System and method for network resource optimization

Abstract

A method for network resource optimization may include selecting a first navigator device from a plurality of wireless devices in a wireless communication. The first navigator device is in wireless communication with at least one follower device of the plurality of wireless devices. The method may also include transferring the follower data processing task from the at least one follower device to the first pilot device. The follower data processing task includes at least one of a computing task and a communication task. The method may also include performing a follower data processing task using the first navigator device.

Description

Technical Field

[0001] The present disclosure relates to systems and methods for network resource optimization, and more particularly, to systems and methods for optimizing the offloading of computing and communication tasks to remote computing resources. Background Art

[0002] To enhance occupant awareness and convenience, vehicles and / or mobile devices can be equipped with various sensors and systems, such as sensing sensors (e.g., cameras, radars, ultrasonic distance sensors, etc.), microphones, navigation systems, advanced driver assistance systems (ADAS), autonomous driving systems (ADS), etc. Such sensors and systems can generate data that requires further computing tasks to be performed, such as computer vision processing, video processing, natural language processing, speech recognition, navigation routing, autonomous driving path planning, etc. Accordingly, computing tasks can be offloaded to remote computing resources (e.g., remote data centers). However, some vehicles and / or mobile devices may lack the computing and / or communication capabilities required to coordinate the offloading and / or transmission of tasks and related data to remote computing resources. Thus, current offloading systems may not optimally allocate remote computing resources.

[0003] Therefore, while current computing systems and methods achieve their intended purposes, there is still a need for a new and improved system and method for network resource optimization. Summary of the Invention

[0004] According to several aspects, a method for network resource optimization is provided. The method can include selecting a first leader device from among a plurality of wireless devices in a wireless communication. The first leader device wirelessly communicates with at least one follower device among the plurality of wireless devices. The method can further include transferring a follower data processing task from the at least one follower device to the first leader device. The follower data processing task includes at least one of a computing task and a communication task. The method can further include using the first leader device to execute the follower data processing task.

[0005] In another aspect of the present disclosure, selecting the first leader device can further include determining a leader matrix. The leader matrix includes one or more candidate leader devices. Each of the one or more candidate leader devices is one of the plurality of wireless devices. The leader matrix is determined based at least in part on at least one of the data processing task capabilities of each of the plurality of wireless devices and the historical performance reliability of each of the plurality of wireless devices. Selecting the first leader device can further include selecting the first leader device from the leader matrix.

[0006] In another aspect of the present disclosure, determining the leader matrix may further include determining the data processing task capabilities of each of the plurality of wireless devices. A capabilities estimation machine learning model is used to determine the data processing task capabilities of each of the plurality of wireless devices. The data processing task capabilities include at least one of computing capabilities and communication capabilities. Determining the leader matrix may further include determining the historical performance reliability of each of the plurality of wireless devices. Determining the leader matrix may further include selecting one or more candidate leader devices from the plurality of wireless devices at least partially based on the data processing task capabilities of each of the plurality of wireless devices and the historical performance reliability of each of the plurality of wireless devices. Determining the leader matrix may further include determining the leader matrix at least partially based on the one or more candidate leader devices.

[0007] In another aspect of the present disclosure, determining the leader matrix may further include determining the leader matrix, where the leader matrix is a sorted list including each of the one or more candidate leader devices. The ranking order of the leader matrix is determined at least partially based on the data processing task capabilities and historical performance reliability of each of the one or more candidate leader devices.

[0008] In another aspect of the present disclosure, selecting a first leader device from the leader matrix may further include determining the planned data processing task capabilities of each of the one or more candidate leader devices. A capabilities prediction machine learning model is used to determine the planned data processing task capabilities of each of the one or more candidate leader devices. Selecting a first leader device from the leader matrix may further include determining the ranking order of the leader matrix at least partially based on the planned data processing task capabilities of each of the one or more candidate leader devices.

[0009] In another aspect of the present disclosure, determining the leader matrix may further include calculating the cost of each of the one or more candidate leader devices using a cost function:

[0010] t c =(t0*z)+P0

[0011]

[0012] where T c is the cost, t0 is the number of interactions allowed within a given time period without exceeding the device's low-power limit, z is the total cost of performing t0 interactions within the given time period, P0 is the cost per unit of power consumed within the given time period, r c，Mb is the cost of transmitting 1 megabyte of registration information within the given time period, c c，Mbis the cost of transmitting 1 megabyte of payload data over a given period of time, i is the number of interactions performed over a given period of time, and ∫i(t)dt is the integral of i evaluated over multiple periods of time. Determining the leader matrix can also include determining a ranking order of the leader matrix based at least in part on the cost of each of one or more candidate leader devices.

[0013] In another aspect of the present disclosure, transferring the follower data processing task can also include transferring a computing task from at least one follower device to a first leader device. The computing task includes an offloading optimization computing task. Transferring the follower data processing task can also include transferring a communication task from at least one follower device to the first leader device. The communication task includes a server uplink communication task.

[0014] In another aspect of the present disclosure, transferring the follower data processing task can also include identifying an obstacle that impedes transmission between a first follower device and a first leader device. Transferring the follower data processing task can also include relaying the follower data processing task from the first follower device through a second follower device to the first leader device to avoid the obstacle.

[0015] In another aspect of the present disclosure, transferring the computing task can also include transferring the offloading optimization computing task from at least one follower device to a first leader device. The offloading optimization computing task includes performing a device-specific offloading machine learning model.

[0016] In another aspect of the present disclosure, performing the follower data processing task can also include determining the data processing task capabilities of a first leader device. Performing the follower data processing task can also include transferring the follower data processing task to a second leader device based at least in part on the data processing task capabilities of the first leader device.

[0017] According to several aspects, a system for optimizing network resources for a vehicle is provided. The system can include a first follower vehicle that includes a follower vehicle communication system. The system can also include a follower vehicle controller in electrical communication with the follower vehicle communication system. The follower vehicle controller is programmed to establish a first wireless connection with a leader vehicle using the follower vehicle communication system. The follower vehicle controller is programmed to transfer a follower data processing task to the leader vehicle using the follower vehicle communication system and the first wireless connection.

[0018] In another aspect of the present disclosure, to establish a first wireless connection with a leader vehicle, the follower vehicle controller is further programmed to determine a leader matrix. The leader matrix includes one or more candidate leader vehicles. The leader matrix is determined at least in part based on at least one of the data processing task capabilities of each of the plurality of vehicles and the historical performance reliability of each of the plurality of vehicles. To establish a first wireless connection with a leader vehicle, the follower vehicle controller is further programmed to select a leader vehicle from the leader matrix.

