A UAV carrier aggregation resource allocation method based on extreme learning machine

Abstract

This invention discloses a UAV carrier aggregation resource allocation method based on Extreme Learning Machine (ELM), belonging to the field of UAV communication technology. The method includes the following steps: S1, preprocessing the sub-channels in the UAV communication system to filter out the available sub-channel set; S2, allocating sub-channels to users according to channel quality ranking; S3, performing power allocation for users in the initially allocated sub-channels based on an ELM network; S4, dynamically adjusting resources for users whose minimum rate requirements are not met after power allocation. This invention has clear logic, low computational complexity, and fast response speed, making it suitable for resource-constrained UAV communication systems and enabling efficient and fair resource allocation in dynamic environments.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) communication technology, specifically a UAV carrier aggregation resource allocation method based on extreme learning machine. Background Technology

[0002] With the widespread application of unmanned aerial vehicles (UAVs) in scenarios such as communication relay, emergency communication, and network coverage enhancement, higher demands are placed on the transmission rate and dynamic resource allocation capabilities of UAV communication systems. Carrier aggregation technology, by aggregating multiple discrete sub-channels, can effectively improve the transmission capacity of UAV links and meet the needs of high-bandwidth services.

[0003] However, the UAV communication environment is characterized by high node mobility, rapid channel changes, and limited energy and computing resources. Traditional resource allocation methods based on the Lagrange multiplier method or iterative optimization are difficult to execute in real time on UAV platforms due to their high computational complexity, making them unsuitable for dynamic deployment. Furthermore, existing deep learning-based resource allocation methods often require extensive parameter training and have high deployment costs, making them difficult to implement on small UAV onboard processors.

[0004] Therefore, there is an urgent need for a lightweight, low-computational-complexity, and dynamic-environment-adaptable resource allocation method that can quickly complete the sub-channel and power allocation in UAV carrier aggregation systems while ensuring system fairness and quality of service. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a UAV carrier aggregation resource allocation method based on extreme learning machine, applicable to dynamic resource allocation scenarios in unmanned aerial vehicle (UAV) communication systems.

[0006] The technical solution of this invention is: a UAV carrier aggregation resource allocation method based on extreme learning machine, comprising the following steps:

[0007] S1. Preprocess the sub-channels in the UAV communication system and filter out the set of usable sub-channels;

[0008] S2. Allocate sub-channels to users based on channel quality ranking;

[0009] S3. Based on the extreme learning machine network, power allocation is performed on users in the sub-channels after the initial allocation;

[0010] S4. Dynamically adjust resources for users who do not meet the minimum rate requirement after power allocation.

[0011] Furthermore, in S1, the preprocessing specifically includes: excluding sub-channels occupied by the primary user; excluding sub-channels whose frequency interval with the primary user channel is less than the protection band requirement; excluding sub-channels whose interference power exceeds the system's allowable threshold; and excluding sub-channels whose spectrum sensing timestamps have exceeded their validity period.

[0012] Furthermore, S2 includes the following sub-steps:

[0013] S21. For each user whose sub-channel quantity requirement is not met, calculate the benefit estimate for each available sub-channel.

[0014] S22. In the set of available sub-channels for each user, select the sub-channel with the largest difference between the estimated benefit and the current channel price as the target channel with the highest net benefit;

[0015] S23. Calculate the bid for the target channel, enable all users to submit bids in parallel, and temporarily allocate each channel to the user with the highest bid.

[0016] S24. Update the allocated channel price to the winning user's bid and reduce the number of unmet subchannel demands of the winning user;

[0017] S25. Repeat S21 to S24 until all users meet the sub-channel requirements or no user is willing to bid for the target channel.

[0018] Furthermore, in S21, the user's benefit estimate for the available sub-channels is determined based on the user's priority weight and the channel quality of the sub-channels for the user.

[0019] Furthermore, in S3, the Extreme Learning Machine Network combines the channel quality information of the user on each assigned sub-channel with the user's minimum rate requirement information into an input vector through a single forward propagation calculation, and outputs the power allocation value of the user on each assigned sub-channel.

[0020] Furthermore, S4 includes the following sub-steps:

[0021] S41. Calculate the current total data transmission rate for all users after all power allocations, and identify all users who have not met the minimum rate requirement;

[0022] S42. From the users who have met the minimum rate requirement, select the user with the lowest communication priority and reclaim part or all of the sub-channels occupied by that user.

