A multi-beam satellite service calculation method and system based on joint allocation of transmission and calculation resources
By jointly allocating computing resources in a multi-beam satellite communication system, the problem of unbalanced resource allocation is solved, dynamic adjustment and optimization of resources is realized, the efficiency of satellite communication and computing is improved, and the energy consumption and time of terminals are reduced. Delay.
Patent Information
- Application Number
- CN202210884957.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-26
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-07-26
AI Technical Summary
In the existing multi-beam satellite communication system, the resource allocation algorithm fails to make full use of the space-based computing platform, resulting in large amounts of data transmission, long time delays, high terminal energy consumption, and unbalanced communication resource use, which cannot effectively realize transmission and computing resources. Joint optimization.
A multi-beam satellite service calculation method based on the joint allocation of computing resources is proposed. Through satellites, the computing task demand information is received, the resource list is generated, and the uplink and downlink wireless resources and processor computing resources are jointly allocated, real-time resource scheduling and unloading is realized. Strategy, dynamically adjust beam resources, adapt to network traffic and computing load changes.
It improves the utilization rate of satellite resources, reduces the delay and energy consumption of terminal services, extends the battery life, and realizes joint optimization and balanced use of resources.
Smart Images

Figure CN115396952B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a multi-beam satellite service calculation method and system based on joint allocation of transmission and calculation resources, belonging to the technical field of satellite communication. Background Art
[0002] In the current satellite system, the satellite forwards the data obtained by the terminal or the data generated by itself to the ground cloud platform for processing. With the development of technology and the continuous improvement of demand, the amount of data generated by sensors is increasing, and the data processing calculation is becoming more and more complex. In order to achieve more accurate information extraction, the calculation dependence is becoming more and more serious. In addition, the services provided by satellites are highly diverse, the types and demands are diverse, and the time and space distribution are seriously uneven, resulting in extremely unbalanced use of communication resources. Some beams are running at high load, and some beams are idle. This puts high demands on the efficient use of satellite network transmission and computing resources. In the traditional multi-beam satellite communication system, the power and frequency resources allocated to each beam are fixed and only part of the total satellite resources. In addition, the existing resource allocation algorithm is mainly based on data communication. During the data transmission process, the data content is not considered, the data is not processed, and the computing resources of the space-based computing platform are not fully utilized, resulting in a large amount of data transmission, extended transmission time, high terminal energy consumption, and extended business processing time. In the new generation of satellite communication systems, satellites have on-orbit processing capabilities. The time and energy consumption cost of satellite terminal services are greatly related to terminal resource allocation. At present, satellite communication resource allocation and computing resource allocation are independent, and data processing delay and energy consumption are not considered together with communication delay and energy consumption. Therefore, how to jointly consider limited space-based transmission resources, computing resources, and power resources to achieve the transmission and processing of terminal services has become an urgent problem that needs to be solved. Summary of the invention
[0003] The technical problem solved by the present invention is: in view of the deficiencies existing in the current prior art, a multi-beam satellite service calculation method and system based on the joint allocation of transmission and computing resources are proposed, which can realize the real-time allocation of payload resources, adapt to changes in network traffic and node computing load, and improve the utilization rate of satellite resources.
[0004] The present invention solves the above technical problems by the following technical solutions:
[0005] A multi-beam satellite service calculation method based on joint allocation of transmission and calculation resources, comprising:
[0006] The satellite receives computing task demand information sent by all terminals within a scheduling period T, generates a service application list, and generates a satellite available resource list based on available bandwidth resources and processor computing resources;
[0007] Perform joint allocation of transmission and computing resources based on the available resource list and service application list, and determine the final offloading terminal list and the final wireless resource allocation plan;
[0008] Transmitting the final unloading terminal list and the final wireless resource allocation plan to each terminal;
[0009] Perform resource scheduling according to the final wireless resource allocation plan;
[0010] Generate and update the satellite available resource list in real time based on the resources occupied by the satellite's scheduled tasks. Each terminal transmits data or performs data calculation in each terminal based on the satellite's resource scheduling results.
[0011] After completing the data calculation task, updating the satellite available resource list;
[0012] The data calculation results are transmitted from the satellite to each terminal to complete the business calculation of the current scheduling cycle and enter the next scheduling cycle.
[0013] The joint allocation of the computing resources is specifically as follows:
[0014] According to the available resource list, uplink wireless resources, downlink wireless resources, and processor computing resources are obtained and jointly allocated. The uplink wireless resources include the bandwidth and transmission power of each terminal connected to the satellite. The downlink wireless resources include the bandwidth of each terminal connected to the satellite and the available time of each beam received by the satellite, and the unloading strategy is executed.
[0015] The uninstallation strategy is as follows:
[0016] Calculate the satellite system cost when each terminal is unloaded individually;
[0017] The satellite system costs when each terminal is unloaded individually are sorted from small to large;
[0018] Create an uninstall terminal list s And add a null initial value, the corresponding initial satellite system cost is the system cost when all terminals are not uninstalled, and add the uninstall terminal list U one by one according to the sorted satellite system cost s , each time a new terminal and the corresponding satellite system cost are added, the uninstall terminal list U is re- s Allocate wireless resources to all terminals.