[0019] In another aspect of the present disclosure, to determine the leader matrix, the follower vehicle controller is further programmed to determine the data processing task capabilities of each of the plurality of vehicles. A capability estimation machine learning model is used to determine the data processing task capabilities of each of the plurality of vehicles. The data processing task capabilities include at least one of computing capability and communication capability. To determine the leader matrix, the follower vehicle controller is further programmed to determine the historical performance reliability of each of the plurality of vehicles. To determine the leader matrix, the follower vehicle controller is further programmed to select one or more candidate leader vehicles from the plurality of vehicles at least in part based on the data processing task capabilities of each of the plurality of vehicles and the historical performance reliability of each of the plurality of vehicles. To determine the leader matrix, the follower vehicle controller is further programmed to determine the leader matrix at least in part based on the one or more candidate leader vehicles. The leader matrix is a sorted list including each of the one or more candidate leader vehicles. The ranking order of the leader matrix is determined at least in part based on the data processing task capabilities and historical performance reliability of each of the one or more candidate leader vehicles.

[0020] In another aspect of the present disclosure, to select a leader vehicle from the leader matrix, the follower vehicle controller is further programmed to determine the planned data processing task capabilities for each of the one or more candidate leader vehicles. A capability prediction machine learning model is used to determine the planned data processing task capabilities for each of the one or more candidate leader vehicles. To select a leader vehicle from the leader matrix, the follower vehicle controller is further programmed to calculate the cost for each of the one or more candidate leader vehicles using a cost function:

[0021] t c =(t o *z)+P0

[0022]

[0023] where t cis the cost, t0 is the number of interactions allowed within a given time period without exceeding the device's low - power limit, z is the total cost of performing t0 interactions within a given time period, P0 is the cost per unit of power consumed within a given time period, r c，Mb is the cost of transmitting 1 megabyte of registration information within a given time period, c c，Mb is the cost of transmitting 1 megabyte of payload data within a given time period, i is the number of interactions performed within a given time period, and ∫i(t)dt is the integral of i evaluated over multiple time periods. To select a leader vehicle from the leader matrix, the follower vehicle controller is also programmed to determine a ranking order of the leader matrix based at least in part on the planned data - processing task capabilities of each of one or more candidate leader vehicles and the cost of each of the one or more candidate leader vehicles.

[0024] In another aspect of the present disclosure, to transfer the follower data - processing task to the leader vehicle, the follower vehicle controller is also programmed to use the follower vehicle communication system to transfer an offloading optimization computing task from the follower vehicle to the leader vehicle. The offloading optimization computing task includes executing a vehicle - specific offloading machine - learning model.

[0025] In another aspect of the present disclosure, the system further includes a leader vehicle that includes a leader vehicle communication system and a leader vehicle controller in electrical communication with the leader vehicle communication system. The leader vehicle controller is programmed to establish a first wireless connection with a first follower vehicle using the leader vehicle communication system. The leader vehicle controller is also programmed to receive a follower data - processing task from the first follower vehicle using the leader vehicle communication system and the first wireless connection. The leader vehicle controller is also programmed to establish a second wireless connection with a server system using the leader vehicle communication system. The leader vehicle controller is also programmed to perform the follower data - processing task using the leader vehicle communication system and the second wireless connection.

[0026] In another aspect of the present disclosure, to receive a follower data - processing task from the first follower vehicle, the leader vehicle controller is also programmed to identify an obstacle that impedes transmission between the first follower vehicle and the leader vehicle. To receive a follower data - processing task from the first follower vehicle, the leader vehicle controller is also programmed to terminate the first wireless connection between the leader vehicle and the first follower vehicle in response to identifying the obstacle.

[0027] A method for optimizing network resources for a vehicle is provided in several aspects. The method may include selecting a first leader vehicle from a plurality of vehicles in wireless communication. The first leader vehicle communicates wirelessly with at least one follower vehicle among the plurality of vehicles. The method may further include transferring follower data processing tasks from at least one follower vehicle to the first leader vehicle. The follower data processing tasks include executing a vehicle-specific offloading machine learning model. The method may further include using the first leader vehicle to execute the follower data processing tasks.

[0028] In another aspect of the present disclosure, selecting the first leader vehicle may further include determining a leader matrix. The leader matrix includes one or more candidate leader vehicles. Each of the one or more candidate leader vehicles is one of the plurality of vehicles. The leader matrix is determined at least in part based on at least one of the data processing task capabilities of each of the plurality of vehicles and the historical performance reliability of each of the plurality of vehicles. Selecting the first leader vehicle may further include selecting the first leader vehicle from the leader matrix.

[0029] In another aspect of the present disclosure, determining the leader matrix may further include determining the data processing task capabilities of each of the plurality of vehicles. A capability estimation machine learning model is used to determine the data processing task capabilities of each of the plurality of vehicles. The data processing task capabilities include at least one of computing power and communication power. Determining the leader matrix may further include determining the historical performance reliability of each of the plurality of vehicles. Determining the leader matrix may further include selecting one or more candidate leader vehicles from the plurality of vehicles at least in part based on the data processing task capabilities of each of the plurality of vehicles and the historical performance reliability of each of the plurality of vehicles. Determining the leader matrix may further include calculating the cost of each of the one or more candidate leader vehicles using a cost function:

[0030] t c =(t0*z)+P0

[0031]

[0032] where t c is the cost, t0 is the number of interactions allowed within a given time period without exceeding the device's low power limit, z is the total cost of performing t0 interactions within a given time period, P0 is the cost per unit of power consumed within a given time period, r c，Mb is the cost of transmitting 1 megabyte of registration information within a given time period, c c，Mbis the cost of transmitting 1 megabyte of payload data over a given time period, i is the number of interactions performed over a given time period, and ∫i(t)dt is the integral of i evaluated over multiple time periods. Determining the leader matrix may also include determining the leader matrix at least in part based on one or more candidate leader vehicles. The leader matrix is a sorted list including each of the one or more candidate leader vehicles. The ranking order of the leader matrix is determined at least in part based on the data processing task capabilities, the historical performance reliability of each of the one or more candidate leader vehicles, and the cost of each of the one or more candidate leader vehicles.

[0033] Based on the description provided herein, further applicable fields will become apparent. It should be understood that the specification and specific examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The drawings described herein are for illustrative purposes only and are not intended to limit the scope of the present disclosure in any way.

[0035] Figure 1 is a schematic diagram of a system for network resource optimization according to an exemplary embodiment;

[0036] Figure 2 is a flowchart of a method for network resource optimization according to an exemplary embodiment;

[0037] Figure 3 is a flowchart of a method for reducing obstacles according to an exemplary embodiment; and

[0038] Figure 4 is a flowchart of a method for selecting a representative leader device according to an exemplary embodiment. DETAILED DESCRIPTION

[0039] The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses.