[0023] S43. The recovered sub-channels are preferentially allocated to the users with the highest communication priority among those who have not met the minimum rate requirement;

[0024] S44. Perform local power allocation for users who have been allocated new sub-channels in S43;

[0025] S45, Repeat S41-S44 until all users meet the minimum rate requirement or there are no more sub-channel resources available for recycling.

[0026] The beneficial effects of this invention are:

[0027] (1) Significantly improved computational efficiency: This invention uses an extreme learning machine to replace the traditional iterative algorithm for power allocation. The result can be output through a single forward propagation, which greatly reduces the computational complexity and can meet the real-time requirements of UAVs in resource-constrained environments.

[0028] (2) More efficient and fair resource allocation: This invention uses an auction algorithm to allocate channels, which achieves efficient and fair allocation of channel resources; combined with a dynamic adjustment mechanism, it can quickly respond to system changes, ensure that resources are given priority to users with the most urgent needs, and improve the overall spectrum efficiency and user satisfaction of the system.

[0029] (3) Strong system robustness: The present invention shows good adaptability to changes in system parameters such as the number of users, UAV transmission power, and minimum user rate requirements, and can work stably in a variety of dynamic scenarios;

[0030] (4) Low implementation cost and easy deployment: The present invention has a clear logic, a high degree of modularity of the device, and a simple extreme learning machine network structure, which is easy to implement and deploy on UAVs and is conducive to large-scale application. Attached Figure Description

[0031] Figure 1 This is a flowchart of a UAV carrier aggregation resource allocation method based on extreme learning machine;

[0032] Figure 2 A flowchart for resource allocation. Detailed Implementation

[0033] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0034] like Figure 1 As shown, this invention provides a UAV carrier aggregation resource allocation method based on extreme learning machine, comprising the following steps:

[0035] S1. Preprocess the sub-channels in the UAV communication system and filter out the set of usable sub-channels;

[0036] S2. Allocate sub-channels to users based on channel quality ranking;

[0037] S3. Based on the extreme learning machine network, power allocation is performed on users in the sub-channels after the initial allocation;

[0038] S4. Dynamically adjust resources for users who do not meet the minimum rate requirement after power allocation.

[0039] like Figure 2 As shown, the core of this invention lies in a hierarchical resource allocation process, which decomposes the complex joint resource allocation problem into two relatively independent sub-problems: channel allocation and power allocation. Different lightweight algorithms are used to solve these sub-problems, and a dynamic adjustment mechanism is ultimately used to ensure the fairness and effectiveness of resource allocation.

[0040] In this embodiment of the invention, S1 specifically involves: excluding sub-channels occupied by the primary user; excluding sub-channels whose frequency interval with the primary user channel is less than the protection band requirement; excluding sub-channels whose interference power exceeds the system's allowable threshold; and excluding sub-channels whose spectrum sensing timestamps have exceeded their validity period.

[0041] In this invention, all sub-channels in the system are screened based on the primary user occupancy status, guard band requirements, real-time interference power, and the validity period of spectrum sensing results. Unusable sub-channels are excluded, generating a set of usable sub-channels. This step, as a preprocessing stage, significantly reduces the complexity of subsequent optimization problems.

[0042] In this embodiment of the invention, S2 includes the following sub-steps:

[0043] S21. For each user whose sub-channel quantity requirement is not met, calculate the benefit estimate for each available sub-channel.

[0044] S22. In the set of available sub-channels for each user, select the sub-channel with the largest difference between the estimated benefit and the current channel price as the target channel with the highest net benefit;

[0045] S23. Calculate the bid for the target channel, enable all users to submit bids in parallel, and temporarily allocate each channel to the user with the highest bid.

[0046] S24. Update the allocated channel price to the winning user's bid and reduce the number of unmet subchannel demands of the winning user;

[0047] S25. Repeat S21 to S24 until all users meet the sub-channel requirements or no user is willing to bid for the target channel.

[0048] In this invention, based on the set of available sub-channels obtained in step S1, a simulated economic auction mechanism is used for the initial allocation of sub-channels. This process includes: each user evaluating the benefits of available channels based on their own priority and channel quality; users selecting the channel with the highest net benefit and submitting a bid; the system temporarily allocating the channel to the user with the highest bid and updating the channel price and user demand; and through multiple iterations, the initial allocation of channels is finally completed.

[0049] In this embodiment of the invention, in S21, the user's benefit estimate for the available sub-channels is determined based on the user's priority weight and the channel quality of the sub-channel for the user.

[0050] In this embodiment of the invention, in S3, the Extreme Learning Machine Network combines the channel quality information of the user on each assigned sub-channel with the user's minimum rate requirement information into an input vector through a single forward propagation calculation, and outputs the power allocation value of the user on each assigned sub-channel.