[0019] Re-uninstall terminal list U s The specific steps of allocating wireless resources to all terminals include:
[0020] Re-allocate the on-board computing resources, the available time of each beam received by the satellite, the bandwidth of each terminal, and the transmit power of each terminal;
[0021] According to the re-allocation, the satellite system cost is recalculated. When the satellite system cost of each terminal is no longer reduced, the unloaded terminal list U s The last added terminal is deleted, and a final uninstallation terminal list is obtained after the deletion, and the wireless resource allocation scheme corresponding to the final uninstallation terminal list is used as the final wireless resource allocation scheme.
[0022] The computing task requirement information includes:
[0023] Data location L u , used to describe the data location, including on-board data and terminal local data;
[0024] Amount of data to be calculated The program that performs this calculation u , the amount of program data The amount of data after calculation The number of floating point operations c required for the calculation task u , local computing energy consumption e u , the terminal's maximum data transmission transmission power P u , the computing resources of each terminal are F u Number of floating point operations / s, P s Satellite downlink transmission power per beam, where:
[0025] Local computing energy consumption u The average energy consumption of the actual test in the same computing environment is obtained. The business application list is The satellite available resource list includes {available bandwidth resources, available computing resources}.
[0026] Calculate the amount of data to be transmitted for the task d u Including the amount of data to be calculated for the task and mission program data Mission program data, when the mission program data is already stored on the satellite, the amount of data to be transmitted u Only the amount of data to be processed is included
[0027] The calculation method of satellite system cost S when each terminal is unloaded separately is:
[0028] It is determined according to the terminal service completion time and the terminal service completion energy consumption. The specific formula is:
[0029]
[0030] Where, T u is the terminal service completion time, E u is the energy consumption of the terminal business, α is the optimal value, if
[0031] If only time is optimized, α=1; if only energy consumption is optimized, α=0.
[0032] The terminal service completion time is determined according to the service completion delay of the current terminal, where:
[0033] When the data required for the computing task is stored locally in the terminal, if the computing task is calculated on the satellite, the service completion delay is determined as the sum of the uplink transmission delay, the on-satellite computing delay, and the downlink transmission delay of the computing result, where:
[0034] The uplink transmission delay is The on-board computing delay is in is the computing resource allocated by the satellite to terminal u, and the downlink transmission delay is The ratio of the number of working time slots allocated to the beam where u is located to the total number of time slots, is the uplink transmission rate, is the downlink rate of terminal u; where N0 is the background noise power spectrum density, is the uplink channel gain from terminal u to satellite, is the downlink channel gain from the satellite to the terminal u, including path loss and antenna gain;
[0035] If the computing task is performed locally on the terminal, the service completion delay is c u / F u .
[0036] When the data required for the computing task is stored on the satellite, if the computing task is calculated on the satellite, the service completion delay is the sum of the on-satellite computing delay and the downlink transmission delay of the computing result, where:
[0037] The on-board computing delay is The downlink transmission delay is
[0038] If the computing task is calculated locally on the terminal, the service completion delay is the sum of the downlink transmission delay and the local computing delay. The downlink transmission delay is The local computing delay is c u / F u .
[0039] The calculation method for terminal service completion time is:
[0040]
[0041] The calculation method for energy consumption of terminal business completion is:
[0042]
[0043] In the formula, x u∈{0,1},x u =1 for on-board calculation, x u = 0, it is calculated locally at the terminal, L u ∈{0,1},L u = 1, the data to be calculated is located on the satellite, L u = 0, the data to be processed is located in the terminal. u Calculate energy consumption locally.
[0044] When allocating on-board computing resources, according to the uninstall terminal list U S , the computing resources allocated to each terminal in the list are Determine the calculation delay of the terminal as The objective function is to minimize the sum of computing delays of all terminal services in the unloaded terminal list, and the constraint condition is that the sum of computing resources allocated to all terminals is less than the total computing resources of the satellite. The value of computing resources allocated to each terminal is calculated by a convex optimization method.
[0045] In the downlink beam output by the satellite, the time slot and terminal bandwidth of the downlink beam are allocated, where:
[0046] Determine the number of beams that can work simultaneously in the downlink beam, and determine the residence time of each beam position in the current scheduling cycle as Calculate d′ u For the data to be transmitted by the satellite to terminal u, all beams with a beam interference distance less than R that have not yet been allocated bandwidth resources are set as set U, and the available bandwidth for service transmission is determined as B. The terminal u task can be allocated to use d′ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated.
[0047] In the uplink beam received by the satellite, the time slot and terminal bandwidth of the uplink beam are allocated, where:
[0048] Set all the beams with a distance less than R from the beam interference that have not yet been allocated bandwidth resources as set U, determine the available bandwidth for service transmission as B, and the terminal u task can be allocated to use d″ u is the data to be transmitted to the satellite by terminal u, d″ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated.
[0049] The method for allocating the transmit power of each terminal is as follows:
[0050] First determine and Then, we solve the nonlinear equations by bisection method to obtain p u ;
[0051] The minimization problem is solved with the goal of optimizing the system cost S. The transmit power budget of each terminal cannot exceed its maximum transmit power capability. The transmit power p is determined by the Lagrange multiplier method. u The nonlinear equations are divided into two groups, and the computing resources allocated to each terminal are Terminal u tasks can be assigned using Calculate the transmit power p corresponding to each terminal by binary search u , where: p u is the transmission power of data sent when terminal u unloads the task, 0<p u ≤P u When the task is not unloaded u =0.