[0040] In aspects of the present disclosure, some devices may lack the computing and / or communication capabilities to complete a particular task. For example, low-power Internet of Things (IoT) devices may lack the computing power to execute advanced machine learning algorithms. Additionally, low-power IoT devices may lack the communication capabilities to establish a connection with an external server using a cellular data connection. Accordingly, the present disclosure provides a new and improved system and method for network resource optimization that allows for the transfer of computing and / or communication tasks between devices.

[0041] REFERENCE Figure 1, a system for network resource optimization is shown and generally designated by reference numeral 10. System 10 generally includes at least one leader device and at least one follower device. Within the scope of the present disclosure, the at least one leader device and the at least one follower device can include any electronic device capable of computing and / or communicating, including, for example, smartphones, tablets, personal computers, wearable devices (e.g., smartwatches), server computers, roadside units (RSUs), vehicles, IoT devices, etc. In Figure 1 In the exemplary embodiment shown, the at least one leader device and the at least one follower device are vehicles selected from a plurality of vehicles 12. In one non-limiting example, the at least one leader device is a first leader vehicle 12a, and the at least one follower device includes a first follower vehicle 12b and a second follower vehicle 12c. Each of the plurality of vehicles 12 includes a vehicle system 14. Although the present disclosure is primarily directed to vehicles as an example, it should be understood that the present disclosure is equally applicable to any plurality of wireless devices.

[0042] The vehicle system 14 includes a vehicle controller 20, a plurality of vehicle sensors 22, and a vehicle communication system 24.

[0043] The vehicle controller 20 is configured to implement a method 100 for network resource optimization, as described below. The vehicle controller 20 includes at least one processor and a non-transitory computer-readable storage device or medium. The processor can be a custom or commercially available processor, a central processing unit (CPU), a graphics processing unit (GPU), an auxiliary processor among several processors associated with the vehicle controller 20, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, combinations thereof, or generally a device for executing instructions. The computer-readable storage device or medium can include, for example, volatile and non-volatile storage devices such as read-only memory (ROM), random access memory (RAM), and keep-alive memory (KAM). KAM is persistent or non-volatile memory that can be used to store various operating variables when the processor is powered off. The computer-readable storage device or medium can be implemented using multiple storage devices, such as programmable read-only memory (PROM), electrically erasable PROM (EEPROM), electrically erasable and programmable ROM (EEPROM), flash memory, or other electrical, magnetic, optical, or combined storage devices capable of storing data, some of which represent executable instructions used by the vehicle controller 20 to control various systems of the vehicle 12. The vehicle controller 20 can also be composed of a plurality of controllers that communicate electrically with each other. The vehicle controller 20 can be interconnected with additional systems and / or controllers of the vehicle 12, allowing the vehicle controller 20 to access data such as the speed, acceleration, braking, and steering angle of the vehicle 12.

[0044] In an exemplary embodiment, the capabilities of vehicle controller 20 can vary between vehicles. Thus, the vehicle controller 20 of the first leader vehicle 12a can be referred to as the leader vehicle controller. The vehicle controllers 20 of the first follower vehicle 12b and the second follower vehicle 12c can be referred to as follower vehicle controllers. In a non-limiting example, the leader vehicle controller of the first leader vehicle 12a includes enhanced computing capabilities, including, for example, additional random access memory, additional processing power, and / or speed, etc. In a non-limiting example, the follower vehicle controllers of the first follower vehicle 12b and the second follower vehicle 12c include reduced computing capabilities compared to the leader vehicle controller of the first leader vehicle 12a. Thus, in aspects of the present disclosure, it is advantageous to transfer (i.e., “offload”) computing tasks from the first follower vehicle 12b and the second follower vehicle 12c to the first leader vehicle 12a, as will be discussed in more detail below.

[0045] The vehicle controller 20 communicates electrically with a plurality of vehicle sensors 22 and a vehicle communication system 24. In an exemplary embodiment, electrical communication is established using, for example, a CAN network, a FLEXRAY network, a local area network (e.g., WiFi, Ethernet, etc.), or a serial peripheral interface (SPI) network, etc. It should be understood that various additional wired and wireless technologies and communication protocols for communicating with the vehicle controller 20 are within the scope of the present disclosure.

[0046] The plurality of vehicle sensors 22 are used to acquire telemetry data of the vehicle 12. Within the scope of the present disclosure, telemetry data includes, for example, engine RPM, vehicle speed, fuel level, engine temperature, odometer reading, battery voltage, brake system status, transmission data, tire pressure, GNSS position, acceleration and deceleration, steering angle, suspension system data, exhaust emission level, diagnostic trouble code (DTC), airbag status, windshield wiper status, lights and indicator lights, and cruise control status. In an exemplary embodiment, the plurality of vehicle sensors 22 includes sensors for determining performance data regarding the vehicle 12. In a non-limiting example, the plurality of vehicle sensors 22 further includes at least one of the following: a motor speed sensor, a motor torque sensor, an electric drive motor voltage and / or current sensor, an accelerator pedal position sensor, a brake position sensor, a coolant temperature sensor, a cooling fan speed sensor, a wheel speed sensor, and a transmission oil temperature sensor.

[0047] In another exemplary embodiment, the plurality of vehicle sensors 22 further includes sensors for determining information regarding the environment within the vehicle 12. In a non-limiting example, the plurality of vehicle sensors 22 further includes at least one of a seat occupancy sensor, a cabin air temperature sensor, a cabin motion detection sensor, a cabin camera, a cabin microphone, etc.

[0048] In another exemplary embodiment, the plurality of vehicle sensors 22 further includes sensors for determining information about the environment around the vehicle 12. In one non-limiting example, the plurality of vehicle sensors 22 further includes at least one of an ambient air temperature sensor, an atmospheric pressure sensor, a Global Navigation Satellite System (GNSS), and / or a photographic camera and / or a video camera positioned to observe the environment in front of the vehicle 12.

[0049] The GNSS is used to determine the geographical location of the vehicle 12. In one exemplary embodiment, the GNSS is the Global Positioning System (GPS). In one non-limiting example, the GPS includes a GPS receiver antenna (not shown) and a GPS controller (not shown) in electrical communication with the GPS receiver antenna. The GPS receiver antenna receives signals from multiple satellites, and the GPS controller calculates the geographical location of the vehicle 12 based on the signals received by the GPS receiver antenna. In one exemplary embodiment, the GNSS further includes a map. The map includes information about infrastructure, such as municipal boundaries, roads, railways, sidewalks, buildings, etc. Thus, the map information is used to contextualize the geographical location of the vehicle 12. In one non-limiting example, the map is retrieved from a remote source using a wireless connection. In another non-limiting example, the map is stored in a database of the GNSS. It should be understood that various additional types of satellite-based radio navigation systems, such as the Global Positioning System (GPS), Galileo, GLONASS, and the BeiDou Navigation Satellite System (BDS), are within the scope of the present disclosure. It should be understood that, without departing from the scope of the present disclosure, the GNSS can be integrated with the vehicle controller 20 (e.g., on the same circuit board as the vehicle controller 20 or integrated as part of the vehicle controller 20).

