The invention discloses an unmanned aerial vehicle track and beamforming design method for a communication computing resource fusion network, and the method comprises the following steps: 1, building an unmanned aerial vehicle-assisted communication computing resource fusion system model, and describing a calculation throughput maximization problem under the perception quality constraint; step 2, decoupling the problem into a combined unmanned aerial vehicle trajectory planning and sensing integrated beam forming sub-problem and a computing resource allocation sub-problem, and modeling as a Markov decision process; 3, training an unmanned aerial vehicle trajectory and beam forming strategy by adopting a soft actor commentator SAC architecture, and obtaining a computing resource allocation decision through an interior point method; and step 4, performing joint optimization design of the general inductance calculation resource fusion network by adopting a completely trained SAC framework, and realizing cooperative gain of general inductance calculation performance. According to the method, adaptive improvement is carried out on a deep reinforcement learning framework, complex derivation needed by a traditional convex approximation method is avoided, and the highly non-convex unmanned aerial vehicle trajectory and beamforming joint optimization problem is efficiently solved.
本发明公开了一种通感算资源融合网络无人机轨迹与波束赋形设计方法,步骤如下:步骤一、建立无人机辅助的通感算资源融合系统模型,描述感知质量约束下的计算吞吐量最大化问题;步骤二、将问题解耦为联合无人机轨迹规划与通感一体化波束赋形子问题和计算资源分配子问题,并建模为马尔科夫决策过程;步骤三、采用软演员评论家SAC架构训练无人机轨迹与波束赋形策略,通过内点法获得计算资源分配决策;步骤四、采用训练完备的SAC框架进行通感算资源融合网络的联合优化设计,实现通感算性能的协同增益。本发明通过对深度强化学习框架进行适应性改进,避免传统凸近似方法所需的复杂推导,高效求解高度非凸的无人机轨迹与波束赋形联合优化问题。
Unmanned aerial vehicle trajectory and beamforming design method for communication computing resource fusion network
一种通感算资源融合网络无人机轨迹与波束赋形设计方法
2024-02-23
Patent
Electronic Resource
Chinese
European Patent Office | 2024
|Unmanned aerial vehicle communication equipment and unmanned aerial vehicle
European Patent Office | 2022
|Unmanned aerial vehicle communication method and device and unmanned aerial vehicle
European Patent Office | 2020
|TRAJECTORY TRACKING CONTROL FOR UNMANNED AERIAL VEHICLE
IEEE | 2018
|