[0052] A multi-beam satellite service computing system for joint scheduling of transmission and computing resources, used to implement a multi-beam satellite service computing method, comprising:
[0053] Multi-beam antenna unit, multi-rate digital processing unit, computing unit, management control unit, storage unit, wherein:
[0054] The multi-beam antenna unit is composed of a multi-beam antenna in the form of a phased array or a reflector, and is used for signal transmission and reception;
[0055] A multi-rate digital processing unit adjusts the communication rate according to the bandwidth and transmission power of the terminal;
[0056] The computing unit manages the loading of each terminal according to the computing tasks;
[0057] Management control unit, used for terminal management, resource allocation, and computing program loading management;
[0058] The storage unit is used to store application programs and business data. Newly added application programs are transmitted and stored by the terminal.
[0059] The data exchange unit realizes high-speed data interaction and sharing among the multi-rate digital processing unit, storage unit, computing unit, and management control unit. The interactive data includes business data, program data, and control data.
[0060] The management and control unit includes a task parsing unit, a resource allocation and scheduling unit, and a computing program management unit;
[0061] The task analysis unit analyzes the computing tasks by information type, computing energy consumption and computing time;
[0062] The resource allocation and scheduling unit determines the calculation method according to the calculation complexity and calculation time, and sends the calculation task information data to the corresponding chip of the calculation unit;
[0063] The computing program management unit is used to load, unload, and monitor the operation of the terminal control program, and to apply for and release satellite computing resources.
[0064] The advantages of the present invention compared with the prior art are:
[0065] (1) The present invention provides a multi-beam satellite service calculation method and system based on joint allocation of transmission and computing resources. Satellite beams can cover areas with sparse traffic as well as areas with dense data. This results in extremely unbalanced use of satellite transmission and computing resources, with some beams running at high loads and some beam communication resources idle. This patent proposes a joint beam coverage and computing strategy based on traffic load, computing load, etc., allocates load resources in real time, can adapt to changes in network traffic and node computing load, and adaptively formulates service unloading strategies and beam hopping coverage strategies to improve satellite resource utilization;
[0066] (2) The method of the present invention jointly considers the satellite multi-beam communication load and computing load, and performs differentiated allocation of beam bandwidth according to the network status. Compared with the existing satellite system's four-color or seven-color multiplexing and non-joint allocation, it realizes the joint allocation of spatial coverage, frequency bandwidth, computing resources, and power resources, and can dynamically adjust the resources between beams. Under the same service density conditions, the bandwidth demand of the satellite network can be reduced. In addition, the present invention aims to reduce terminal power consumption and service completion time, and performs joint scheduling of satellite and terminal computing resources, which can effectively reduce the service delay and power consumption of satellite terminals and extend the service life of satellite terminal batteries. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] Figure 1 A flowchart of the steps of the business calculation method provided for the invention;
[0068] Figure 2 A schematic diagram of the unit composition of the business processing system provided for the invention; DETAILED DESCRIPTION
[0069] A multi-beam satellite service calculation method and system based on joint allocation of transmission and computing resources, which performs differentiated allocation of beam bandwidth according to network status, realizes joint allocation of spatial coverage, frequency bandwidth, computing resources, and power resources compared with four-color or seven-color multiplexing and non-joint allocation of existing satellite systems, can dynamically adjust resources between beams, can reduce bandwidth requirements of satellite networks, effectively reduce service delays and energy consumption of satellite terminals, and delay the use time of satellite terminal batteries. The specific method steps are as follows:
[0070] The satellite receives computing task demand information sent by all terminals within a scheduling period T, generates a service application list, and generates a satellite available resource list based on available bandwidth resources and processor computing resources;
[0071] Perform joint allocation of transmission and computing resources based on the available resource list and service application list, and determine the final offloading terminal list and the final wireless resource allocation plan;
[0072] Transmitting the final unloading terminal list and the final wireless resource allocation plan to each terminal;
[0073] Perform resource scheduling according to the final wireless resource allocation plan;
[0074] Generate and update the satellite available resource list in real time based on the resources occupied by the satellite's scheduled tasks. Each terminal transmits data or performs data calculation in each terminal based on the satellite's resource scheduling results.
[0075] After completing the calculation task, update the satellite's available resource list;
[0076] The calculation results of the calculation tasks are transmitted from the satellite to each terminal, completing the business calculation of the current scheduling cycle and entering the next scheduling cycle.
[0077] The specific joint allocation of computing resources is as follows:
[0078] Jointly allocate uplink wireless resources, downlink wireless resources, and computing resources. Uplink wireless resources include bandwidth and transmit power of each terminal connected to the satellite. Downlink wireless resources include bandwidth of each terminal connected to the satellite and available time of each beam received by the satellite, and execute unloading strategies.
