodiment of an application scenario of the present application;
FIG. 2 is a schematic diagram of control channels and service channels;
FIG. 3 is a flow chart of an embodiment of the method of the present application;
FIG. 4 is a flow chart of an embodiment of a method of the present application for a roadside unit device;
fig. 5 is a flowchart of an embodiment of a method for a vehicle-mounted terminal device according to the present application;
FIG. 6 is a schematic diagram of an embodiment of a roadside unit apparatus;
fig. 7 is a schematic diagram of an embodiment of an in-vehicle terminal device;
FIG. 8 is a schematic view of a roadside unit apparatus according to another embodiment of the present invention;
fig. 9 is a block diagram of an in-vehicle terminal device of another embodiment of the present invention.
Detailed Description
For the purposes, technical solutions and advantages of the present application, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
The following describes in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.
Fig. 1 is a system embodiment of an application scenario of the present application.
The system of the application comprises roadside unit equipment 400 and vehicle-mounted terminal equipment 500, and further comprises cloud equipment 100. The invention considers the federal learning process between moving vehicles in a cloud-edge scene. Each roadside unit (RSU) device acts as an edge server, responsible for managing the federal learning process of vehicles within its coverage area. Multiple roadside unit devices can aggregate for the same federal learning task, and then the cloud is responsible for overall aggregation. For example, communication is performed between the roadside unit device and the roadside unit device, and between the roadside unit device and the cloud terminal through a wired network. The roadside unit device communicates with the vehicle using a wireless frequency band.
Fig. 2 is a schematic diagram of control channels and service channels. Referring to the WAVE protocol, the communication procedure between the roadside unit device and the vehicle takes 100ms as one period (denoted as one frame). Each frame is divided into two channel transmission parts, a control channel and a service channel. The two processes are provided with a guide interval before starting, and each node in the Internet of vehicles needs to start working after the guide interval. The main function of the guiding interval is to synchronize the time of each node in the internet of vehicles so that the channel switching without conflict can be realized in a coordinated way. During the pilot interval, neither the control channel nor the service channel is active. The roadside unit equipment and the vehicle can acquire channel resources by utilizing MAC mechanisms such as VeMAC [1], RAMAC [2] and the like, and data transmission can be carried out on a service channel in a TDMA mode. Wherein the roadside unit devices utilize the control channels for broadcasting security messages, control messages (including channel resource allocation messages).
In the embodiment of the application, the roadside unit equipment acquires a special transmission time slot in a service channel, broadcasts a training task and a model parameter update related message in the transmission time slot, and may also include broadcasting an incentive and rewarding related message of a federal learning process, a confirmation message of a receiving state of a model parameter sent by a vehicle and the like.
Fig. 3 is a flow chart of an embodiment of the method of the present application. The application provides a wireless edge federal learning transmission scheduling method, wherein a federal learning scheduling process among cloud equipment, roadside unit equipment and vehicle-mounted terminal equipment comprises the following steps 110-170:
step 110, the cloud end equipment issues training tasks to roadside unit equipment;
the cloud end distributes federal learning training tasks to roadside unit equipment, and the roadside unit equipment distributes the training tasks to vehicles in a coverage area to participate in training. Each roadside unit device may be responsible for different training tasks, or may be responsible for the same training task.
Step 120, broadcasting a training task to the vehicle-mounted terminal by roadside unit equipment;
the roadside unit equipment continuously broadcasts the current channel resource occupation state information on a control channel, and broadcasts training tasks and model parameters on a service channel, and can also comprise broadcasting of motivation and rewarding related messages of the federal learning process.
130, monitoring a training task of the vehicle-mounted terminal and monitoring channel resource occupation state information;
the vehicle first listens for at least one complete frame length, thereby obtaining current channel resource occupancy state information, federal learning training tasks and release information of model parameters, and may also include the reception of incentive and rewarding related messages of the federal learning process. The channel resource occupancy status information will describe the status of each time slot of the current transmission channel and the occupied vehicles.
Step 140, the vehicle-mounted terminal trains task request and service channel time slot acquisition request, the roadside unit apparatus sends training task confirmation indication, allocates the special time slot of the service channel;
if the vehicle decides to join the federal learning process, starting from the second frame, firstly, transmitting information for applying federal training and related time slot acquisition to roadside unit equipment in a control channel through a CSMA transmission mode, wherein the information is used for acquiring idle transmission time slots (at least one time slot is acquired according to the model size) of a service channel.
