仿神經計算是近年的新興領域，該計算架構提供了平行運算與低功耗的特性，同時也有許多利用該計算架構的應用被提出。然而，大部分主流應用架構對仿神經計算中的吸引子特性著墨甚少。因此，本研究將重心放在遍布生物系統的吸引子特性上，且為了完整呈現其應用前景，我們提出三個吸引子仿生神經網路的應用情境。首先，我們在自行開發的Flyintel仿生自走車平台上實作一個簡單的覓食與迴避隨機行走（forage and avoidance random walk, FAR）網路，使自走車在一個固定圓環裡隨機行走並追縱特定標誌。接著我們將重心放在果蠅的中央複合體（central complex, CX）模型，該模型利用吸引子的空間位置來表示果蠅與特定標誌的相對關係，果蠅自身的旋轉會同步轉移吸引子的位置，以維持和標誌間的相對關係。藉此我們建立簡化中央複合體模型（simplified central complex, SCX），並同樣在Flyintel上呈現其功能。自走車能在原始標誌已經移除的情境下，依據記憶中的位置導航前往該地。另一方面，吸引子特性也可以應用於時間序列上，我們提出脈衝式神經網路序列控制（SNN sequence control, SSC）架構，吸引子在這裡代表的是一個任務序列的進度，每當一項任務完成，吸引子就會傳遞至後一個神經網路節點。最後我們透過SSC控制一架模擬四軸飛行器執行一連串飛行動作，並且加上簡單的例外處理功能，以凸顯SSC實際應之可行性。

Neuromorphic computing is an emerging field with the advantages of parallel computing and low power consumption. Recently, researchers have proposed numerous neuromorphic algorithms and application scenarios to harness the new technology’s potential. However, only a few studies have shown dedicated applications of attractor dynamics, a ubiquitous property in biological systems, in neuromorphic engineering. To fully exploit the potential for neuromorphic computing, we build three bio-inspired neural networks by utilizing attractor dynamics. Starting from the forage and avoidance random walk (FAR) network on the Flyintel mobile robotic platform, the robot demonstrates the decision-making capability in an arena with a visual landmark. We further refer to the fruit fly central complex model. A bump attractor denotes a specific position in space, and the attractor is transient and can be repositioned through asymmetric connections. We derive a simplified central complex (SCX) model and realize it on the Flyintel by preserving the same principle. The learned spatial memory leads the robot back to the landmark when the landmark is absent. Next, we chronologically apply the attractor dynamics. In the SNN sequence control (SSC) framework, the bump denotes the progress of a sequence of tasks. The bump advances by one while the task in hand is completed. We demonstrate that the SSC controls a simulated quadcopter to maneuver a preprogrammed mission with essential capability in handling exceptions.

摘要 i
Abstract ii
誌謝 iii
Chapter 1 Introduction 1
1.1 Attractors in SNN 1
1.2 Application subjects 2
1.3 Related Works 3
1.3.1 Decision-making in robot control 3
1.3.2 Spatial memory for robot navigation 4
1.3.3 Memory for sequential task control 5
1.4 Structure of this thesis 6
Chapter 2 SNN Application Platforms 8
2.1 Flyintel 8
2.1.1 FlyintelBot 8
2.1.2 Flysim – the SNN simulator 11
2.1.3 Flyintel – spike data handling 11
2.1.4 Flyintel arena 12
2.2 Simulated Quadcopter 15
2.2.1 SNN simulator and spike handling 15
2.2.2 Drone simulation and piloting 15
Chapter 3 Forage and avoidance random walk network 18
3.1 Network model 18
3.2 Stimulus encoding and activity decoding 21
3.3 Experiment design 23
3.3.1 Risk avoidance task 23
3.3.2 Forage task 23
3.4 Results 24
3.4.1 Risk avoidance task 24
3.4.2 Forage task 25
Chapter 4 Simplified Central Complex Model 27
4.1 The Drosophila central complex model 27
4.2 Simplified Drosophila central complex model 28
4.3 Experiment design 31
4.3.1 Spatial working memory task 31
4.4 Results 33
Chapter 5 SNN Sequence Control 34
5.1 Networks and experiment design 34
5.1.1 Linear sequence network 34
5.1.2 Branched sequence network 37
5.1.3 Task condition of satisfaction unit 39
5.1.4 An example of drone application 41
5.2 Robustness 42
5.3 Results 42
Chapter 6 Conclusion and Discussion 45
Bibliography 47
Appendix A Models and Parameters 52
A.1 Spiking neuron model 52
A.2 Neuron and synapse parameters 53
A.2.1 FAR network 53
A.2.2 SCX network 54
A.2.3 SSC 54
A.3 Spike generation thresholds 55
A.4 Network configurations 56
A.4.1 FAR network (standard 1× inhibitory weights) 56
A.4.2 SCX network 57
A.4.3 Linear sequence network 57
A.4.4 Branched sequence network 57
A.4.5 CoS sequence network 57
Appendix B Assembly Information and Software Implementations 58
B.1 FlyintelBot wiring diagram 58
B.2 Mechanical models in the Flyintel project 59
B.3 Source codes 69
Appendix C Videos 70
C.1 Flyintel videos 70
C.2 SSC videos 70

[1] G. Indiveri and S.-C. Liu, “Memory and Information Processing in Neuromorphic Systems,” Proc. IEEE, vol. 103, no. 8, pp. 1379–1397, Aug. 2015, doi: 10.1109/JPROC.2015.2444094.