[0079] The uninstallation strategies are as follows:
[0080] Calculate the satellite system cost when each terminal is unloaded individually;
[0081] The satellite system costs when each terminal is unloaded individually are sorted from small to large;
[0082] Create an uninstall terminal list s And add a null initial value, the corresponding initial satellite system cost is the system cost when all terminals are not uninstalled, and add the uninstall terminal list U one by one according to the sorted satellite system cost s , each time a new terminal and the corresponding satellite system cost are added, the uninstall terminal list U is re- s Allocate wireless resources to all terminals;
[0083] Re-uninstall terminal list U s The specific steps of allocating wireless resources to all terminals include:
[0084] Re-allocate the on-board computing resources, the available time of each beam received by the satellite, the bandwidth of each terminal, and the transmit power of each terminal;
[0085] According to the re-allocation, the satellite system cost is recalculated. When the satellite system cost of each terminal is no longer reduced, the unloaded terminal list U s Delete the last added terminal, obtain the final uninstall terminal list after deletion, and use the wireless resource allocation scheme corresponding to the final uninstall terminal list as the final wireless resource allocation scheme;
[0086] The computing task requirement information includes:
[0087] Data location L u , used to describe the data location, including on-board data and terminal local data;
[0088] Amount of data to be calculated bits, the program i that performs the calculation u , the amount of program data bits, the amount of data after calculation bits, the number of floating-point operations c required for the calculation task u , local computing energy consumption e u , the terminal is the most
[0089] Large data transmission transmission power P u , the computing resources of each terminal are F u Number of floating point operations / s, P s Satellite downlink transmission power per beam, where:
[0090] Local computing energy consumption u The average energy consumption of the actual test in the same computing environment is obtained. The business application list is The satellite available resource list includes {available bandwidth resources, available computing resources};
[0091] Calculate the amount of data to be transmitted for the task d u Including the amount of data to be calculated for the task and mission program data Mission program data, when the mission program data is already stored on the satellite, the amount of data to be transmitted u Only the amount of data to be processed is included
[0092] The calculation method of satellite system cost S when each terminal is unloaded separately is:
[0093] It is determined according to the terminal service completion time and the terminal service completion energy consumption. The specific formula is:
[0094]
[0095] Where, T u is the terminal service completion time, E u is the energy consumption of the terminal service, α is the optimization value, if only the time is optimized, α=1, if only the energy consumption is optimized, α=0;
[0096] The terminal service completion time is determined according to the service completion delay of the current terminal, where:
[0097] When the data required for the computing task is stored locally in the terminal, if the computing task is calculated on the satellite, the service completion delay is determined as the sum of the uplink transmission delay, the on-satellite computing delay, and the downlink transmission delay of the computing result, where:
[0098] The uplink transmission delay is The on-board computing delay is in is the computing resource allocated by the satellite to terminal u, and the downlink transmission delay is The ratio of the number of working time slots allocated to the beam where u is located to the total number of time slots, is the uplink transmission rate, is the downlink rate of terminal u; where N0 is the background noise power spectrum density, is the uplink channel gain from terminal u to satellite, is the downlink channel gain from the satellite to the terminal u, including path loss and antenna gain;
[0099] If the computing task is performed locally on the terminal, the service completion delay is c u / F u ;
[0100] When the data required for the computing task is stored on the satellite, if the computing task is calculated on the satellite, the service completion delay is the sum of the on-satellite computing delay and the downlink transmission delay of the computing result, where:
[0101] The on-board computing delay is The downlink transmission delay is
[0102] If the computing task is calculated locally on the terminal, the service completion delay is the sum of the downlink transmission delay and the local computing delay. The downlink transmission delay is The local computing delay is c u / F u ;
[0103] The calculation method for terminal service completion time is:
[0104]
[0105] The calculation method for energy consumption of terminal business completion is:
[0106]
[0107] In the formula, x u ∈{0,1},x u =1 for on-board calculation, x u = 0, it is calculated locally at the terminal, L u ∈{0,1},L u = 1, the data to be calculated is located on the satellite, L u = 0, the data to be processed is located in the terminal. u Calculate energy consumption for local areas;
[0108] When allocating on-board computing resources, according to the uninstall terminal list U S , the computing resources allocated to each terminal in the list are Determine the calculation delay of the terminal as Taking the minimization of the sum of computing delays of all terminal services as the objective function and the constraint that the sum of computing resources allocated to all terminals is less than the total computing resources of the satellite, the value of computing resources allocated to each terminal is calculated through the convex optimization method to complete the joint allocation;
[0109] In the downlink beam output by the satellite, the time slot and terminal bandwidth of the downlink beam are allocated, where:
[0110] Determine the number of beams that can work simultaneously in the downlink beam, and determine the residence time of each beam position in the current scheduling cycle as Calculate d′ u For the data to be transmitted by the satellite to terminal u, all beams with a beam interference distance less than R that have not yet been allocated bandwidth resources are set as set U, and the available bandwidth for service transmission is determined as B. The terminal u task can be allocated to use d′ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated;
[0111] In the uplink beam received by the satellite, the time slot and terminal bandwidth of the uplink beam are allocated, where:
[0112] Set all the beams with a distance less than R from the beam interference that have not yet been allocated bandwidth resources as set U, determine the available bandwidth for service transmission as B, and the terminal u task can be allocated to use d″ u is the data to be transmitted to the satellite by terminal u, d″ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated.
