Selected work

Representative Papers

FlightMagNav: An Open Dataset and Probabilistic Map Learning and Validation Framework for Outdoor Magnetic Field-Based Positioning

Oct 06, 2026

This study addresses the absence of open datasets and standardized validation frameworks for geomagnetic localization of low-altitude unmanned aerial vehicles by constructing the first open-source geomagnetic benchmark dataset tailored for aerial platforms. Methodologically, multimodal data are acquired by integrating an optically pumped quantum magnetometer with an inertial navigation system, and a probabilistic framework for geomagnetic map learning and evaluation is proposed. The primary contributions of this work lie in bridging the critical data gap in interference-resilient, infrastructure-free localization, establishing new standards for evaluating geomagnetic positioning performance, and providing a reproducible benchmarking platform to facilitate future research in this domain.

0 citationsRead paper

Residual Kalman Dynamics for Event-Based UAV Forecasting

Sep 01, 2026

研究使用残差卡尔曼滤波方法预测基于事件相机数据的无人机短中期轨迹，提高了预测准确性。

0 citationsRead paper

Autonomous generation of different courses of action in mechanized combat operations

Nov 07, 2025

To address the slow autonomous decision-making and poor adaptability of tactical plans in dynamic, adversarial mechanized ground combat environments, this paper proposes a real-time action plan generation and evolution method based on a sequential decision-making framework. The method integrates rule-based reasoning, battlefield modeling, and simulation techniques to concurrently generate and evaluate thousands of feasible maneuver paths under multidimensional constraints—including force posture, unit composition, offensive/defensive mission types, and advance velocity. A closed-loop feedback mechanism enables dynamic plan revision and performance optimization. Compared with conventional static tactical planning approaches, the proposed method significantly improves both plan generation speed and environmental adaptability. Experimental results demonstrate rapid generation of superior alternative strategies over baseline plans, achieving over 40% improvement in command decision-making efficiency. This work provides a scalable technical foundation for adaptive operations in complex, time-critical battlefield scenarios.

0 citationsRead paper

Active Inference for an Intelligent Agent in Autonomous Reconnaissance Missions

Oct 20, 2025

To address the challenges of autonomous agent path planning and collaborative situational awareness in reconnaissance, this paper proposes a dynamic path planning method grounded in active inference. The approach constructs a generative model integrating Dempster–Shafer evidence theory with a Gaussian sensor model; observation likelihoods are defined via pignistic probabilities, and decision-making is driven by variational free energy minimization. Evidence maps are updated online via Bayesian inference, enabling adaptive trade-offs between exploration and exploitation. Its key innovation lies in embedding evidence theory within the active inference framework and establishing a free energy computation paradigm linking pignistic probabilities to posterior distributions. Simulation results demonstrate significant improvements in wide-area search efficiency and target tracking robustness, while enabling multi-agent persistent co-construction and real-time updating of a unified operational situational map.

0 citationsRead paper

Drone Detection Using a Low-Power Neuromorphic Virtual Tripwire

Sep 16, 2025

Small unmanned aerial vehicles (UAVs) pose increasingly severe threats to military and civilian infrastructure, necessitating low-power, long-term autonomous detection solutions. Method: This paper proposes a fully neuromorphic, low-power virtual perimeter system comprising event-based cameras, spiking neural networks (SNNs), and neuromorphic chips for edge detection. Trained exclusively on synthetic data, the SNN achieves robust geometric UAV recognition—eliminating reliance on rotor motion cues—and supports multi-node networking for spatiotemporally precise intrusion monitoring within restricted zones. Contribution/Results: Compared to GPU-based edge inference systems, the proposed architecture reduces power consumption by two to three orders of magnitude, enabling over one year of battery-only operation. It establishes the first neuromorphic UAV detection paradigm capable of sustained, grid-free, autonomous deployment—marking a significant advance in energy-efficient, real-time aerial threat surveillance.

