弱电鱼生存于浑浊的水域，视力严重退化，这种鱼类可以通过尾部的电器官发射低频电信号（称为EODs），在周围环境中形成自生电场。环境中的物体会扰动此自生电场，弱电鱼躯体上遍布如同阵列形式的感知细胞，通过感知细胞接收这种扰动，可以实现对周围环境的感知。弱电鱼启发了水下极低频主动电感知的研究，具体分为三个研究方向，第一是生物学研究，基于弱电鱼生物实验进行统计与总结规律；第二是仿生学研究，基于弱电鱼感知机制进行水下仿生传感器研究；第三是神经学研究，探索弱电鱼感知器官与细胞模型。这三方面的研究尚未形成一套完整的弱电鱼感知理论体系，并且许多研究存在矛盾，因此本论文结合信号处理、电磁学与生物学开展研究，具体内容如下。

第一，研究了弱电鱼感知理论，包含电子图像理论与弱电鱼仿生偶极子模型。由于电子图像理论源于数据统计，没有形成模型化的感知体系，仿生偶极子模型忽略弱电鱼的生物特征与信号特征，都不足以完整地描述弱电鱼的感知机理。因此从弱电鱼信号出发，构造实验测量弱电鱼信号，使用自适应滤波、低通滤波器进行测量信号去噪。使用COMSOL仿真实验环境和包络提取与对比的方法，对实验测量的信号进行验证。

第二，对弱电鱼信号进行特征分析与处理。基于实验测量信号，对弱电鱼信号的时频性质与波形性质开展研究。首先使用短时傅里叶变换（Short-time Fourier Transform，STFT）、连续小波变换（Continuous Wavelet Transform，CWT）和维格纳-威利分布（Wigner Ville Distribution，WVD）进行时频分析。其次使用脊线提取算法与Hough变换分析时频曲线的变化现象，结果表明弱电鱼信号的时频曲线一定程度上有类似正弦形式的变化。为了探究弱电鱼信号的波形性质，采用经验模态分解（Empirical Mode Decomposition，EMD）与CWT进行处理。弱电鱼信号经过EMD分解为2-3种典型的本征模态函数信号（Intrinsic Mode Functions，IMF），基于IMF信号对弱电鱼信号进行小波变换，结果表明弱电鱼的信号在波形上有变化。这些信号处理工作不仅与弱电鱼生物研究相印证，而且为探索弱电鱼感知特征，研究仿生水下电感知建立基础。

第三，探索了弱电鱼感知生物机理与接收信号模型。基于弱电鱼信号特性，仿真论证了HH模型适用于描述弱电鱼接收器官作用机理的可行性；建立了弱电鱼接收器官细胞级二级信号转换的仿真模型，即感知细胞与传入神经细胞级联模型；在此基础上，针对典型弱电鱼生物参数，仿真实现正弦半波电压信号，经过感知细胞转换为脉冲电流，再由传入神经转换为尖峰脉冲序列的过程，实现了从电压扰动到生物神经信号的转换。为后续弱电鱼环境感知机理研究奠定了基础，为弱电鱼接收器官仿生系统实现提供了依据。

Weakly electric fish, which live in turbid waters with severely degraded vision, can create an autogenous electric field in their surroundings by emitting low-frequency electrical signals (called EODs) through the electrical organ in their tail. Objects in the environment perturb this self-generated electric field, and the weakly electric fish's carapace is dotted with sensory cells in the form of arrays, which receive this perturbation and allow them to perceive their surroundings. Weakly electric fish have inspired research on underwater active electrical perception at very low frequencies, which is divided into three research directions: firstly, biological research, based on biological experiments with weakly electric fish, to conduct statistics and summarise patterns; secondly, bionomics research, based on weakly electric fish perception mechanisms, to conduct research on underwater bionic sensors; and thirdly, neurological research, to explore weakly electric fish perceptual organs and cell models. These three aspects of research have not yet formed a complete theoretical system of weakly electric fish perception, and many studies are contradictory, so this paper combines signal processing, electromagnetism and biology to carry out research, as follows.