[0051] In this invention, after the sub-channel allocation is determined, a pre-trained Extreme Learning Machine (ELM) network is used to rapidly allocate power to each user on their assigned sub-channel. The network's input consists of the user's channel quality information and minimum rate requirement on each sub-channel. Through a single forward propagation, the network directly outputs the user's power allocation value. ELM networks are characterized by few training parameters, fast learning speed, and good generalization ability, making them highly suitable for real-time processing in UAVs.

[0052] In this embodiment of the invention, S4 includes the following sub-steps:

[0053] S41. Calculate the current total data transmission rate for all users after all power allocations, and identify all users who have not met the minimum rate requirement;

[0054] S42. From the users who have met the minimum rate requirement, select the user with the lowest communication priority and reclaim part or all of the sub-channels occupied by that user.

[0055] S43. The recovered sub-channels are preferentially allocated to the users with the highest communication priority among those who have not met the minimum rate requirement;

[0056] S44. Perform local power allocation for users who have been allocated new sub-channels in S43;

[0057] S45, Repeat S41-S44 until all users meet the minimum rate requirement or there are no more sub-channel resources available for recycling.

[0058] In this invention, after initial resource allocation, the system checks whether the actual data transmission rate of all users meets their minimum requirements. For users whose requirements are not met, the system reclaims some sub-channels from low-priority users whose requirements have been met and reallocates these channels to high-priority users whose requirements are not met, followed by local power reallocation. This process is repeated iteratively until all user requirements are met or there are no more resources available for reallocation in the system.

[0059] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A UAV carrier aggregation resource allocation method based on extreme learning machine, characterized in that, Includes the following steps: S1. Preprocess the sub-channels in the UAV communication system and filter out the set of usable sub-channels; S2. Allocate sub-channels to users based on channel quality ranking; S3. Based on the extreme learning machine network, power allocation is performed on users in the sub-channels after the initial allocation; S4. Dynamically adjust resources for users who do not meet the minimum rate requirement after power allocation.

2. The UAV carrier aggregation resource allocation method based on extreme learning machine according to claim 1, characterized in that, In S1, the preprocessing specifically includes: excluding sub-channels occupied by the primary user; excluding sub-channels whose frequency interval with the primary user channel is less than the protection band requirement; excluding sub-channels whose interference power exceeds the system's allowable threshold; and excluding sub-channels whose spectrum sensing timestamps have exceeded their validity period.

3. The UAV carrier aggregation resource allocation method based on extreme learning machine according to claim 1, characterized in that, S2 includes the following sub-steps: S21. For each user whose sub-channel quantity requirement is not met, calculate the benefit estimate for each available sub-channel. S22. In the set of available sub-channels for each user, select the sub-channel with the largest difference between the estimated benefit and the current channel price as the target channel with the highest net benefit; S23. Calculate the bid for the target channel, enable all users to submit bids in parallel, and temporarily allocate each channel to the user with the highest bid. S24. Update the allocated channel price to the winning user's bid and reduce the number of unmet subchannel demands of the winning user; S25. Repeat S21 to S24 until all users meet the sub-channel requirements or no user is willing to bid for the target channel.

4. The UAV carrier aggregation resource allocation method based on extreme learning machine according to claim 3, characterized in that, In step S21, the user's benefit estimate for the available sub-channels is determined based on the user's priority weight and the channel quality of the sub-channels for the user.

5. The UAV carrier aggregation resource allocation method based on extreme learning machine according to claim 1, characterized in that, In S3, the Extreme Learning Machine Network combines the channel quality information of the user on each assigned sub-channel with the user's minimum rate requirement information into an input vector through a single forward propagation calculation, and outputs the power allocation value of the user on each assigned sub-channel.

6. The UAV carrier aggregation resource allocation method based on extreme learning machine according to claim 1, characterized in that, S4 includes the following sub-steps: S41. Calculate the current total data transmission rate for all users after all power allocations, and identify all users who have not met the minimum rate requirement; S42. From the users who have met the minimum rate requirement, select the user with the lowest communication priority and reclaim part or all of the sub-channels occupied by that user. S43. The recovered sub-channels are preferentially allocated to the users with the highest communication priority among those who have not met the minimum rate requirement; S44. Perform local power allocation for users who have been allocated new sub-channels in S43; S45, Repeat S41-S44 until all users meet the minimum rate requirement or there are no more sub-channel resources available for recycling.

Images