Since a message collision with another vehicle is likely to occur when a vehicle transmits a slot acquisition message in a control channel through a transmission scheme of CSMA. Thus requiring roadside unit equipment for validation. When the roadside unit device successfully receives the message of the vehicle, the vehicle is agreed to enter the federal learning training process and the time slot is still in an idle state, the roadside unit device allocates the time slot to the vehicle and notifies the vehicle through the channel resource occupation state information. By continuously monitoring messages from roadside units, the vehicle, when acknowledging that it successfully acquires the transmission time slot of the dedicated service channel
Step 150, the vehicle-mounted terminal executes a training task and reports the training task through a special time slot and roadside unit equipment;
after the vehicle-mounted terminal confirms that the vehicle-mounted terminal successfully acquires the transmission time slot of the special service channel, the vehicle-mounted terminal starts to utilize local data to carry out model training and update of model parameters, and updates and reports the model parameters in the transmission time slot occupied by the vehicle-mounted terminal;
for example, after the vehicle successfully receives the latest model from the roadside unit device, the vehicle performs message confirmation of successful reception of the ith round of model aggregation parameters in the transmission time slot of the dedicated service channel, otherwise, performs message transmission of failed reception of the ith round of model aggregation parameters.
For another example, after the vehicle completes a new round (i+1st round) of model training, reporting model parameters by using a transmission time slot of a dedicated service channel. If the vehicle cannot finish reporting all the model parameters in one frame, a mark of 'i+1st round of transmission incompletion' is sent simultaneously to prompt roadside unit equipment to continue updating the model parameters in the next frame. If the new training is not completed, the vehicle sends a message of 'i+1st training is not completed' in the transmission time slot to prompt that the roadside unit equipment is still in the new training process of the model.
In order to protect own data privacy, the vehicle can communicate with the roadside unit device in an encrypted manner when transmitting the model parameters in a special time slot.
And 160, aggregating and reporting the model parameters, generating an aggregation result and broadcasting.
The roadside unit equipment receives model updating parameters sent by the vehicle in the coverage area in the occupied time slot, aggregates the parameters, broadcasts the aggregated model parameters and the current iteration round number (i-th round) to the vehicle node, and reports the model parameters to the cloud.
The vehicle-mounted terminal receives the latest model parameters from the roadside unit equipment in the service channel, and the roadside unit equipment marks the current training turns.
The cloud end can receive model parameters from (one or more) roadside unit devices participating in the same training task in an asynchronous mode, and after cloud side aggregation is completed, a final aggregation result is issued to all the roadside unit devices, and the roadside unit devices are responsible for broadcasting to all vehicles in the coverage range of the roadside unit devices.
At this time, a round of global training process is ended, and the above steps 130 to 150 are continued until the model training reaches a preset accuracy or the whole algorithm converges.
Step 70, exiting the federal learning process and releasing the special time slot.
When the vehicle decides to exit the federal learning process, a message of 'learning termination' is sent in a time slot special for the vehicle, contribution degree of the vehicle is calculated by roadside unit equipment and rewarded to the vehicle, and the special time slot of the vehicle for model parameter sending is released in the next frame.
When the vehicle does not send a message for m continuous frames, the roadside unit equipment judges that the vehicle leaves the coverage area of the vehicle or is interrupted, releases a special time slot of the vehicle, reports the contribution state of the vehicle to the cloud, and sends relevant rewards after the vehicle is re-networked by the cloud. The value setting of m can be flexibly set according to the running speed of the vehicle, the network state and the like in the application process.
It should be noted that, the network entity used in the wireless communication system includes the vehicle-mounted terminal device, the roadside unit device or other intermediate devices; the above steps may also be used for providing information processing service means for the network entity device; the steps can be used for any device, system, subsystem, circuit, chip or software entity for providing information receiving, transmitting, identifying and processing for the vehicle-mounted terminal equipment or roadside unit equipment.
Fig. 4 is a flowchart of an embodiment of a method of the present application for a roadside unit device.
The method according to any one of the embodiments of the first aspect of the present application, for use in a roadside unit device, comprises at least some of the following steps 210 to 280.
Step 210, receiving a federal learning training task from the cloud;
step 220, broadcasting the training task and model parameters in a service channel;
step 230, receiving a request message from the vehicle-mounted terminal equipment on a control channel, wherein the request message comprises a training task request;
preferably, the request message further comprises a service channel slot acquisition request.