 

Firstly, we have studied the theory of weakly electric fish perception, which consists of electronic image theory and the bionic dipole model of weakly electric fish. As the electronic image theory originates from data statistics, it does not form a modeled perception system, and the bionic dipole model ignores the biological and signal characteristics of weakly electric fish, both of which are insufficient to completely describe the perception mechanism of weakly electric fish. Therefore, the experimental measurement of the weakly electric fish signal is constructed from the weakly electric fish signal, and the measured signal is denoised using adaptive filtering and low-pass filtering. The experimentally measured signals are validated using a COMSOL simulation experimental environment and an envelope extraction and comparison method.

 

Secondly, the weakly electric fish signal is characterised and processed. Based on the experimental measurements, the time-frequency and waveform properties of the weakly electric fish signal are investigated. At first, the short-time Fourier transform (STFT), continuous wavelet transform (CWT) and Wigner ville distribution (WVD) are used for time-frequency analysis. Next, the ridge extraction algorithm and the Hough transform were used to analyse the variation of the time-frequency curves, and the results showed that the time-frequency curves of the weakly electric fish signals have a sinusoidal-like variation to some extent. In order to investigate the waveform nature of the weakly electric fish signal, empirical mode decomposition (EMD) with CWT was used. The weakly electric fish signal is decomposed by EMD into 2-3 typical intrinsic mode functions (IMF) signals, and the wavelet transform of the weakly electric fish signal is based on the IMF signal. These signal processing works not only corroborate with the biological studies of weakly electric fish, but also establish the basis for exploring the perceptual characteristics of weakly electric fish and studying bionic underwater electrical perception.

 

Thirdly, we explored the biological mechanism of perception and the receiving signal model of weakly electric fish. Based on the signal characteristics of the weakly electric fish, the simulation demonstrates the feasibility of the HH model to describe the action mechanism of the receiving organ of the weakly electric fish; the simulation model of the cellular-level secondary signal conversion of the receiving organ of the weakly electric fish, i.e. the sensing cell and the afferent nerve cell cascade model, is established; on the basis of this model, a sinusoidal half-wave voltage signal is simulated and converted into a pulse current by the sensing cell and then into a spike pulse sequence by the afferent nerve, realizing the conversion from voltage perturbation to biological nerve signal. This has established the basis for subsequent research on the environmental perception mechanism of weakly electric fish and provided a basis for the implementation of a bionic system for the receiving organ of weakly electric fish.