[0113] The method for allocating the transmit power of each terminal is as follows:
[0114] The minimization problem is solved with the goal of optimizing the system cost S. The transmit power budget of each terminal cannot exceed its maximum transmit power capability. The transmit power p is determined by the Lagrange multiplier method. u The nonlinear equations are divided into two groups, and the computing resources allocated to each terminal are Terminal u tasks can be assigned using Calculate the transmit power p corresponding to each terminal by binary search u ,in:
[0115] p u is the transmission power of data sent when terminal u unloads the task, 0<p u ≤P u When the task is not unloaded u =0.
[0116] A multi-beam satellite service computing system, comprising a multi-beam antenna unit, a multi-rate digital processing unit, a computing unit, a management control unit, and a storage unit, wherein:
[0117] The multi-beam antenna unit is composed of a multi-beam antenna in the form of a phased array or a reflector, and is used for signal transmission and reception;
[0118] A multi-rate digital processing unit adjusts the communication rate according to the bandwidth and transmission power of the terminal;
[0119] The computing unit manages the loading of each terminal according to the computing tasks;
[0120] Management control unit, used for terminal management, resource allocation, and computing program loading management;
[0121] The storage unit is used to store application programs and business data. Newly added application programs are transmitted and stored by the terminal.
[0122] The data exchange unit realizes high-speed data interaction and sharing among the multi-rate digital processing unit, storage unit, computing unit, and management control unit. The interactive data includes business data, program data, and control data.
[0123] The management and control unit includes a task analysis unit, a resource allocation and scheduling unit, and a computing program management unit;
[0124] The task analysis unit analyzes the computing tasks by information type, computing energy consumption and computing time;
[0125] The resource allocation and scheduling unit determines the calculation method according to the calculation complexity and calculation time, and sends the calculation task information data to the corresponding chip of the calculation unit;
[0126] The computing program management unit is used to load, unload, and monitor the operation of the terminal control program, and to apply for and release satellite computing resources.
[0127] The following is further described based on specific embodiments:
[0128] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0129] The method steps in the embodiments of the present application are described in detail below with reference to the accompanying drawings.
[0130] Example:
[0131] Terminals include data collection terminals, vehicles, aircraft, ships, buoys, etc. The computing services applied by the terminals include making quick decisions based on the collected data, data analysis, etc.
[0132] If the business to be processed is generated by a terminal, the data storage location is located at the terminal, such as data collected by a data collection terminal and data generated by a vehicle sensor.
[0133] If the service to be processed is generated by a satellite, the data storage location is on the satellite, such as data generated by a satellite with an optical payload or a SAR payload.
[0134] Business computing programs can be uploaded to the satellite as needed. The satellite stores the business computing programs and maintains a list of business computing programs. Business computing programs are based on CPU, FPGA or GPU, and can be loaded, unloaded and configured.
[0135] The satellite is a multi-beam satellite that covers different areas through multiple beams. Due to the limitations of satellite weight, volume and energy consumption, the satellite's transmit power is limited. Under the condition of multi-beam large bandwidth, it is difficult for all downlink beams to handle the working status. Therefore, the downlink adopts beam hopping technology. For example, for a satellite with a total of 19 beams, there can be 10 beams working simultaneously in the downlink. The satellite's transmit power limitation is met by making different beams work in different time slots.
[0136] The terminal computing power is set to 1G flops, the satellite capacity is set to 10G flops, the amount of data to be processed is 30-130Mbit, and the amount of data to be processed is 3-13G flops. The number of terminals in each beam varies. The satellite operates in the Ka band, and the total available bandwidth for both uplink and downlink is 2GHz.
[0137] The implementation process of this patent is as follows:
[0138] (1) The service requested by terminal u is a data computing task. There are two cases according to the different data storage locations: one is that the service data to be calculated is located locally in the terminal, and the other is that the service data to be processed is located in the satellite connected to the satellite. The terminal needs to obtain the result of data processing. The terminal reports its task computing demand information to the satellite it is connected to. Suppose the task computing demand information includes the data location L u , used to describe whether the data is located locally or on the satellite, and the amount of data to be calculated bits, the program i that performs the calculation u and program data volume bits, the amount of data after calculation bits, the number of floating-point operations c required for this task u , local computing energy consumption e u , the terminal's maximum data transmission transmission power P u , the computing resources provided by the terminal are F u Number of floating point operations / s; P s Satellite downlink transmission power per beam.
[0139] (2) The satellite collects the computing tasks of all the terminals connected to its own beam within a scheduling period T and forms a service application queue The satellite monitors available bandwidth resources and processor computing resources, and forms an available resource list {available bandwidth resources, available computing resources};
[0140] (3) Based on the available resource list and the service application list, terminal service offloading decisions are made and transmission and computing resources are jointly allocated to resources. The joint allocation of transmission and computing resources includes the joint allocation of uplink wireless resources, downlink wireless resources, and computing resources. The uplink wireless resources include the bandwidth and transmit power allocated to each terminal, the downlink wireless resources include the bandwidth of each terminal and the available time of each beam, and the computing offloading strategy for determining the computing location of the terminal task. The specific process is as follows:
[0141] 31) Calculate the system cost when each terminal is uninstalled separately
[0142] For services where data is local to the terminal, if the service is calculated on the satellite, the service completion delay is the sum of the uplink transmission delay, the on-satellite calculation delay, and the downlink transmission delay of the calculation result. The uplink transmission delay is The on-board computing delay is The downlink transmission delay is in The ratio of the number of working time slots allocated to the beam where u is located to the total number of time slots, is the uplink transmission rate. is the downlink rate of terminal u; if the service is calculated locally at the terminal, the service completion delay is c u / F u .