Step 240, in response to the training task request, sending a training task confirmation instruction to the vehicle-mounted terminal, and allocating a dedicated time slot of a service channel;
step 250, receiving a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters;
preferably, the dedicated time slots communicate in an encrypted manner.
Preferably, in the dedicated time slot, at least one message from the terminal device is also received: the ith round of model aggregation parameters receive successful confirmation information; the ith round of model aggregation parameter receiving failure message; the i+1th update parameter transmits an incomplete message; training incomplete messages in the (i+1) th round; learning termination message. These messages may all be included in the training task report message.
Step 260, in one embodiment of the present application, further comprises the steps of: and according to the updated model parameters, parameter aggregation is carried out, and the ith round of model aggregation parameters are broadcasted in a service channel.
Step 270, reporting model parameters to the cloud and receiving an aggregation result;
sending the ith round of model aggregation parameters to the cloud; and receiving an aggregation result from the cloud and broadcasting the aggregation result to the vehicle-mounted terminal equipment in the coverage area.
In some embodiments of the present application, preferably, further comprising:
step 280, releasing the special time slot.
In one embodiment of the present application, the dedicated time slot is released in response to a learning termination message from the dedicated time slot of the service channel, and the contribution of the terminal device is calculated.
In one embodiment of the present application, the dedicated time slot is released to report vehicle contributions to the cloud in response to the continuous m frames not receiving the training task report message.
Fig. 5 is a flowchart of an embodiment of the method for a vehicle-mounted terminal device.
The embodiment of the application also provides a wireless edge federal learning transmission scheduling method, which is used for the vehicle-mounted terminal equipment and comprises at least one part of the following steps 310-360:
step 310, acquiring a broadcast message of training task and model parameters;
step 320, transmitting a request message to a roadside unit on a control channel, wherein the request message comprises a training task request;
preferably, the request message further comprises a service channel slot acquisition request.
Step 330, receiving training task confirmation indication and channel resource occupation state information in a control channel;
step 340, determining a dedicated time slot of the service channel according to the channel resource occupation state information;
step 350, performing model training;
step 360, a training task report message is sent in the dedicated time slot of the service channel, wherein the training task report message contains updated model parameters.
Preferably, the dedicated time slots communicate in an encrypted manner.
And in the special time slot, the method is further used for sending at least one of the following messages to roadside unit equipment: the ith round of model aggregation parameters receive successful confirmation information; the ith round of model aggregation parameter receiving failure message; the i+1th update parameter transmits an incomplete message; training incomplete messages in the (i+1) th round; learning termination message.
Fig. 6 is a schematic diagram of an embodiment of a roadside unit apparatus.
The embodiment of the application also provides roadside unit equipment, which is used for realizing the method of any one embodiment of the application, wherein at least one module in the roadside unit equipment is used for at least one of the following functions: receiving a federal learning training task; broadcasting the training task and model parameters; receiving a request message from a vehicle-mounted terminal; transmitting a training task confirmation instruction; transmitting channel resource occupation state information; receiving a training task report message; performing parameter aggregation, and determining and transmitting parameters of an ith round of model set; receiving an aggregation result from a cloud; determining to release the dedicated time slot; determining the contribution degree of the terminal equipment; a contribution bonus message is sent.
In order to implement the above technical solution, the roadside unit device 400 provided in the present application includes a network transmitting module 401, a network determining module 402, and a network receiving module 403 that are connected to each other.
The network sending module is used for broadcasting the training task and the model parameters; transmitting a training task confirmation instruction; transmitting channel resource occupation state information; transmitting the model set parameters of the ith round; a contribution bonus message is sent.
The network determining module is used for performing parameter aggregation and determining the parameters of the ith round of model set; determining to release the dedicated time slot; and determining the contribution degree of the terminal equipment.
The network receiving module is used for receiving federal learning training tasks from the cloud end equipment; receiving a request message from a vehicle-mounted terminal; receiving a training task report message; and receiving an aggregation result from the cloud.
Specific methods for implementing the functions of the network sending module, the network determining module and the network receiving module are described in the embodiments of the methods of the present application, and are not described here again.
The roadside unit device can refer to an intelligent traffic system roadside device (RSU), a traffic wireless communication system base station facility, roadside unit devices or servers connected with a base station, can also be a system for providing services for the devices, and can also be any system, subsystem, module, circuit, chip or software running device for providing information receiving, transmitting, identifying and processing for the devices.