[0050] In another exemplary embodiment, at least one of the plurality of vehicle sensors 22 is a sensing sensor capable of sensing objects in the environment around the vehicle 12 and / or measuring distances. In one non-limiting example, the plurality of vehicle sensors 22 includes a stereo camera having distance measurement capabilities. In one example, at least one of the plurality of vehicle sensors 22 is fixed inside the vehicle 12, for example, fixed in the roof lining of the vehicle 12 and having a field of view through the windshield of the vehicle 12. In another example, at least one of the plurality of vehicle sensors 22 is fixed outside the vehicle 12, for example, fixed on the roof of the vehicle 12 and having a field of view of the environment around the vehicle 12. It should be understood that various additional types of sensing sensors, such as LiDAR sensors, ultrasonic ranging sensors, radar sensors, and / or time-of-flight sensors, are within the scope of the present disclosure. The plurality of vehicle sensors 22 is in electrical communication with the vehicle controller 20, as described above.

[0051] The vehicle controller 20 uses the vehicle communication system 24 to communicate with other systems external to the vehicle 12. For example, the vehicle communication system 24 includes the ability to communicate with vehicles (“V2V” communication), with infrastructure (“V2I” communication), with remote systems at a remote call center (e.g., ON-STAR of General Motors), and / or with personal devices. Generally, the term vehicle-to-everything communication (“V2X” communication) refers to the communication between the vehicle 12 and any remote system (e.g., vehicle, infrastructure, and / or remote system). In some embodiments, the vehicle communication system 24 is a wireless communication system that is configured to communicate via a wireless local area network (WLAN) using the IEEE 802.11 standard or by using cellular data communication (e.g., using GSMA standards such as SGP.02, SGP.22, SGP.32, etc.). Thus, the vehicle communication system 24 may also include an embedded universal integrated circuit card (eUICC) that is configured to store at least one cellular connection configuration profile, such as an embedded subscriber identity module (eSIM) profile. The vehicle communication system 24 is also configured to communicate via a personal area network (e.g., Bluetooth) and / or near field communication (NFC). However, additional or alternative communication methods such as dedicated short range communication (DSRC) channels and / or mobile telecommunication protocols based on the 3rd Generation Partnership Project (3GPP) standards are also considered within the scope of the present disclosure. A DSRC channel refers to a one-way or two-way short-to-medium range wireless communication channel designed specifically for automotive use, as well as a set of corresponding protocols and standards. 3GPP refers to a partnership between multiple standards organizations that develop mobile telecommunication protocols and standards. The structure of the 3GPP standards is “releases”. Thus, communication methods based on 3GPP releases 14, 15, 16, and / or future 3GPP releases are considered within the scope of the present disclosure. Thus, the vehicle communication system 24 may include one or more antennas and / or communication transceivers for receiving and / or transmitting signals, such as cooperative sensing messages (CSM). The vehicle communication system 24 is configured to wirelessly transmit information between the vehicle 12 and another vehicle. In addition, the vehicle communication system 24 is configured to wirelessly transmit information between the vehicle 12 and infrastructure or other vehicles. It should be understood that, without departing from the scope of the present disclosure, the vehicle communication system 24 may be integrated with the vehicle controller 20 (e.g., integrated on the same circuit board as the vehicle controller 20 or integrated as part of the vehicle controller 20).

[0052] In an exemplary embodiment, the capabilities of the vehicle communication system 24 can vary between vehicles. Thus, the vehicle communication system 24 of the first leader vehicle 12a can be referred to as a leader vehicle communication system. The vehicle communication systems 24 of the first follower vehicle 12b and the second follower vehicle 12c can be referred to as follower vehicle communication systems. In a non-limiting example, the leader vehicle communication system of the first leader vehicle 12a includes full communication capabilities, including local area (e.g., Bluetooth, WiFi, NFC, etc.) and wide area (e.g., cellular data) communication capabilities. In a non-limiting example, the follower vehicle communication systems of the first follower vehicle 12b and the second follower vehicle 12c include reduced communication capabilities, such as only local area communication capabilities. Thus, in aspects of the present disclosure, it is advantageous to transfer (i.e., "offload") communication tasks that require wide area communication from the first follower vehicle 12b and the second follower vehicle 12c to the first leader vehicle 12a via local area communication, as will be discussed in more detail below.

[0053] Continuing to refer Figure 1 , in an exemplary embodiment, the system 10 further includes a server system 30. The server system 30 includes a server controller 32a that is in electrical communication with a database 34 and a server communication system 36. In a non-limiting example, the server system 30 is located in a server farm, or a data center, etc., and is connected to the Internet using the server communication system 36. The server controller 32a includes at least one server processor 32b and a server non-transitory computer-readable storage device or server medium 32c. The description of the type and configuration given above for the vehicle controller 20 also applies to the server controller 32a. In some examples, the server controller 32a differs from the vehicle controller 20 in that the server controller 32a can have a higher processing speed, include more memory, include more input / output, etc. In a non-limiting example, the server processor 32b and the server medium 32c of the server controller 32a are similar in structure and / or function to the processor and medium of the vehicle controller 20, as described above. The server controller 32a is used in combination with the vehicle controller 20 to implement a method 100 for network resource optimization, as will be discussed in more detail below. The database 34 is used to store telemetry data received from multiple vehicles 12, as will be discussed in more detail below. The server communication system 36 is used to communicate with an external system (e.g., the vehicle controller 20) via the vehicle communication system 24. In a non-limiting example, the server communication system 36 is similar in structure and / or function to the vehicle communication system 24 of the vehicle system 14, as described above. In some examples, the server communication system 36 differs from the vehicle communication system 24 in that the server communication system 36 can perform higher power signal transmission, more sensitive signal reception, higher bandwidth transmission, additional transmission / reception protocols, etc.

[0054] Reference Figure 2 ,shows a flowchart of a method 100 for network resource optimization. Method 100 begins at block 102. After block 102, method 100 proceeds to blocks 104, 106, 108, and 110. At block 104, the data processing task capabilities of each of the plurality of vehicles 12 are determined. Within the scope of the present disclosure, the data processing task capability is the ability of a wireless device (e.g., one of the plurality of vehicles 12) to perform computing and / or communication tasks. The data processing task capability may include computing capability and communication capability. Within the scope of the present disclosure, the computing capability is the ability of a wireless device to perform computing tasks (e.g., storage / retrieval of data in memory, mathematical calculations, execution of algorithms such as machine learning algorithms, computer vision algorithms, etc.).