[1] 吕文平. 深海伺服阀控缸系统[D]. 东北大学, 2011.
[2] 梁婷蓉. 基于水下传感器网络的目标被动定位算法研究[D]. 天津城建大学, 2022.
[3] 王天宇. 基于弱电鱼仿生的水下电流场探测技术研究[D]. 哈尔滨工程大学, 2019.
[4] 刘智辉，何俊华，王国富. 二维扫描式水下电视系统的设计与实现[J].光子学报，2008, 37(2): 410-412.
[5] 白广明, 卢建旗. 水下电视系统设备研制及图像处理方法研究现状综述[J]. 黑龙江水利, 2016, 2(9):1-5.
[6] Lissmann H W. Electric Location by Fishes[J]. Scientific American, 1963, 208(3):50-59.
[7] Lissmann H W. On the function and evolution of electric organs in fish[J]. Journal of experimental biology, 1958, 35(1): 156-191.
[8] Scheich H, Bullock T H, Hamstra Jr R H. Coding properties of two classes of afferent nerve fibers: high-frequency electroreceptors in the electric fish, Eigenmannia[J]. Journal of Neurophysiology, 1973, 36(1): 39-60.
[9] Von d E G , Schwarz S , Gomez L , et al. Electric fish measure distance in the dark.[J]. Nature, 1998, 395(6705):890-4.
[10] Von d E G. Discrimination of objects through electrolocation in the weakly electric fish, Gnathonemus petersii[J]. Journal of Comparative Physiology A, 1990, 167(3): 413-421.
[11] Von d E G, Schwarz S, Gomez L, et al. Electric fish measure distance in the dark[J]. Nature, 1998, 395(6705): 890-894.
[12] Caputi A A, Budelli R, Grant K, et al. The electric image in weakly electric fish: physical images of resistive objects in Gnathonemus petersii[J]. The Journal of experimental biology, 1998, 201(14): 2115-2128.
[13] Budelli R, Caputi A A. The electric image in weakly electric fish: perception of objects of complex impedance[J]. Journal of Experimental Biology, 2000, 203(3): 481-492.
[14] Von d E G. Distance and shape: perception of the 3-dimensional world by weakly electric fish[J]. Journal of Physiology-Paris, 2004, 98(1-3): 67-80.
[15] Fechler K, Holtkamp D, Neusel G, et al. Mind the gap: the minimal detectable separation distance between two objects during active electrolocation[J]. Journal of Fish Biology, 2012, 81(7): 2255-2276.
[16] Fechler K, von der Emde G. Figure–ground separation during active electrolocation in the weakly electric fish, Gnathonemus petersii[J]. Journal of Physiology-Paris, 2013, 107(1-2): 72-83.
[17] Caputi A A, Aguilera P A, Pereira A C, et al. On the haptic nature of the active electric sense of fish[J]. Brain research, 2013, 1536: 27-43.
[18] Gottwald M, Singh N, Haubrich A N, et al. Electric-color sensing in weakly electric fish suggests color perception as a sensory concept beyond vision[J]. Current Biology, 2018, 28(22): 3648-3653. e2.
[19] Rasnow B. The effects of simple objects on the electric field of Apteronotus[J]. Journal of Comparative Physiology A, 1996, 178(3): 397-411.
[20] Solberg J R, Lynch K M, MacIver M A. Active electrolocation for underwater target localization[J]. The International Journal of Robotics Research, 2008, 27(5): 529-548.
[21] Underwater robot navigation around a sphere using electrolocation sense and Kalman filter
[22] 彭杰钢, 吴俊. 基于 Cole-Cole 模型的水下主动电场定位系统有限元模拟[J]. 科学通报, 2016 (23): 2647-2658.
[23] Ren Q, Peng J, Chen H. Amplitude information-frequency characteristics for multi-frequency excitation of underwater active electrolocation systems[J]. Bioinspiration & Biomimetics, 2019, 15(1): 016004.
[24] Bastian J. Electrolocation[J]. Journal of comparative physiology, 1981, 144(4): 481-494.
[25] Wessel R, Koch C, Gabbiani F. Coding of time-varying electric field amplitude modulations in a wave-type electric fish[J]. Journal of neurophysiology, 1996, 75(6): 2280-2293.
[26] Nelson M E, Xu Z, Payne J R. Characterization and modeling of P-type electrosensory afferent responses to amplitude modulations in a wave-type electric fish[J]. Journal of Comparative Physiology A, 1997, 181(5): 532-544.
[27] Benda J, Herz A V M. A universal model for spike-frequency adaptation[J]. Neural computation, 2003, 15(11): 2523-2564.
[28] Benda J, Longtin A, Maler L. Spike-frequency adaptation separates transient communication signals from background oscillations[J]. Journal of Neuroscience, 2005, 25(9): 2312-2321.
[29] Chacron M J. Nonlinear information processing in a model sensory system[J]. Journal of Neurophysiology, 2006, 95(5): 2933-2946.