[2] W. Maass, “Networks of spiking neurons: The third generation of neural network models,” Neural Netw., vol. 10, no. 9, pp. 1659–1671, Dec. 1997, doi: 10.1016/S0893-6080(97)00011-7.
[3] B. Rueckauer, I.-A. Lungu, Y. Hu, M. Pfeiffer, and S.-C. Liu, “Conversion of Continuous-Valued Deep Networks to Efficient Event-Driven Networks for Image Classification,” Front. Neurosci., vol. 11, 2017, doi: 10.3389/fnins.2017.00682.
[4] S. B. Shrestha and G. Orchard, “SLAYER: Spike Layer Error Reassignment in Time,” Adv. Neural Inf. Process. Syst., vol. 31, 2018, Accessed: Jun. 05, 2021. [Online]. Available: https://papers.nips.cc/paper/2018/hash/82f2b308c3b01637c607ce05f52a2fed-Abstract.html
[5] F. Ponulak and J. J. Hopfield, “Rapid, parallel path planning by propagating wavefronts of spiking neural activity,” Front. Comput. Neurosci., vol. 7, p. 98, Jul. 2013, doi: 10.3389/fncom.2013.00098.
[6] G. Gallego et al., “Event-based Vision: A Survey,” IEEE Trans. Pattern Anal. Mach. Intell., pp. 1–1, 2020, doi: 10.1109/TPAMI.2020.3008413.
[7] Y. Sandamirskaya, “Dynamic neural fields as a step toward cognitive neuromorphic architectures,” Front. Neurosci., vol. 7, 2014, doi: 10.3389/fnins.2013.00276.
[8] E. T. Rolls, “Attractor networks,” WIREs Cogn. Sci., vol. 1, no. 1, pp. 119–134, 2010, doi: 10.1002/wcs.1.
[9] C. D. Brody, R. Romo, and A. Kepecs, “Basic mechanisms for graded persistent activity: discrete attractors, continuous attractors, and dynamic representations,” Curr. Opin. Neurobiol., vol. 13, no. 2, pp. 204–211, Apr. 2003, doi: 10.1016/S0959-4388(03)00050-3.
[10] D. Durstewitz, J. K. Seamans, and T. J. Sejnowski, “Neurocomputational models of working memory,” Nat. Neurosci., vol. 3, no. 11, pp. 1184–1191, Nov. 2000, doi: 10.1038/81460.
[11] J. J. Hopfield, “Neural networks and physical systems with emergent collective computational abilities,” Proc. Natl. Acad. Sci., vol. 79, no. 8, pp. 2554–2558, Apr. 1982, doi: 10.1073/pnas.79.8.2554.
[12] M. Davies et al., “Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook,” Proc. IEEE, vol. 109, no. 5, pp. 911–934, May 2021, doi: 10.1109/JPROC.2021.3067593.
[13] O. Krestinskaya, A. P. James, and L. O. Chua, “Neuromemristive Circuits for Edge Computing: A Review,” IEEE Trans. Neural Netw. Learn. Syst., vol. 31, no. 1, pp. 4–23, Jan. 2020, doi: 10.1109/TNNLS.2019.2899262.
[14] T. Bailey and H. Durrant-Whyte, “Simultaneous localization and mapping (SLAM): part II,” IEEE Robot. Autom. Mag., vol. 13, no. 3, pp. 108–117, Sep. 2006, doi: 10.1109/MRA.2006.1678144.
[15] E. I. Moser, E. Kropff, and M.-B. Moser, “Place Cells, Grid Cells, and the Brain’s Spatial Representation System,” Annu. Rev. Neurosci., vol. 31, no. 1, pp. 69–89, 2008, doi: 10.1146/annurev.neuro.31.061307.090723.
[16] A. Honkanen, A. Adden, J. da Silva Freitas, and S. Heinze, “The insect central complex and the neural basis of navigational strategies,” J. Exp. Biol., vol. 222, no. Suppl_1, Feb. 2019, doi: 10.1242/jeb.188854.