[0055] Within the scope of the present disclosure, the communication capability is the ability of a wireless device to perform communication tasks (e.g., wireless and / or wired transmission / reception of data). Within the scope of the present disclosure, the capability refers to the total amount of tasks that can be processed simultaneously. In one non-limiting example, the computing capability is measured, for example, in floating-point operations per second, instructions per second, etc. In one non-limiting example, the communication capability is measured, for example, in bits per second, etc.

[0056] In one exemplary embodiment, each of the plurality of vehicles 12 determines the computing capability of the vehicle controller 20 of the vehicle system 14 and the communication capability of the vehicle communication system 24 of the vehicle system 14. In one non-limiting example, the computing capability is determined at least in part based on the number and / or characteristics of one or more computing tasks currently being executed by the vehicle controller 20. In one non-limiting example, the communication capability is determined at least in part based on the number and / or characteristics of one or more communication tasks currently being executed by the vehicle communication system 24.

[0057] In another non-limiting example, a capability estimation machine learning model executed by the vehicle controller 20 of each of the plurality of vehicles 12 is used to determine the data processing task capability. In one non-limiting example, the capability estimation machine learning model includes multiple layers, including an input layer and an output layer, and one or more hidden layers. The input layer receives as inputs the physical computing capability of the vehicle controller 20 (e.g., the maximum number of floating-point operations per second of the vehicle controller 20), the physical communication capability of the vehicle communication system 24 (e.g., the maximum number of bits per second of the vehicle communication system 24), the characteristics of one or more computing tasks currently being executed by the vehicle controller 20, and the characteristics of one or more communication tasks currently being executed by the vehicle communication system 24. The input is then passed to the hidden layer. Each hidden layer applies a transformation (e.g., a non-linear transformation) to the data and passes the result to the next hidden layer until the last hidden layer. The output layer produces the data processing task capability.

[0058] To train an ability estimation machine learning model, a dataset of inputs and their corresponding data processing task abilities is used. The model is trained by adjusting the internal weights between nodes in each hidden layer to minimize the prediction error. During training, optimization techniques (such as gradient descent) are used to adjust the internal weights to reduce the prediction error. The training process is repeated for the entire dataset until the prediction error is minimized, and then the resulting trained model is used to process new input data.

[0059] After the ability estimation machine learning model is sufficiently trained, the model is able to accurately and precisely determine the data processing task ability based on the physical computing ability of the vehicle controller 20, the physical communication ability of the vehicle communication system 24, the characteristics of one or more computing tasks currently executed by the vehicle controller 20, and the characteristics of one or more communication tasks currently executed by the vehicle communication system 24. By adjusting the weights between nodes in each hidden layer during the training process, the model "learns" to identify patterns in the data that indicate the data processing task ability.

[0060] In some exemplary embodiments, each of the plurality of vehicles 12 determines the data processing task ability and broadcasts the data processing task ability using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the data processing task ability of each of the plurality of vehicles 12. In another exemplary embodiment, the server system 30 remotely determines the data processing task ability of each of the plurality of vehicles 12. After block 104, method 100 proceeds to block 112, as will be discussed in more detail below.

[0061] In block 106, the historical performance reliability of each of the plurality of vehicles 12 is determined. Within the scope of the present disclosure, the historical performance reliability quantifies the variation in the performance and / or reliability of a wireless device (e.g., one of the plurality of vehicles 12) used to perform computing and / or communication tasks. In one non-limiting example, the historical performance reliability includes metrics such as computing system uptime and / or availability, statistical variance of computing power, network jitter, network availability, network packet loss rate, etc. In one exemplary embodiment, the historical performance reliability includes a weighted average of one or more of the foregoing metrics.

[0062] In some exemplary embodiments, each of the plurality of vehicles 12 determines historical performance reliability and broadcasts the historical performance reliability using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the historical performance reliability of each of the plurality of vehicles 12. In another exemplary embodiment, the server system 30 remotely determines the historical performance reliability of each of the plurality of vehicles 12. After block 106, method 100 proceeds to block 112, as will be discussed in more detail below.

[0063] At block 108, the planned data processing task capabilities of each of the plurality of vehicles 12 are determined. Within the scope of the present disclosure, the planned data processing task capabilities are the predicted future data processing task capabilities of each of the plurality of vehicles 12 based on the scheduled or predicted usage. In one non-limiting example, known and / or scheduled future events such as over-the-air (OTA) updates, system backups, network outages, etc. may affect the planned data processing task capabilities. In another non-limiting example, the characteristics of the currently executing tasks may affect the planned data processing task capabilities. For example, if the currently executing task includes downloading an OTA update, the predicted future data processing task capabilities may take into account the computational load due to installing the update after the download task is completed.

[0064] In one exemplary embodiment, a capability prediction machine learning model is used to determine the planned data processing task capabilities. In one non-limiting example, the capability prediction machine learning model includes multiple layers, including an input layer and an output layer, and one or more hidden layers. The input layer receives as inputs the current data processing task capabilities, known future events and / or tasks, the characteristics of the currently executing tasks, etc. The input is then passed to the hidden layer. Each hidden layer applies a transformation (e.g., a non-linear transformation) to the data and passes the result to the next hidden layer until the last hidden layer. The output layer produces the planned data processing task capabilities.

[0065] To train the capability prediction machine learning model, a dataset of inputs and their corresponding scheduled data processing task capabilities is used. The model is trained by adjusting the internal weights between the nodes in each hidden layer to minimize the prediction error. During the training process, optimization techniques (e.g., gradient descent) are used to adjust the internal weights to reduce the prediction error. The training process is repeated for the entire dataset until the prediction error is minimized, and then the generated trained model is used to process new input data.

[0066] After the ability prediction machine learning model is sufficiently trained, the model can accurately and precisely determine the planned data processing task ability based on the current data processing task ability, known future events and / or tasks, characteristics of the currently executed task, and so on. By adjusting the weights between the nodes in each hidden layer during the training process, the model "learns" to identify patterns in the data that indicate the planned data processing task ability.

[0067] In some exemplary embodiments, each of the plurality of vehicles 12 determines the planned data processing task ability and broadcasts the planned data processing task ability using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the planned data processing task ability of each of the plurality of vehicles 12. In another exemplary embodiment, the server system 30 remotely determines the planned data processing task ability of each of the plurality of vehicles 12. After block 108, method 100 proceeds to block 112, which will be discussed in more detail below.

[0068] At block 110, the cost of each of the plurality of vehicles 12 is determined. Within the scope of the present disclosure, the cost represents the expected monetary cost of establishing and transmitting data using a wireless connection.

