2 January 2025 in Expert Systems with Applications
3 June 2023 in Expert Systems with Applications
25 June 2025 in Knowledge-Based Systems
27 November 2025 in IEEE Internet of Things Journal
High-accuracy forecasting of short-term electricity load (STEL) is essential for real-time industrial operations, particularly within the context of the Internet of Things (IoT), yet it remains challenging due to the complex and nonlinear nature of power system fluctuations. Given that the brain analysis capabilities far surpass those of neural networks in many scenarios, this article proposes a brain-inspired network, namely BI-Net, that integrates distinct brain mechanisms for enhanced STEL forecasting. On the one hand, the STEL forecasting is framed as a cognitive process, and the BI-Net’s flowchart is designed to mimic the human brain’s layer-by-layer environmental cognition, thereby improving structural rationality. On the other hand, by leveraging multiple brain-inspired mechanisms, including parallel, cooperative, asynchronous, and gating mechanisms, the BI-Net effectively captures nonlinear and long short-term dependencies, enabling enhanced extraction and integration of diverse features. Specifically, the network architecture incorporates a gating module derived from the gating mechanism, a high-level feature extraction module inspired by parallel and asynchronous mechanisms, and an interactive coordination module based on the cooperative mechanism. Evaluated on three real-world STEL datasets, BI-Net demonstrates superior forecasting performance and serves as an effective tool for power management, combining the analytical strengths of brain science with rationality in its multilevel design.
9 July 2025 in Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering
Trajectory prediction (TP) is critical to the aircraft flight safety especially in few-shot situations, for example, flight of private or company aircraft, take-off or landing on aircraft carriers, etc., however most existing high-accuracy TP methods rely heavily on big data. For achieving few-shot TP performance, this paper proposes a few-shot TP structure consisting of two modules: the three-mode TP for the source domain and the three-mode transfer learning for the target domain. In the first module, instead of learning evolutionary characteristics of the whole flight process by a Bi-directional Long Short-Term Memory Network (Bi-LSTM), three Bi-LSTMs corresponding to climbing mode, declining mode, and whole-flight mode are designed, respectively, to extract different flight evolutionary rules. Particularly, the whole-flight mode can improve TP in transition processes between different flight modes. Additionally, outputs from the above three Bi-LSTMs are integrated via Convolutional Neural Networks (CNNs) driven by the source domain big data. In the second module, to avoid few-shot TP performance degradation caused by the negative transfer, transfer learning is carried out between matched flight modes, then the fine-tuning way is applied to improve these new models, and finally the linear combination instead of the CNN is designed to fuse outputs in few-shot situations. Compared with many existing methods, the proposed few-shot TP structure not only learn time-varying non-linear features but also extract multiple flight models separately based on transferred weights. Two kinds of real-world few-shot flight trajectories from Xiamen Airport to Shanghai Airport, and from Xi’an Airport to Beijing Airport are taken as examples to demonstrate the effectiveness of the proposed few-shot TP structure.
19 October 2025 in Applied Soft Computing