[0143] Among them, N0 is the background noise power spectrum density, is the uplink channel gain from terminal u to satellite, is the downlink channel gain from satellite to terminal u, including path loss and antenna gain, p u is the transmission power of data sent when terminal u unloads the task, 0<p u ≤P u When the task is not unloaded u =0.
[0144] For services with data on satellites, if the services are calculated on board, the service completion delay is the sum of the on-board calculation delay and the downlink transmission delay of the calculation result. The downlink transmission delay is If the service is calculated locally at the terminal, the service completion delay is the sum of the downlink transmission delay and the local calculation delay. The downlink transmission delay is The local computing delay is c u / F u .
[0145] The terminal service completion time can be expressed as
[0146]
[0147] The energy consumption of terminal business completion can be expressed as
[0148]
[0149] where x u ∈{0,1},x u =1 for on-board calculation, x u = 0, it is calculated locally at the terminal, L u ∈{0,1},L u = 1, the data to be processed is located on the satellite, L u =0, the data to be processed is located locally in the terminal.
[0150] System cost Take α∈[0,1], where the value of α is a decimal between 0 and 1 according to the system optimization goal. If only latency is optimized, take α=1, and if only energy consumption is optimized, take α=0.
[0151] 32) Sort the terminals according to the system cost of uninstalling each terminal individually from small to large
[0152] 33) Uninstall Terminal List U s Initially empty, the initial system cost is the system cost when all terminals are not uninstalled, and the terminals are added to the uninstall terminal list U one by one in the order determined in 32). s , uninstall terminal list U s Every time a new terminal is added, the uninstall terminal list is re-checked. s Allocate wireless resources to all terminals in the system, including:
[0153] a) On-board computing resource allocation
[0154] Suppose the list of unloading terminals performing task calculations on the satellite is U s , the computing resources allocated to each terminal u in the set are The calculation delay of the terminal service is Take U s The objective function is to minimize the sum of the computing delays of all terminal services, and the constraint condition is that the sum of the computing resources allocated to all terminals is less than the total computing resources of the satellite. The convex optimization problem solution is used to solve the minimization problem, and the allocation of each terminal can be calculated. Specifically, the KKT condition can be used to solve this problem.
[0155] b) Downlink beam available time allocation and terminal bandwidth allocation
[0156] The available time allocation for downlink beam is as follows:
[0157] Assume that there are N = 19 beams in the downlink, and at most K = 10 beams can work simultaneously. The residence time of each beam position b in the scheduling cycle is d′ u It is the data to be transmitted from the satellite to the terminal u.
[0158] The specific method for allocating transmission bandwidth is as follows:
[0159] For the unallocated beams, find all the beams with unallocated bandwidth resources whose interference distance is less than R = 4 beams. Let the terminal set accessing these unallocated bandwidth resources be U, and let the available bandwidth for service transmission be B. The task of terminal u can be allocated to use d′ u =0 o'clock Repeat this process until all terminal resources are allocated.
[0160] d′ u is the data to be transmitted from the satellite to terminal u. If the computing service of terminal u is calculated on the satellite, then d′ u is the amount of data of the calculation result; if the calculation service of terminal u is calculated at the terminal and the data to be calculated is located on the satellite, then d′ u is the amount of data to be calculated. If the computing service of terminal u is calculated at the terminal and the data to be calculated is located at the terminal, then d′ u =0.
[0161] c) Uplink terminal bandwidth and terminal transmit power allocation
[0162] For the uplink beams that have not been allocated, find all the beams that have not been allocated bandwidth resources and whose interference distance with the uplink beam is less than R = 4. Let the terminal set accessing these beams with unallocated bandwidth resources be U, and let the available bandwidth for service transmission be B. The task of terminal u can be allocated to use d″ u is the data to be transmitted to the satellite by terminal u, d″ u =0 o'clock Repeat this process until all terminal resources are allocated.
[0163] The minimization problem is solved with the goal of optimizing the system cost S. The transmit power budget of each terminal cannot exceed its maximum transmit power capability as a constraint. Through the Lagrange multiplier method, we can get the following about p u The nonlinear equations, combined with f u and Using the dichotomy method, we can find p u The value of .
[0164] d) Recalculate the system cost according to the reallocated wireless resource allocation scheme. When the system cost is no longer reduced compared to the original value, delete the last added terminal from the uninstall terminal list. The uninstall terminal list is the final uninstall terminal list, and the wireless resource allocation scheme corresponding to the final uninstall terminal list is the final wireless resource allocation scheme.
[0165] (4) sending the final offloading terminal list and the final wireless resource allocation plan to the terminal via the control channel;
[0166] (5) Perform resource scheduling according to the final wireless resource allocation plan;
[0167] (6) Update the available resource table based on the resources occupied by the satellite's scheduled tasks. The satellite transmits the decision results and the frequency band and power allocation results to the terminal through the downlink. The terminal transmits the calculation data or performs local calculation in the specified frequency band according to the scheduling results.