[0069] In one exemplary embodiment, a cost function is used to determine the cost:

[0070]

[0071]

[0072] In equation (1), t c is the cost, t0 is the number of interactions allowed within a given time period (e.g., one day) without exceeding the device's low power limit, z is the total cost of performing t0 interactions within the given time period, P0 is the cost per unit of power consumed within the given time period, r c，Mb is the cost of transmitting 1 megabyte of registration information within the given time period, r c，total is the total registration cost, if the vehicle is in the ignition-off state, bool(ignition state) is defined as 1 (i.e., true), if the vehicle is in the ignition-on state, bool(ignition state) is defined as 0 (i.e., false), p l is the device low power limit (i.e., the power consumption allowed within the given time period), n is the number of reconnection times within the given time period, and τ u is the duty cycle of the registration process (i.e., the time required to perform the registration process as a part of the given time period). In equation (1), the number of one or more devices (e.g., one or more controllers of the vehicle) in the vehicle that may consume power to establish a connection Perform summation. Additionally, in Equation (2), c c，Mb is the cost of transmitting 1 megabyte of payload data over a given time period, i is the number of interactions performed over a given time period, and ∫i(t)dt is the integral of i evaluated over multiple time periods (e.g., fourteen days).

[0073] Within the scope of the present disclosure, establishing a wireless connection between multiple devices includes a registration process. The registration process establishes a connection between the devices and includes the transmission of registration information. In one non-limiting example, the registration information includes, for example, device identification data, handshake data, security data, connection parameters, etc. After completion of the registration process, the wireless connection can be used to send payload data. In one non-limiting example, the payload data can include, for example, computational tasks as described above. Within the scope of the present disclosure, an interaction is any wireless connection between two or more devices that includes at least the transmission of payload data. Within the scope of the present disclosure, a device low power limit is the power consumption allowed for a given device over a given time period. In one non-limiting example, the device low power limit can be determined at least in part based on an energy budget (i.e., the total amount of energy consumption allowed over a given time period).

[0074] In another exemplary embodiment, a cost determination machine learning model is used to determine the cost. In one non-limiting example, the cost determination machine learning model includes multiple layers, including an input layer and an output layer, and one or more hidden layers. The input layer receives device power consumption, electricity cost, and data transmission cost information as inputs. The inputs are then passed to the hidden layers. Each hidden layer applies a transformation (e.g., a non-linear transformation) to the data and passes the result to the next hidden layer until the last hidden layer. The output layer produces the cost.

[0075] To train the cost determination machine learning model, a dataset of inputs and their corresponding costs is used. The model is trained by adjusting the internal weights between the nodes in each hidden layer to minimize the prediction error. During the training process, optimization techniques (e.g., gradient descent) are used to adjust the internal weights to reduce the prediction error. The training process is repeated for the entire dataset until the prediction error is minimized, and then the generated trained model is used to process new input data.

[0076] After the cost determination machine learning model has been adequately trained, the model is able to accurately and precisely determine the cost based on device power consumption, electricity cost, and data transmission cost information. By adjusting the weights between the nodes in each hidden layer during the training process, the model "learns" to identify patterns in the data that indicate the cost.

[0077] In some exemplary embodiments, each of the plurality of vehicles 12 determines a cost and broadcasts the cost using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the cost of each of the plurality of vehicles 12. In another exemplary embodiment, the server system 30 remotely determines the cost of each of the plurality of vehicles 12. After block 110, method 100 proceeds to block 112.

[0078] At block 112, a leader matrix is determined. Within the scope of the present disclosure, the leader matrix is a sorted list that includes one or more candidate leader vehicles (also referred to as candidate leader devices). Each of the one or more candidate leader vehicles is selected from the plurality of vehicles 12. The one or more candidate leader vehicles are selected and a ranking order of the leader matrix is determined based at least in part on at least one of the following: the data processing task capabilities of each of the plurality of vehicles 12 determined at block 104, the historical performance reliability of each of the plurality of vehicles 12 determined at block 106, the planned data processing task capabilities of each of the plurality of vehicles 12 determined at block 108, and the cost of each of the plurality of vehicles 12 determined at block 110.

[0079] In one exemplary embodiment, each of the one or more candidate leader vehicles is selected based on a performance threshold. In one non-limiting example, the one or more candidate leader vehicles are selected as one or more of the plurality of vehicles 12 having a data processing task capability greater than or equal to a predetermined data processing task capability threshold. In another non-limiting example, the one or more candidate leader vehicles are selected as one or more of the plurality of vehicles 12 also having a historical performance reliability greater than or equal to a predetermined historical performance reliability threshold.

[0080] In another exemplary embodiment, a machine learning model is used to select each of the one or more candidate leader vehicles. In one non-limiting example, a candidate selection machine learning model is used to select the one or more candidate leader vehicles, and the candidate selection machine learning model is trained to select the one or more candidate leader vehicles based at least in part on at least one of the following: the data processing task capabilities of each of the plurality of vehicles 12 determined at block 104, the historical performance reliability of each of the plurality of vehicles 12 determined at block 106, the planned data processing task capabilities of each of the plurality of vehicles 12 determined at block 108, and the cost of each of the plurality of vehicles 12 determined at block 110.

[0081] In an exemplary embodiment, the ranking order of the leader matrix is determined based on performance metrics. In a non - limiting example, the ranking order of the leader matrix is determined such that candidate vehicles with higher data - processing task capabilities are ranked higher in the leader matrix (i.e., closer to the first position). In another non - limiting example, the ranking order of the leader matrix is determined such that candidate vehicles with higher historical performance reliability are ranked higher in the leader matrix (i.e., closer to the first position). In another non - limiting example, the ranking order of the leader matrix is determined such that candidate vehicles with higher planned data - processing task capabilities are ranked higher in the leader matrix (i.e., closer to the first position). In another non - limiting example, the ranking order of the leader matrix is determined such that candidate vehicles with lower costs are ranked higher in the leader matrix (i.e., closer to the first position).

[0082] In another exemplary embodiment, a machine - learning model is used to determine the ranking order of the leader matrix. In a non - limiting example, a candidate - ranking machine - learning model is used to determine the ranking order of the leader matrix, and the candidate - ranking machine - learning model is trained to rank one or more candidate leader vehicles based at least in part on at least one of the following: the data - processing task capabilities of each of the plurality of vehicles 12 determined in block 104, the historical performance reliability of each of the plurality of vehicles 12 determined in block 106, the planned data - processing task capabilities of each of the plurality of vehicles 12 determined in block 108, and the cost of each of the plurality of vehicles 12 determined in block 110.

[0083] In some exemplary embodiments, each of the plurality of vehicles 12 determines the leader matrix and broadcasts the leader matrix using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the leader matrix. In another exemplary embodiment, the server system 30 remotely determines the leader matrix. After block 112, method 100 proceeds to block 114.

