Loading...
2 results
Search Results
Now showing 1 - 2 of 2
- Exploring Deep Reinforcement Learning- Assisted Federated Learning for Online Resource Allocation in Privacy-Preserving EdgeIoTPublication . Zheng, Jingjing; Li, Kai; Ni, Wei; Tovar, Eduardo; Guizani, Mohsen; Mhaisen, NaramFederated learning (FL) has been increasingly considered to preserve data training privacy from eavesdropping attacks in mobile edge computing-based Internet of Thing (EdgeIoT). On the one hand, the learning accuracy of FL can be improved by selecting the IoT devices with large datasets for training, which gives rise to a higher energy consumption. On the other hand, the energy consumption can be reduced by selecting the IoT devices with small datasets for FL, resulting in a falling learning accuracy. In this paper, we formulate a new resource allocation problem for privacy-preserving EdgeIoT to balance the learning accuracy of FL and the energy consumption of the IoT device. We propose a new federated learning-enabled twin-delayed deep deterministic policy gradient (FLDLT3) framework to achieve the optimal accuracy and energy balance in a continuous domain. Furthermore, long short term memory (LSTM) is leveraged in FL-DLT3 to predict the time-varying network state while FL-DLT3 is trained to select the IoT devices and allocate the transmit power. Numerical results demonstrate that the proposed FL-DLT3 achieves fast convergence (less than 100 iterations) while the FL accuracy-to-energy consumption ratio is improved by 51.8% compared to existing state-of-the-art benchmark.
- Deep Q-Learning based Resource Management in UAV-assisted Wireless Powered IoT NetworksPublication . Li, Kai; Ni, Wei; Tovar, Eduardo; Jamalipour, AbbasIn Unmanned Aerial Vehicle (UAV)-assisted Wireless Powered Internet of Things (IoT), the UAV is employed to charge the IoT nodes remotely via Wireless Power Transfer (WPT) and collect their data. A key challenge of resource management for WPT and data collection is preventing battery drainage and butter overflow of the ground IoT nodes in the presence of highly dynamic airborne channels. In this paper, we consider the resource management problem in practical scenarios, where the UAV has no a-prior information on battery levels and data queue lengths of the nodes. We formulate the resource management of UAV-assisted WPT and data collection as Markov Decision Process (MDP), where the states consist of battery levels and data queue lengths of the IoT nodes, channel qualities, and positions of the UAV. A deep Q-learning based resource management is proposed to minimize the overall data packet loss of the IoT nodes, by optimally deciding the IoT node for data collection and power transfer, and the associated modulation scheme of the IoT node.