Institution profile

Nanjing University of Aeronautics and Astronautics

Academic institutionasia · cn
Official website
Research library465linked papers
Opportunities0open roles
Selected work

Representative Papers

SpectrumFM: A Foundation Model for Intelligent Spectrum Management

May 02, 2025arXiv.org

To address the low recognition accuracy, slow convergence, and poor generalization of existing small-scale models in dynamic spectrum environments, this paper proposes SpectrumFM—a spectral foundation model. Methodologically, SpectrumFM integrates CNNs with multi-head self-attention to enhance IQ-signal representation learning; introduces the first foundation-model paradigm for spectrum analysis, featuring dual self-supervised pretraining tasks—masked signal reconstruction and next-time-step signal prediction; and employs parameter-efficient fine-tuning (e.g., LoRA) for cross-task transfer. Experiments demonstrate significant improvements: 12.1% higher accuracy in automatic modulation classification (AMC), 9.3% gain in wireless technology classification (WTC), an AUC of 0.97 for spectrum sensing at −4 dB SNR, over 10% improvement in anomaly detection performance, faster convergence, and markedly enhanced few-shot adaptation capability.

3 citations1 influentialRead paper

MM-LINS: a Multi-Map LiDAR-Inertial System for Over-Degenerate Environments

Mar 25, 2025IEEE Transactions on Intelligent Vehicles

In highly degraded environments—such as warehouse logistics and food delivery—where human occlusion, airborne debris, and cooking fumes cause severe LiDAR point cloud sparsity (“over-degradation”), existing LiDAR-inertial SLAM systems suffer catastrophic map drift. To address this, we propose the first multi-map LiDAR-inertial SLAM framework specifically designed for over-degraded scenarios. Our method features: (1) a novel dynamic multi-map mechanism that detects degradation in real time and freezes active maps to prevent error accumulation; (2) a constraint-enhanced cross-map fusion strategy leveraging Scan Context for identifying dormant maps and optimizing pose alignment via overlapping trajectories; and (3) an integrated pipeline comprising an iterative error-state Kalman filter frontend, dynamic initialization, and a graph-optimization backend. Evaluated on public over-degradation datasets and real-world complex environments, our system significantly suppresses drift, achieving high-precision localization and dense mapping. The source code is publicly available.

3 citationsRead paper

Hallucination Mitigating for Medical Report Generation

Jan 22, 2026

This work addresses the challenge of hallucinations in large vision-language models that compromise clinical reliability in medical report generation. To mitigate this issue, the authors propose the Knowledge-Enhanced Report Generation with Retrieval and Mitigation (KERM) framework, which first retrieves relevant lesion-related knowledge using MedCLIP, then employs a context-aware filtering module to select knowledge consistent with the patient’s imaging findings and medical history, and finally integrates a fine-grained reinforcement learning reward mechanism to guide the model toward generating accurate, evidence-based medical descriptions. Experimental results on the IU-Xray and MIMIC-CXR datasets demonstrate that KERM significantly reduces hallucination rates while improving both clinical accuracy and overall report quality.

1 citationsRead paper
Recent publications

Latest Papers