Fig. 7 is a schematic diagram of an embodiment of the in-vehicle terminal device.
The application further proposes a vehicle-mounted terminal device, configured to implement the method according to any one of the embodiments of the application, where at least one module in the vehicle-mounted terminal device is configured to implement at least one of the following functions: receiving the broadcast message; sending the request message; receiving the training task confirmation indication; receiving the channel resource occupation state information; determining a dedicated time slot of a service channel; determining updated model parameters; sending the training task report message; a contribution rewards message is received.
In order to implement the above technical solution, the vehicle-mounted terminal device 500 provided in the present application includes a terminal sending module 501, a terminal determining module 502, and a terminal receiving module 503 that are connected to each other.
The terminal receiving module is used for receiving the broadcast message; receiving the training task confirmation indication; receiving the channel resource occupation state information; a contribution rewards message is received.
The terminal determining module is used for determining a special time slot of a service channel; updated model parameters are determined.
The terminal sending module is used for sending the request message; and sending the training task report message.
Specific methods for implementing the functions of the terminal sending module, the terminal determining module and the terminal receiving module are described in the embodiments of the methods of the present application, and are not described herein.
The vehicle-mounted terminal device can refer to an intelligent network-connected automobile communication unit, an on-vehicle User Equipment (UE), a personal mobile terminal, an intelligent terminal, a mobile phone and a computer with a communication function, can also be a system for providing services for the device, and can also be any system, subsystem, module, circuit, chip or software running device for providing information receiving, transmitting, identifying and processing for the device.
Fig. 8 is a schematic view showing a construction of a roadside unit apparatus according to another embodiment of the present invention. As shown, the roadside unit device 600 includes a processor 601, a wireless interface 602, and a memory 603. Wherein the wireless interface may be a plurality of components, i.e. comprising a transmitter and a receiver, providing a means for communicating with various other apparatuses over a transmission medium. The wireless interface realizes the communication function with the vehicle-mounted terminal equipment, processes wireless signals through the receiving and transmitting device, and the data carried by the signals are communicated with the memory or the processor through the internal bus structure. The memory 603 contains a computer program for executing any of the embodiments of the present application, which computer program runs or changes on the processor 601. When the memory, the processor, and the wireless interface circuit are connected through a bus system, the bus system includes a data bus, a power bus, a control bus, and a status signal bus, which are not described herein again.
Fig. 9 is a block diagram of an in-vehicle terminal device of another embodiment of the present invention. The in-vehicle terminal device 700 includes at least one processor 701, a memory 702, a user interface 703, and at least one network interface 704. The various components in the in-vehicle terminal device 700 are coupled together by a bus system. Bus systems are used to enable connected communication between these components. The bus system includes a data bus, a power bus, a control bus, and a status signal bus.
The user interface 703 may include a display, keyboard, or pointing device, such as a mouse, trackball, touch pad, or touch screen, among others.
The memory 702 stores executable modules or data structures. The memory may store an operating system and application programs. The operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks. The application programs include various application programs such as a media player, a browser, etc. for implementing various application services.
In an embodiment of the present invention, the memory 702 contains a computer program that executes any of the embodiments of the present application, the computer program running or changing on the processor 701.
The memory 702 contains a computer readable storage medium, and the processor 701 reads the information in the memory 702 and performs the steps of the above method in combination with its hardware. In particular, the computer readable storage medium has stored thereon a computer program which, when executed by the processor 701, implements the steps of the method embodiments as described in any of the embodiments above.
The processor 701 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the methods of the present application may be performed by integrated logic circuitry in hardware or instructions in software in processor 701. The processor 701 may be a general purpose processor, a digital signal processor, an application specific integrated circuit, an off-the-shelf programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. The disclosed methods, steps, and logic blocks in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be embodied directly in the execution of a hardware decoding processor, or in the execution of a combination of hardware and software modules in a decoding processor.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. In one typical configuration, the device of the present application includes one or more processors (CPUs), an input/output user interface, a network interface, and memory.
Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Accordingly, the present application also proposes a computer readable medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method according to any of the embodiments of the present application. For example, the memory 603, 702 of the present invention may include non-volatile memory in a computer-readable medium, random Access Memory (RAM) and/or non-volatile memory, etc., such as read-only memory (ROM) or flash RAM.