0 citationsRead paper

Recent publications

Latest Papers

FlightMagNav: An Open Dataset and Probabilistic Map Learning and Validation Framework for Outdoor Magnetic Field-Based Positioning

Oct 06, 2026

This study addresses the absence of open datasets and standardized validation frameworks for geomagnetic localization of low-altitude unmanned aerial vehicles by constructing the first open-source geomagnetic benchmark dataset tailored for aerial platforms. Methodologically, multimodal data are acquired by integrating an optically pumped quantum magnetometer with an inertial navigation system, and a probabilistic framework for geomagnetic map learning and evaluation is proposed. The primary contributions of this work lie in bridging the critical data gap in interference-resilient, infrastructure-free localization, establishing new standards for evaluating geomagnetic positioning performance, and providing a reproducible benchmarking platform to facilitate future research in this domain.

0 citationsRead paper

Residual Kalman Dynamics for Event-Based UAV Forecasting

Sep 01, 2026

研究使用残差卡尔曼滤波方法预测基于事件相机数据的无人机短中期轨迹，提高了预测准确性。

0 citationsRead paper

Autonomous generation of different courses of action in mechanized combat operations

Nov 07, 2025

To address the slow autonomous decision-making and poor adaptability of tactical plans in dynamic, adversarial mechanized ground combat environments, this paper proposes a real-time action plan generation and evolution method based on a sequential decision-making framework. The method integrates rule-based reasoning, battlefield modeling, and simulation techniques to concurrently generate and evaluate thousands of feasible maneuver paths under multidimensional constraints—including force posture, unit composition, offensive/defensive mission types, and advance velocity. A closed-loop feedback mechanism enables dynamic plan revision and performance optimization. Compared with conventional static tactical planning approaches, the proposed method significantly improves both plan generation speed and environmental adaptability. Experimental results demonstrate rapid generation of superior alternative strategies over baseline plans, achieving over 40% improvement in command decision-making efficiency. This work provides a scalable technical foundation for adaptive operations in complex, time-critical battlefield scenarios.

0 citationsRead paper

Active Inference for an Intelligent Agent in Autonomous Reconnaissance Missions

Oct 20, 2025

To address the challenges of autonomous agent path planning and collaborative situational awareness in reconnaissance, this paper proposes a dynamic path planning method grounded in active inference. The approach constructs a generative model integrating Dempster–Shafer evidence theory with a Gaussian sensor model; observation likelihoods are defined via pignistic probabilities, and decision-making is driven by variational free energy minimization. Evidence maps are updated online via Bayesian inference, enabling adaptive trade-offs between exploration and exploitation. Its key innovation lies in embedding evidence theory within the active inference framework and establishing a free energy computation paradigm linking pignistic probabilities to posterior distributions. Simulation results demonstrate significant improvements in wide-area search efficiency and target tracking robustness, while enabling multi-agent persistent co-construction and real-time updating of a unified operational situational map.

0 citationsRead paper

Drone Detection Using a Low-Power Neuromorphic Virtual Tripwire

Sep 16, 2025

Small unmanned aerial vehicles (UAVs) pose increasingly severe threats to military and civilian infrastructure, necessitating low-power, long-term autonomous detection solutions. Method: This paper proposes a fully neuromorphic, low-power virtual perimeter system comprising event-based cameras, spiking neural networks (SNNs), and neuromorphic chips for edge detection. Trained exclusively on synthetic data, the SNN achieves robust geometric UAV recognition—eliminating reliance on rotor motion cues—and supports multi-node networking for spatiotemporally precise intrusion monitoring within restricted zones. Contribution/Results: Compared to GPU-based edge inference systems, the proposed architecture reduces power consumption by two to three orders of magnitude, enabling over one year of battery-only operation. It establishes the first neuromorphic UAV detection paradigm capable of sustained, grid-free, autonomous deployment—marking a significant advance in energy-efficient, real-time aerial threat surveillance.

0 citationsRead paper