[0168] (7) After completing the terminal service, release the resources occupied by the terminal and update the satellite available resource list {available bandwidth resources, available computing resources} according to the satellite mission status;
[0169] (8) After the calculation results of the services calculated on the satellite are obtained, the results of the service calculations are transmitted to the service application terminal through the transmission time and bandwidth allocated for downlink;
[0170] (9) After one scheduling cycle ends, the next scheduling cycle begins.
[0171] In one embodiment, a multi-beam satellite service computing system for joint scheduling of computing resources includes: a multi-beam antenna unit, a multi-rate digital processing unit, a computing unit, a management control unit, and a storage unit.
[0172] (1) Multi-beam antenna unit: a multi-beam antenna in the form of a phased array or a reflector to transmit and receive communication signals.
[0173] (2) a multi-rate digital processing unit, based on digital signal processing, to achieve variable communication rates based on bandwidth and transmission power;
[0174] (3) Computing unit, including reconfigurable FPGA, CPU, and GPU. The computing business program is based on containers and implemented in the form of microservices. The compiled application files are pre-stored in the satellite storage unit and loaded from the storage unit according to the business request.
[0175] (4) A management control unit, which is used for terminal management, resource allocation and management, business management, computing program loading management, etc. It further includes a task analysis unit, which is used to analyze the application business information type, computing energy consumption and computing time; a resource allocation and scheduling unit, which is used to determine the computing processing method according to the computing complexity and computing time, and send the application business information data to the corresponding processing chip according to the computing processing method; a computing program management unit, which is used to monitor the loading, unloading and running process of heterogeneous processor programs and apply for and release computing resources.
[0176] (5) A storage unit, used to store application programs and business data. Newly added application programs are transmitted and stored by the terminal.
[0177] (6) Data exchange unit, which realizes high-speed data exchange and sharing among the multi-rate digital processing unit, storage unit, computing unit, and management control unit. The exchanged data includes business data, program data, and control data.
[0178] Although the present invention has been disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art may make possible changes and modifications to the technical solution of the present invention by using the methods and technical contents disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.
[0179] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.
Claims
1. A multi-beam satellite service calculation method based on joint allocation of transmission and calculation resources, characterized in that include: The satellite receives computing task demand information sent by all terminals within a scheduling period T, generates a service application list, and generates a satellite available resource list based on available bandwidth resources and processor computing resources; Perform joint allocation of transmission and computing resources based on the available resource list and service application list, and determine the final offloading terminal list and the final wireless resource allocation plan; Transmitting the final unloading terminal list and the final wireless resource allocation plan to each terminal; Perform resource scheduling according to the final wireless resource allocation plan; Generate and update the satellite available resource list in real time based on the resources occupied by the satellite's scheduled tasks. Each terminal transmits data or performs data calculation in each terminal based on the satellite's resource scheduling results. After completing the data calculation task, updating the satellite available resource list; The data calculation results are transmitted from the satellite to each terminal, completing the business calculation of the current scheduling cycle and entering the next scheduling cycle; The joint allocation of the computing resources is specifically as follows: According to the available resource list, uplink wireless resources, downlink wireless resources, and processor computing resources are obtained and jointly allocated. The uplink wireless resources include the bandwidth and transmission power of each terminal connected to the satellite, and the downlink wireless resources include the bandwidth of each terminal connected to the satellite and the available time of each beam received by the satellite, and the unloading strategy is executed; The uninstallation strategy is as follows: Calculate the satellite system cost when each terminal is unloaded individually; The satellite system costs when each terminal is unloaded individually are sorted from small to large; Create an uninstall terminal list s And add a null initial value, the corresponding initial satellite system cost is the system cost when all terminals are not uninstalled, and add the uninstall terminal list U one by one according to the sorted satellite system cost s , each time a new terminal and the corresponding satellite system cost are added, the uninstall terminal list U is re- s Allocate wireless resources to all terminals; Re-uninstall terminal list U s The specific steps of allocating wireless resources to all terminals include: Re-allocate the on-board computing resources, the available time of each beam received by the satellite, the bandwidth of each terminal, and the transmit power of each terminal; According to the re-allocation, the satellite system cost is recalculated. When the satellite system cost of each terminal is no longer reduced, the unloaded terminal list U s Delete the last added terminal, obtain the final uninstall terminal list after deletion, and use the wireless resource allocation scheme corresponding to the final uninstall terminal list as the final wireless resource allocation scheme; The computing task requirement information includes: Data location L u , used to describe the data location, including on-board data and terminal local data; Amount of data to be calculated The program that performs this calculation u , the amount of program data The amount of data after calculation The number of floating point operations c required for the calculation task u , local computing energy consumption e u , the terminal's maximum data transmission transmission power P u , the computing resources of each terminal are F u Number of floating point operations / s, P s Satellite downlink transmission power per beam, where: Local computing energy consumption u The average energy consumption of the actual test in the same computing environment is obtained. The business application list is The satellite available resource list includes {available bandwidth resources, available computing resources}; Calculate the amount of data to be transmitted for the task d u Including the amount of data to be calculated for the task and mission program data Mission program