Based on the embodiment of the device of the application, the embodiment of the application also provides a system for wireless edge federal learning transmission scheduling, as shown in fig. 1, which comprises at least 1 roadside unit device according to any one embodiment of the application and/or at least 1 vehicle-mounted terminal device according to any one embodiment of the application.
Preferably, the system for wireless edge federal learning transmission scheduling further comprises a cloud server, wherein the cloud server is used for at least one of the following functions: issuing a training task to the roadside unit equipment; receiving an ith round of model aggregation parameters from the roadside unit device; and performing cloud side aggregation, and sending an aggregation result to the roadside unit equipment.
It should be noted that the specific mobile communication technology described in the present invention is not limited, and may be WCDMA, CDMA2000, TD-SCDMA, wiMAX, LTE/LTE-A, LAA, muLTEfire, and fifth generation, sixth generation, and nth generation mobile communication technologies that may occur subsequently.
The terminal described in the present invention refers to a terminal side product capable of supporting a communication protocol of a land mobile communication system, and a Modem module (Wireless Modem) for special communication, which can be integrated by various types of terminal forms such as a mobile phone, a tablet computer, a data card, etc. to complete a communication function.
For convenience of description, a fourth generation mobile communication system LTE/LTE-a and its derivative multewire are taken as an example, wherein a mobile communication terminal may be denoted as UE (User Equipment), and an access device at a network side may be denoted as a base station or an access point.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.
Claims (16)
1. A wireless edge federal learning transmission scheduling method is used for roadside units and is characterized by comprising the following steps:
receiving a federal learning training task from the cloud;
broadcasting the training task and model parameters on a service channel;
receiving a request message from a vehicle-mounted terminal device on a control channel, wherein the request message comprises a training task request;
responding to the training task request, sending a training task confirmation instruction to the vehicle-mounted terminal, and allocating a special time slot of a service channel;
and receiving a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters.
2. The wireless edge federal learning transmission scheduling method of claim 1, further comprising the steps of:
and according to the updated model parameters, parameter aggregation is carried out, and the ith round of model aggregation parameters are broadcasted in a service channel.
3. The wireless edge federal learning transmission scheduling method of claim 1, further comprising the steps of:
receiving, in the dedicated time slot, a message from the terminal device of at least one of:
the ith round of model aggregation parameters receive successful confirmation information;
the ith round of model aggregation parameter receiving failure message;
the i+1th update parameter transmits an incomplete message;
training incomplete messages in the (i+1) th round;
learning termination message.
4. The wireless edge federal learning transmission scheduling method of claim 1, further comprising the steps of:
sending the ith round of model aggregation parameters to the cloud;
and receiving an aggregation result from the cloud and broadcasting the aggregation result to the vehicle-mounted terminal equipment in the coverage area.
5. The wireless edge federal learning transmission scheduling method of claim 1, further comprising the steps of:
and responding to the learning termination message from the special time slot of the service channel, releasing the special time slot, and calculating the contribution degree of the terminal equipment.
6. The wireless edge federal learning transmission scheduling method of claim 1, further comprising the steps of:
and releasing the special time slot to report vehicle contribution to the cloud in response to the fact that the training task report message is not received by m continuous frames.
7. The wireless edge federal learning transmission scheduling method is used for vehicle-mounted terminal equipment and is characterized by comprising the following steps of:
acquiring broadcast messages of training tasks and model parameters;
transmitting a request message to a roadside unit on a control channel, wherein the request message comprises a training task request;
receiving training task confirmation indication and channel resource occupation state information in a control channel;
determining a special time slot of a service channel according to the channel resource occupation state information;
performing model training;
and sending a training task report message in the special time slot of the service channel, wherein the training task report message comprises updated model parameters.
8. The wireless edge federal learning transmission scheduling method of claim 7, further comprising the steps of:
at the dedicated time slot, sending at least one of the following messages to the roadside unit device:
the ith round of model aggregation parameters receive successful confirmation information;
the ith round of model aggregation parameter receiving failure message;
the i+1th update parameter transmits an incomplete message;
training incomplete messages in the (i+1) th round;
learning termination message.
9. The wireless edge federal learning transmission scheduling method according to any one of claims 1 to 8, wherein the request message further comprises a service channel slot acquisition request.
10. The wireless edge federal learning transmission scheduling method according to any one of claims 1 to 8, wherein the dedicated time slots communicate in an encrypted manner.