[30] Drew P J, Abbott L F. Models and properties of power-law adaptation in neural systems[J]. Journal of neurophysiology, 2006, 96(2): 826-833.
[31] Gussin D, Benda J, Maler L. Limits of linear rate coding of dynamic stimuli by electroreceptor afferents[J]. Journal of Neurophysiology, 2007, 97(4): 2917-2929.
[32] Chacron M J, Toporikova N, Fortune E S. Differences in the time course of short-term depression across receptive fields are correlated with directional selectivity in electrosensory neurons[J]. Journal of neurophysiology, 2009, 102(6): 3270-3279.
[33] Chacron M J, Longtin A, Maler L. Efficient computation via sparse coding in electrosensory neural networks[J]. Current opinion in neurobiology, 2011, 21(5): 752-760.
[34] Muller S Z, Zadina A N, Abbott L F, et al. Continual learning in a multi-layer network of an electric fish[J]. Cell, 2019, 179(6): 1382-1392. e10.
[35] Hofmann M H . Detection and generation of electric signals | Physiology of Ampullary Electrosensory Systems[J]. Encyclopedia of Fish Physiology, 2011:359-365.
[36] EE Josberger, Hassanzadeh P , Deng Y , et al. Proton conductivity in ampullae of Lorenzini jelly[J]. Science Advances, 2016, 2(5):e1600112-e1600112.
[37] Bellono N W , Leitch D B , Julius D . Molecular basis of ancestral vertebrate electroreception[J]. Nature, 2017, 543(7645):391-396.
[38] Clarke S E, Longtin A, Maler L. Contrast coding in the electrosensory system: parallels with visual computation[J]. Nature Reviews Neuroscience, 2015, 16(12): 733-744.
[39] Engelmann J, Wallach A, Maler L. Linking active sensing and spatial learning in weakly electric fish[J]. Current opinion in neurobiology, 2021, 71: 1-10.
[40] Shen Z, Song Y, Wang Y, et al. Study on the detection characteristics of Ag/AgCl and carbon fiber marine electric field electrodes[J]. Chin. J. Sci. Instrum, 2018, 39(2): 211-217.
[41] Benda J, Longtin A, Maler L. Spike-frequency adaptation separates transient communication signals from background oscillations[J]. Journal of Neuroscience, 2005, 25(9): 2312-2321.
[42] Post N, von der Emde G. The “novelty response” in an electric fish: response properties and habituation[J]. Physiology & behavior, 1999, 68(1-2): 115-128.
[43] Field C E, Petersen T A, Alves-Gomes J A, et al. A JAR of chirps: the gymnotiform chirp can function as both a communication signal and a jamming avoidance response[J]. Frontiers in integrative neuroscience, 2019: 55.
[44] Hupé G J, Lewis J E, Benda J. The effect of difference frequency on electrocommunication: chirp production and encoding in a species of weakly electric fish, Apteronotus leptorhynchus[J]. Journal of Physiology-Paris, 2008, 102(4-6): 164-172.
[45] Forlim C G, Pinto R D. Automatic realistic real time stimulation/recording in weakly electric fish: long time behavior characterization in freely swimming fish and stimuli discrimination[J]. PLoS One, 2014, 9(1): e84885.
[46] Stankovic L J.A method for time-frequency analysis.IEEE Trans on Signal Process, 1994, 42( 1) :225-229 .
[47] Stankovic L.S -class of time-frequency distributions.IEE Proc.-Vis .Image Signal Process, 1997, 144( 2) :57-64 .
[48] Carmona R A, Hwang Wen L Torresani R.Multiridge detection and time-frequency reconstruction .IEEE Trans on Signal Processing, 1999, 47( 2) :480- 492.
[49] Engelmann J, Wallach A, Maler L. Linking active sensing and spatial learning in weakly electric fish[J]. Current opinion in neurobiology, 2021, 71: 1-10.
[50] Shuai J, Kashimori Y, Kambara T. Electroreceptor model of the weakly electric fish Gnathonemus petersii. I. The model and the origin of differences between A-and B-receptors[J]. Biophysical journal, 1998, 75(4): 1712-1726.
[51] Hodgkin A L, Huxley A F. A quantitative description of membrane current and its application to conduction and excitation in nerve[J]. The Journal of physiology, 1952, 117(4): 500.
[52] Benda J, Maler L. Limits of linear rate coding of dynamic stimuli by electroreceptor afferents[J]. Journal of Neurophysiology, 2007, 97(4): 2917-2929.