[0084] In block 114, a first leader vehicle 12a is selected from the leader matrix determined in block 112. In an exemplary embodiment, the first leader vehicle 12a is selected as the candidate vehicle with the highest rank in the leader matrix (i.e., the candidate vehicle in the first position). In some exemplary embodiments, each of the plurality of vehicles 12 determines the first leader vehicle 12a and broadcasts the first leader vehicle 12a using the vehicle communication system 24. In other exemplary embodiments, one or more of the plurality of vehicles 12 act as host devices and remotely determine the first leader vehicle 12a. In another exemplary embodiment, the server system 30 remotely determines the first leader vehicle 12a. After block 114, method 100 proceeds to block 116.

[0085] At block 116, the first lead vehicle 12a selected at block 114 establishes a first wireless connection with one or more of the first and second follower vehicles 12b, 12c. In an exemplary embodiment, the first wireless connection is a local connection such as a Bluetooth connection, a Near Field Communication (NFC) connection, a Wireless Local Area Network (WLAN / WiFi) connection, a short-range radio connection, etc. It should be understood that the first wireless connection can utilize any connection protocol configured for peer-to-peer connection between wireless devices. After establishing the first wireless connection, one or more of the first and second follower vehicles 12b, 12c transfer follower data processing tasks to the first lead vehicle 12a. In an exemplary embodiment, the follower data processing tasks include at least one of a computing task and a communication task.

[0086] Within the scope of the present disclosure, computing tasks include, for example, storage / retrieval of data in memory, mathematical calculations, execution of algorithms such as machine learning algorithms, computer vision algorithms, video coding algorithms, etc. In a non-limiting example, the computing tasks include offloading optimization computing tasks. Within the scope of the present disclosure, offloading optimization computing tasks include executing an algorithm that is configured to determine which (if any) computing and / or communication tasks should be offloaded from one or more of the first and second follower vehicles 12b, 12c to the first lead vehicle 12a in order to optimize performance, resource usage, etc. In an exemplary embodiment, the offloading optimization computing task is a deterministic, rule-based algorithm. In another exemplary embodiment, the offloading optimization computing task is a vehicle-specific offloading machine learning model. Examples of vehicle-specific offloading machine learning models are discussed in a U.S. application filed on September 6, 2023, titled "ALLOCATING COMPUTING RESOURCES FOR A VEHICLE APPLICATION", application number No. 18 / 461,682, the entire content of which is incorporated herein by reference.

[0087] Within the scope of the present disclosure, communication tasks include, for example, sending and / or receiving of data. In a non-limiting example, the communication tasks include server uplink communication tasks. Within the scope of the present disclosure, server uplink communication tasks include uploading data to and / or downloading data from the server system 30. For example, server uplink communication tasks can include uploading telemetry data from one or more of the first and second follower vehicles 12b, 12c. In another example, server uplink communication tasks can include downloading OTA updates for one or more of the first and second follower vehicles 12b, 12c. After block 116, method 100 proceeds to block 118.

[0088] At block 118, the first leader vehicle 12a performs the follower data processing task received at block 116. In one exemplary embodiment, to perform the communication task, the vehicle controller 20 of the first leader vehicle 12a uses the vehicle communication system 24 to establish a second wireless connection between the first leader vehicle 12a and the server system 30. In one exemplary embodiment, the second wireless connection is a wide area connection, such as a cellular data connection or the like. It should be understood that the second wireless connection can utilize any connection protocol configured for medium-range and / or long-range connections between devices. After establishing the second wireless connection, the first leader vehicle 12a uploads to perform the communication task (e.g., server uplink communication task), as described above. After block 118, method 100 proceeds to the standby state of block 120.

[0089] In one exemplary embodiment, method 100 repeatedly exits the standby state of block 120 and restarts method 100 at block 102. In a non-limiting example, method 100 is restarted on a timer, such as every three hundred milliseconds.

[0090] Reference Figure 3 , a flowchart of a method 300 for reducing obstacles is shown. Method 300 begins at block 302 and proceeds to block 304. At block 304, an obstacle that impedes transmission between the first and second follower vehicles 12b, 12c and the first leader vehicle 12a is identified. Within the scope of the present disclosure, an obstacle includes any condition that impedes transmission, including, for example, physical obstacles (e.g., an intruding vehicle, a building, etc.), electromagnetic obstacles (e.g., network / wireless interference), hardware obstacles (e.g., device malfunction / failure), etc. In one exemplary embodiment, an obstacle is identified based on network transmission metrics, which include, for example, an increase in the number of discarded data packets, an increase in network latency, a decrease in transmission speed, a decrease in transmission bandwidth, etc. If no obstacle is identified at block 304, method 300 proceeds to the standby state of block 306. If an obstacle is identified at block 304, method 300 advances to block 308.

[0091] At block 308, in a first exemplary embodiment, a follower data processing task is relayed from one of the first and second follower vehicles 12b, 12c to the other of the first and second follower vehicles 12b, 12c (i.e., from the first follower vehicle through the second follower vehicle 12c or from the second follower vehicle 12c through the first follower vehicle 12b) to the first leader vehicle 12a to avoid the obstacle identified at block 304. In one exemplary embodiment, direct communication between the first follower vehicle 12b and the first leader vehicle 12a is blocked, but direct communication between the first follower vehicle 12b and the second follower vehicle 12c is not blocked. Additionally, communication between the second follower vehicle 12c and the first leader vehicle 12a is not blocked. Thus, the follower data processing task can be relayed from the first follower vehicle 12b to the second follower vehicle 12c and then from the second follower vehicle 12c to the first leader vehicle 12a to avoid the obstacle.

[0092] In another exemplary embodiment, if no other follower vehicle is available to relay the follower data processing task, the first leader vehicle 12a terminates communication with the obstructed follower vehicle (i.e., one of the first and second follower vehicles 12b, 12c) in response to identifying the obstacle. After block 308, method 300 proceeds to the standby state of block 306.

[0093] In one exemplary embodiment, method 300 is periodically performed by one or more of the plurality of vehicles 12 before, during, and / or after performing the above-described method 100. In a non-limiting example, method 300 repeatedly exits the standby state of block 306 and restarts method 300 at block 302. In a non-limiting example, method 300 is restarted on a timer, for example, every three hundred milliseconds.

[0094] Reference Figure 4 , a flowchart of a method 400 for selecting a representative leader vehicle is shown. Method 400 begins at block 402 and proceeds to block 404. At block 404, in a first exemplary embodiment, the data processing task capabilities of the first leader vehicle 12a are re-evaluated as discussed above with reference to block 104. In a second exemplary embodiment, the power supply capabilities of the first leader vehicle 12a over time are evaluated. In a non-limiting example, a decrease in power supply capabilities is determined. Within the scope of the present disclosure, a decrease in power supply capabilities is a reduction in the percentage of the initial supplied power that the first leader vehicle 12a supplies over a predetermined time period (e.g., one minute). In a third exemplary embodiment, the remaining energy budget of the first leader vehicle 12a is evaluated. In a non-limiting example, the energy budget is the total amount of energy consumption allowed over a given time period.