11. A roadside unit apparatus for carrying out the method as claimed in any one of claims 1 to 6, characterized in that,
at least one module in the roadside unit device is used for at least one of the following functions: receiving a federal learning training task; broadcasting the training task and model parameters; receiving a request message from a vehicle-mounted terminal; transmitting a training task confirmation instruction; transmitting channel resource occupation state information; receiving a training task report message; performing parameter aggregation, and determining and transmitting parameters of an ith round of model set; receiving an aggregation result from a cloud; determining to release the dedicated time slot; determining the contribution degree of the terminal equipment; a contribution bonus message is sent.
12. A vehicle terminal device for implementing the method according to any one of claims 7 to 8, characterized in that,
at least one module in the vehicle-mounted terminal equipment is used for realizing at least one of the following functions: receiving the broadcast message; sending the request message; receiving the training task confirmation indication; receiving the channel resource occupation state information; determining a dedicated time slot of a service channel; determining updated model parameters; sending the training task report message; a contribution rewards message is received.
13. A communication device, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, which when executed by the processor performs the steps of the method according to any one of claims 1 to 10.
14. A computer readable medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method according to any of claims 1 to 10.
15. A system for wireless edge federal learning transmission scheduling, comprising at least 1 roadside unit device according to claim 11 and/or at least 1 vehicle-mounted terminal device according to claim 12.
16. The system for wireless edge federal learning transmission scheduling of claim 15, further comprising a cloud server,
the cloud server is used for
Issuing a training task to the roadside unit equipment;
receiving an ith round of model aggregation parameters from the roadside unit device;
and performing cloud side aggregation, and sending an aggregation result to the roadside unit equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311402807.3A CN117425137A (en) | 2023-10-26 | 2023-10-26 | Wireless edge federal learning transmission scheduling method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311402807.3A CN117425137A (en) | 2023-10-26 | 2023-10-26 | Wireless edge federal learning transmission scheduling method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117425137A true CN117425137A (en) | 2024-01-19 |
Family
ID=89532091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311402807.3A Pending CN117425137A (en) | 2023-10-26 | 2023-10-26 | Wireless edge federal learning transmission scheduling method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117425137A (en) |
-
2023
- 2023-10-26 CN CN202311402807.3A patent/CN117425137A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113965908B (en) | Resource collision solving method, device and storage medium | |
| CN105657842B (en) | A kind of method and roadside unit of resource allocation | |
| CN111277973B (en) | Auxiliary link communication method, device and storage medium | |
| Zhuang et al. | On the uplink MAC performance of a drive-thru Internet | |
| CN105516214A (en) | Information reporting method, resource allocating method, information reporting device and resource allocating device in Internet-of-vehicles system | |
| Ni et al. | Vehicular beacon broadcast scheduling based on age of information (AoI) | |
| CN109803385A (en) | Request the method and device of resource, the method and device of process resource request | |
| CN103595461B (en) | A kind of resource allocation methods of ultra-large VSAT system frame structure | |
| CN106060957A (en) | CSMA/TDMA hybrid access control method | |
| CN111654348A (en) | Internet of things data transmission method and device, computer equipment and storage medium | |
| Nguyen et al. | Improving time slot acquisition through RSU's coordination for TDMA-based MAC protocol in VANETs | |
| CN107295561B (en) | Resource application method and device | |
| CN110381586A (en) | Resource allocation methods, apparatus and system | |
| CN106412834B (en) | Television white spectrum broadband communication system based on WiFi chip and communication method thereof | |
| US20220353030A1 (en) | Configuration method and device for data transmission structure | |
| CN117425137A (en) | Wireless edge federal learning transmission scheduling method and device | |
| Azizian et al. | Improved multi-channel operation for safety messages dissemination in vehicular networks | |
| Cao et al. | OGCMAC: A novel OFDM based group contention MAC for VANET control channel | |
| Sun et al. | Adaptive beaconing for collision avoidance and tracking accuracy in vehicular networks | |
| CN107370703B (en) | Information receiving and transmitting method, device and system | |
| WO2016115932A1 (en) | Channel occupation method and system | |
| TWI451780B (en) | Service providing apparatus, service consuming apparatus, and service transmitting method thereof | |
| CN102035706B (en) | Method for carrying out communication capacity expansion on static segment of vehicle-mounted electronics bus | |
| CN112655171A (en) | Reference signal transmission method and device | |
| Segata et al. | Distributed EDCA bursting: improving cluster-based communication in IVC |
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 |