This research introduces a Reinforcement Learning (RL) based testbed designed for real-time coordination among multiple Unmanned Aerial Vehicles (UAVs) using actor-critic models to control a randomly manoeuvring target UAV, and a swarm of tracking UAVs. The system simulates realistic multi-agent interactions within a scalable and distributed architecture built on FlightGear and JSBSim. A key feature is the ability of multiple agents to learn independently yet collaboratively, enabling synchronized behaviors and adaptive tracking in dynamic environments. Collision avoidance, swarm cohesion, and target tracking efficiency are ensured through a 3D ellipsoidal spatial model and a Gaussian reward function, which guide agents to maintain optimal relative positions. The integration of an Intrinsic Curiosity Module (ICM) enhances exploration in sparse-reward settings. Experimental results show that the multi-agent swarm UAVs effectively tracks complex and evasive target paths, demonstrating robustness, scalability, and real-time coordination capabilities essential for advanced UAV operations.

Multiple input multiple output (MIMO) communication systems have emerged as a key technol-ogy to enhance spectral efficiency and reliability in wireless communications. In recent years, deep neural network (DNN)-based approaches have shown promise in addressing the challenges of MIMO signal detection. Among these approaches, the Transformer architecture, known for its effectiveness in capturing long-range dependencies in sequential data, has gained significant attention. Therefore, this paper proposes a revolutionary DNN-based MIMO signal detection scheme using the Transformer-based architecture. This novel scheme leverages the multi-head self-attention mechanism inherent in Transformer architectures, which enables the model to capture both spatial and temporal dependencies in MIMO channels, thereby improving symbol detection accuracy and robustness under varying channel conditions. The proposed scheme's bit error rate (BER) performance is compared with traditional methods through simulations. The results show that the proposed method achieves a signal-to-noise ratio (SNR) gain of nearly 1.5 ​dB against the traditional detection methods, with the optimal maximum likelihood detector (MLD) only outperforming it ​by ​< ​0.5 ​dB. • Transformer-based MIMO signal detection. • Deep learning for wireless communications. • Open access journals on signal processing. • Next-generation wireless communication systems.

Accurate prediction of the remaining useful life (RUL) of lithium-ion batteries is important for optimizing battery management strategies and extending battery service cycles. This study proposes a hybrid prediction framework based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and frequency-division modeling. First, the original capacity sequence is decomposed into multiple intrinsic mode functions (IMFs) via CEEMDAN. Subsequently, the components are reconstructed into high-frequency, medium-frequency, and low-frequency segments based on sample entropy to reduce input network dimensions and computational costs. For the high-frequency components characterized by short-term fluctuations, a Temporal Convolutional Network (TCN) integrated with a multi-head attention mechanism is developed to capture local dependencies. The medium-frequency components, representing steady-state attenuation, are modeled using Least Squares Support Vector Regression (LSSVR). Meanwhile, the long-term trend of low-frequency components is predicted via an Online Sequential Regularized Extreme Learning Machine (OSRELM). Finally, the predicted results of all components are linearly reconstructed to achieve capacity estimation. Experimental validation conducted on the Center for Advanced Life Cycle Engineering (CALCE) dataset and the National Aeronautics and Space Administration (NASA) battery dataset demonstrates that the proposed method achieves consistently accurate prediction performance across both datasets.