data, when the mission program data is already stored on the satellite, the amount of data to be transmitted u Only the amount of data to be processed is included The calculation method of satellite system cost S when each terminal is unloaded separately is: Determined according to the terminal service completion time and terminal service completion energy consumption, the specific formula is: Where, T u is the terminal service completion time, E u is the energy consumption of the terminal service, α is the optimization value, if only the time is optimized, α=1, if only the energy consumption is optimized, α=0; The terminal service completion time is determined according to the service completion delay of the current terminal, where: When the data required for the computing task is stored locally in the terminal, if the computing task is calculated on the satellite, the service completion delay is determined as the sum of the uplink transmission delay, the on-satellite computing delay, and the downlink transmission delay of the computing result, where: The uplink transmission delay is The on-board computing delay is in is the computing resource allocated by the satellite to terminal u, and the downlink transmission delay is The ratio of the number of working time slots allocated to the beam where u is located to the total number of time slots, is the uplink transmission rate, is the downlink rate of terminal u; where N0 is the background noise power spectrum density, is the uplink channel gain from terminal u to satellite, is the downlink channel gain from the satellite to the terminal u, including path loss and antenna gain; If the computing task is performed locally on the terminal, the service completion delay is c u / F u ; When the data required for the computing task is stored on the satellite, if the computing task is calculated on the satellite, the service completion delay is the sum of the on-satellite computing delay and the downlink transmission delay of the computing result, where: The on-board computing delay is The downlink transmission delay is If the computing task is calculated locally on the terminal, the service completion delay is the sum of the downlink transmission delay and the local computing delay. The downlink transmission delay is The local computing delay is c u / F u ; The calculation method for terminal service completion time is: The calculation method for energy consumption of terminal business completion is: In the formula, x u ∈{0,1},x u =1 for on-board calculation, x u = 0, it is calculated locally at the terminal, L u ∈{0,1},L u = 1, the data to be calculated is located on the satellite, L u = 0, the data to be processed is located in the terminal; e u Calculate energy consumption for local areas; When allocating on-board computing resources, according to the uninstall terminal list U S , the computing resources allocated to each terminal in the list are Determine the calculation delay of the terminal as The objective function is to minimize the sum of the computing delays of all terminal services in the unloaded terminal list, and the constraint condition is that the sum of the computing resources allocated to all terminals is less than the total computing resources of the satellite. The value of the computing resources allocated to each terminal is calculated using a convex optimization method. In the downlink beam output by the satellite, the time slot and terminal bandwidth of the downlink beam are allocated, where: Determine the number of beams that can work simultaneously in the downlink beam, and determine the residence time of each beam position in the current scheduling cycle as Calculate d u ′ is the data to be transmitted by the satellite to terminal u. All beams with a beam interference distance less than R that have not yet been allocated bandwidth resources are set as set U. The available bandwidth for service transmission is determined as B. The terminal u task can be allocated to use d′ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated; In the uplink beam received by the satellite, the time slot and terminal bandwidth of the uplink beam are allocated, where: Set all the beams with a distance less than R from the beam interference that have not yet been allocated bandwidth resources as set U, determine the available bandwidth for service transmission as B, and the terminal u task can be allocated to use d″ u is the data to be transmitted to the satellite by terminal u, d″ u =0 o'clock According to the calculation results Repeat the calculation until all terminals are allocated; The method for allocating the transmit power of each terminal is as follows: First determine and Then, we solve the nonlinear equations by bisection method to obtain p u ; The minimization problem is solved with the goal of optimizing the system cost S. The transmit power budget of each terminal cannot exceed its maximum transmit power capability. The transmit power p is determined by the Lagrange multiplier method. u The nonlinear equations are divided into two groups, and the computing resources allocated to each terminal are Terminal u tasks can be assigned using Calculate the transmit power p corresponding to each terminal by binary search u , where: p u is the transmission power of data sent when terminal u unloads the task, 0<p u ≤P u ; When the task is not unloaded, p u =0.
2. A multi-beam satellite service computing system for joint scheduling of computing resources, characterized in that: The method for calculating the multi-beam satellite service according to claim 1 comprises: A multi-beam antenna unit, a multi-rate digital processing unit, a computing unit, a management control unit, and a storage unit, wherein: The multi-beam antenna unit is composed of a multi-beam antenna in the form of a phased array or a reflector, and is used for signal transmission and reception; A multi-rate digital processing unit adjusts the communication rate according to the bandwidth and transmission power of the terminal; The computing unit manages the loading of each terminal according to the computing tasks; Management control unit, used for terminal management, resource allocation, and computing program loading management; The storage unit is used to store application programs and business data. Newly added application programs are transmitted and stored by the terminal. The data exchange unit realizes high-speed data interaction and sharing among the multi-rate digital processing unit, storage unit, computing unit, and management control unit. The interactive data includes business data, program data, and control data.
3. A multi-beam satellite service computing system for joint scheduling of computing resources according to claim 2, characterized in that: The management and control unit includes a task parsing unit, a resource allocation and scheduling unit, and a computing program management unit; The task analysis unit analyzes the computing tasks by information type, computing energy consumption and computing time; The resource allocation and scheduling unit determines the calculation method according to the calculation complexity and calculation time, and sends the calculation task information data to the corresponding chip of the calculation unit; The computing program management unit is used to load, unload, and monitor the operation of the terminal control program, and to apply for and release satellite computing resources.
