A Comparison of Meta-learning and Transfer-learning for Few-shot Jamming Signal Classification
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Typical anti-jamming technologies based on array antennas, Space Time Adaptive Process (STAP) & Space Frequency Adaptive Process (SFAP), are very effective algorithms to perform nulling and beamforming. However, it does not perform equally well for all types of jamming signals. If the anti-jamming algorithm is not optimized for each signal type, antijamming performance deteriorates and the operation stability of the system become worse by unnecessary computation. Therefore, jamming classification technique is required to obtain optimal anti-jamming performance. Machine learning, which has recently been in the spotlight, can be considered to classify jamming signal. In general, performing supervised learning for classification requires a huge amount of data and new learning for unfamiliar signal. In the case of jamming signal classification, it is difficult to obtain large amount of data because outdoor jamming signal reception environment is difficult to configure and the signal type of attacker is unknown. Therefore, this paper proposes few-shot jamming signal classification technique using meta-learning and transfer-learning to train the model using a small amount of data. A training dataset is constructed by anti-jamming algorithm input data within the GNSS receiver when jamming signals are applied. For metalearning, Model-Agnostic Meta-Learning (MAML) algorithm with a general Convolution Neural Networks (CNN) model is used, and the same CNN model is used for transfer-learning. They are trained through episodic training using training datasets on developed our Python-based simulator. The results show both algorithms can be trained with less data and immediately respond to new signal types. Also, the performances of two algorithms are compared to determine which algorithm is more suitable for classifying jamming signals.
Abstract
Typical anti-jamming technologies based on array antennas, Space Time Adaptive Process (STAP) & Space Frequency Adaptive Process (SFAP), are very effective algorithms to perform nulling and beamforming. However, it does not perform equally well for all types of jamming signals. If the anti-jamming algorithm is not optimized for each signal type, antijamming performance deteriorates and the operation stability of the system become worse by unnecessary computation. Therefore, jamming classification technique is required to obtain optimal anti-jamming performance. Machine learning, which has recently been in the spotlight, can be considered to classify jamming signal. In general, performing supervised learning for classification requires a huge amount of data and new learning for unfamiliar signal. In the case of jamming signal classification, it is difficult to obtain large amount of data because outdoor jamming signal reception environment is difficult to configure and the signal type of attacker is unknown. Therefore, this paper proposes few-shot jamming signal classification technique using meta-learning and transfer-learning to train the model using a small amount of data. A training dataset is constructed by anti-jamming algorithm input data within the GNSS receiver when jamming signals are applied. For metalearning, Model-Agnostic Meta-Learning (MAML) algorithm with a general Convolution Neural Networks (CNN) model is used, and the same CNN model is used for transfer-learning. They are trained through episodic training using training datasets on developed our Python-based simulator. The results show both algorithms can be trained with less data and immediately respond to new signal types. Also, the performances of two algorithms are compared to determine which algorithm is more suitable for classifying jamming signals.
참고문헌 (23)
- Bazec, M, "GNSS Jammer detection, classification and spectrum analysis", Proceeding of the International Conference on Transport Science (2018)
- Fadaei, N, "Detection, characterization and mitigation of GNSS jamming interference using pre-correlation methods", 박사, pp.2016
- Ferre, R. M., "Jammer classification in GNSS bands via machine learning algorithms", Sensors, 19, pp.4841~ (2019)[DOI]
- Finn, C., "Model-agnostic meta-learning for fast adaptation of deep networks", Proceedings of the 34th International conference on machine learning, 70, pp.1126~1135 (2017)
- Gross, J. N., "GNSS spoofing, jamming, and multipath interference classification using a maximum-likelihood multi-tap multipath estimator", Proceedings of the 2017 International Technical Meeting of the Institute of Navigation, pp.662~670 (2017)
- He, H, "Learning from imbalanced data", IEEE Transactions on knowledge and data engineering, 21, pp.1263~1284 (2009)[DOI]
- 진미현, "Jammer Identification: Spectral Correlation Function and Wavelet Coherence", Journal of Positioning, Navigation, and Timing, 7(3), pp.147~153 (2018), 사단법인 항법시스템학회[DOI]
- Jin, M. H., "Meta-learning Based Few-shot Jamming Signal Classification Technique", 2021 IPNT conference (2021)
- Kingma, D. P, "Adam: A method for stochastic optimization", ICLR (2015)
- Koch, G., "Siamese neural networks for one-shot image recognition", 석사, pp.2015
- 이영중, "SVM을 이용한 GPS 재밍 탐지에 대한 연구", 한국정보기술학회논문지, 17(1), pp.11~20 (2019), 한국정보기술학회[DOI]
- Moore, T. D., "Analytic study of space-time and spacefrequency adaptive processing for radio frequency interference suppression", 박사, pp.2002
- Munkhdalai, T, "Meta networks", Proceedings of the 34th International conference on machine learning, 70, pp.2554~2563 (2017)
- Pan, S. J, "A survey on transfer learning", IEEE Transactions on knowledge and data engineering, 22, pp.1345~1359 (2010)[DOI]
- Ravi, S., "Optimization as a model for few-shot learning", ICLR (2017)
- Santoro, A, "Meta-learning with memoryaugmented neural networks", International conference on machine learning, 48, pp.1842~1850 (2016)
- Schmidt, R, "Multiple emitter location and signal parameter estimation", IEEE transactions on antennas and propagation, 34, pp.276~280 (1986)[DOI]
- "learning literature survey, Computer Sciences Technical Report 1648" (2009)[URL]
- Smith, J, "Closed-form least-squares source location estimation from range-difference measurements", IEEE Transactions on Acoustics, Speech, and Signal Processing, 35, pp.1661~1669 (1987)[DOI]
- Vinyals, O., "Matching networks for one shot learning", NIPS (2016)
- Wang, Y., "Generalizing from a few examples: A survey on few-shot learning", ACM Computing Surveys, 53, pp.1~34 (2021)[DOI]
- Space-Time Adaptive Processing for Airborne Radar, "Ward, J", Massachusetts Institute of Technology Lincoln Laboratory, ~, pp.Technical Report 1015 (1994)
- 유승수, "합성곱 신경망 기반 향상된 GNSS 재밍 식별기법", 제어.로봇.시스템학회 논문지, 26(11), pp.1016~1027 (2020), 제어·로봇·시스템학회[DOI]