[0095] In an exemplary embodiment, if the data processing task capability is greater than or equal to a predetermined data processing task capability threshold, or if the power supply capability decrease is less than a predetermined power supply capability decrease threshold, or if the remaining energy budget is greater than a predetermined remaining energy budget threshold, then method 400 proceeds to the standby state at entry block 406. If the data processing task capability is less than the predetermined data processing task capability threshold, or the power supply capability decrease is greater than or equal to the predetermined power supply capability decrease threshold, or the remaining energy budget is less than or equal to the predetermined remaining energy budget threshold, then method 400 proceeds to block 408.

[0096] At block 408, the follower data processing task is transferred to a second leader vehicle. In an exemplary embodiment, the second leader vehicle is selected from the leader matrix determined at block 112. In an exemplary embodiment, the second leader vehicle is selected as the candidate vehicle with the highest rank in the leader matrix other than the first leader vehicle 12a. In a non-limiting example, the second leader vehicle is selected as one of the first follower vehicle 12b and the second follower vehicle 12c. In an exemplary embodiment, a local connection (such as a Bluetooth connection, a Near Field Communication (NFC) connection, a Wireless Local Area Network (WLAN / WiFi) connection, a short-range radio connection, etc.) is used to transfer the follower data processing task to the second leader vehicle. After block 408, method 400 proceeds to the standby state at entry block 406.

[0097] In an exemplary embodiment, method 400 is periodically executed by one or more of the plurality of vehicles 12 before, during, and / or after performing the above-described method 100. In a non-limiting example, method 400 repeatedly exits the standby state at block 406 and restarts method 400 at block 402. In a non-limiting example, method 400 is restarted on a timer, such as every three hundred milliseconds.

[0098] The systems 10 and methods 100, 300, 400 of the present disclosure provide several advantages. For example, a device with reduced communication capabilities (e.g., a device with only short-range communication capabilities) can establish communication with the server system 30 via a leader device. Additionally, a device with reduced computing capabilities can transfer a computing task to a leader device with enhanced computing capabilities. Further, a task can be transferred to a leader device with a more efficient, reliable, and / or higher-performance connection to the server system 30, resulting in improved performance and reduced resource usage.

[0099] The description of the present disclosure is merely exemplary in nature, and variations that do not depart from the gist of the present disclosure are intended to fall within the scope of the present disclosure. Such variations should not be regarded as departing from the scheme and scope of the present disclosure.

Claims

1. A method for network resource optimization, the method comprising: Selecting a first pilot device from a plurality of wireless devices in wireless communication, wherein the first pilot device wirelessly communicates with at least one follower device among the plurality of wireless devices; Transferring a follower data processing task from the at least one follower device to the first pilot device, wherein the follower data processing task includes at least one of a computing task and a communication task; And Using the first pilot device to execute the follower data processing task.

2. The method according to claim 1, wherein Selecting the first pilot device further includes: Determining a pilot matrix, wherein the pilot matrix includes one or more candidate pilot devices, wherein each of the one or more candidate pilot devices is one of the plurality of wireless devices, and wherein the pilot matrix is determined at least in part based on at least one of the data processing task capabilities of each of the plurality of wireless devices and the historical performance reliability of each of the plurality of wireless devices; and Selecting the first pilot device from the pilot matrix.

3. The method according to claim 2, wherein Determining the pilot matrix further includes: Determining the data processing task capabilities of each of the plurality of wireless devices, wherein a capability estimation machine learning model is used to determine the data processing task capabilities of each of the plurality of wireless devices, and wherein the data processing task capabilities include at least one of computing capabilities and communication capabilities; Determining the historical performance reliability of each of the plurality of wireless devices; Selecting the one or more candidate pilot devices from the plurality of wireless devices at least in part based on the data processing task capabilities of each of the plurality of wireless devices and the historical performance reliability of each of the plurality of wireless devices; and Determining the pilot matrix at least in part based on the one or more candidate pilot devices.

4. The method according to claim 3, wherein Determining the pilot matrix further includes: Determining the pilot matrix, wherein the pilot matrix is a sorted list including each of the one or more candidate pilot devices, and wherein the ranking order of the pilot matrix is determined at least in part based on the data processing task capabilities and the historical performance reliability of each of the one or more candidate pilot devices.

5. The method according to claim 4, wherein, Selecting the first pilot device from the pilot matrix further includes: Determining the planned data processing task capabilities of each of the one or more candidate pilot devices, wherein a capability prediction machine learning model is used to determine the planned data processing task capabilities of each of the one or more candidate pilot devices; and Determining the ranking order of the pilot matrix at least in part based on the planned data processing task capabilities of each of the one or more candidate pilot devices.

6. The method according to claim 4, wherein Determining the pilot matrix further includes: Calculating the cost of each of the one or more candidate pilot devices using a cost function: t c = (t0 * z) + P0 where t c is the cost, t0 is the number of interactions allowed within a given time period without exceeding the device's low power limit, z is the total cost of performing t0 interactions within a given time period, P0 is the cost per unit of power consumed within a given time period, r c,Mb is the cost of transmitting 1 megabyte of registration information within a given time period, c c,Mb is the cost of transmitting 1 megabyte of payload data within a given time period, i is the number of interactions performed within a given time period, and ∫i(t)dt is the integral of i evaluated over multiple time periods; and Determining the ranking order of the pilot matrix at least in part based on the cost of each of the one or more candidate pilot devices.

7. The method according to claim 1, wherein, Transferring the follower data processing task further includes: Transferring the computing task from the at least one follower device to the first leader device, where the computing task includes an offloading optimization computing task; and Transferring the communication task from the at least one follower device to the first leader device, where the communication task includes a server uplink communication task.

8. The method according to claim 7, wherein Transferring the follower data processing task further includes: Identifying an obstacle that hinders the transmission between the first follower device and the first leader device; and Relaying the follower data processing task from the first follower device to the first leader device through a second follower device to avoid the obstacle.

9. The method according to claim 7, wherein, Transferring the computing task further includes: Transferring the offloading optimization computing task from the at least one follower device to the first leader device, where the offloading optimization computing task includes executing a device-specific offloading machine learning model.

10. The method according to claim 1, wherein, Executing the follower data processing task further includes: Determining the data processing task capability of the first leader device; and Transferring the follower data processing task to a second leader device at least partially based on the data processing task capability of the first leader device.

Images