[17] N. Burgess, C. Barry, and J. O’Keefe, “An oscillatory interference model of grid cell firing,” Hippocampus, vol. 17, no. 9, pp. 801–812, 2007, doi: 10.1002/hipo.20327.
[18] T.-S. Su, W.-J. Lee, Y.-C. Huang, C.-T. Wang, and C.-C. Lo, “Coupled symmetric and asymmetric circuits underlying spatial orientation in fruit flies,” Nat. Commun., vol. 8, no. 1, p. 139, Jul. 2017, doi: 10.1038/s41467-017-00191-6.
[19] Y. Sandamirskaya and G. Schoner, “Dynamic field theory of sequential action: A model and its implementation on an embodied agent,” in 2008 7th IEEE International Conference on Development and Learning, Aug. 2008, pp. 133–138. doi: 10.1109/DEVLRN.2008.4640818.
[20] E. Neftci, J. Binas, U. Rutishauser, E. Chicca, G. Indiveri, and R. J. Douglas, “Synthesizing cognition in neuromorphic electronic systems,” Proc. Natl. Acad. Sci., vol. 110, no. 37, pp. E3468–E3476, Sep. 2013.
[21] G. Tang and K. P. Michmizos, “Gridbot: An Autonomous Robot Controlled by a Spiking Neural Network Mimicking the Brain’s Navigational System,” in Proceedings of the International Conference on Neuromorphic Systems, New York, NY, USA, 2018, p. 4:1-4:8. doi: 10.1145/3229884.3229888.
[22] E. Nichols, L. J. McDaid, and N. Siddique, “Biologically Inspired SNN for Robot Control,” IEEE Trans. Cybern., vol. 43, no. 1, pp. 115–128, Feb. 2013, doi: 10.1109/TSMCB.2012.2200674.
[23] M. B. Milde et al., “Obstacle Avoidance and Target Acquisition for Robot Navigation Using a Mixed Signal Analog/Digital Neuromorphic Processing System,” Front. Neurorobotics, vol. 11, 2017, doi: 10.3389/fnbot.2017.00028.
[24] J. P. Mitchell, G. Bruer, M. E. Dean, J. S. Plank, G. S. Rose, and C. D. Schuman, “NeoN: Neuromorphic control for autonomous robotic navigation,” in 2017 IEEE International Symposium on Robotics and Intelligent Sensors (IRIS), Oct. 2017, pp. 136–142. doi: 10.1109/IRIS.2017.8250111.
[25] C. Jimenez-Romero and J. Johnson, “SpikingLab: modelling agents controlled by Spiking Neural Networks in Netlogo,” Neural Comput. Appl., vol. 28, no. 1, pp. 755–764, Dec. 2017, doi: 10.1007/s00521-016-2398-1.
[26] A. Cyr, F. Thériault, M. Ross, N. Berberian, and S. Chartier, “Spiking Neurons Integrating Visual Stimuli Orientation and Direction Selectivity in a Robotic Context,” Front. Neurorobotics, vol. 12, 2018, doi: 10.3389/fnbot.2018.00075.
[27] Z. Bing, C. Meschede, F. Röhrbein, K. Huang, and A. C. Knoll, “A Survey of Robotics Control Based on Learning-Inspired Spiking Neural Networks,” Front. Neurorobotics, vol. 12, Jul. 2018, doi: 10.3389/fnbot.2018.00035.
[28] S. A. Lobov, A. N. Mikhaylov, M. Shamshin, V. A. Makarov, and V. B. Kazantsev, “Spatial Properties of STDP in a Self-Learning Spiking Neural Network Enable Controlling a Mobile Robot,” Front. Neurosci., vol. 14, 2020, doi: 10.3389/fnins.2020.00088.
[29] J. S. Taube, R. U. Muller, and J. B. Ranck, “Head-direction cells recorded from the postsubiculum in freely moving rats. I. Description and quantitative analysis,” J. Neurosci., vol. 10, no. 2, pp. 420–435, Feb. 1990, doi: 10.1523/JNEUROSCI.10-02-00420.1990.
[30] S. M. Stringer, T. P. Trappenberg, E. T. Rolls, and I. E. T. d Araujo, “Self-organizing continuous attractor networks and path integration: one-dimensional models of head direction cells,” Netw. Comput. Neural Syst., vol. 13, no. 2, pp. 217–242, Jan. 2002, doi: 10.1080/net.13.2.217.242.
[31] A. Arleo and W. Gerstner, “Modeling Rodent Head-direction Cells and Place Cells for Spatial Learning in Bio-mimetic Robotics,” in In, pp. 236–245.
[32] M. J. Milford, G. F. Wyeth, and D. Prasser, “RatSLAM: a hippocampal model for simultaneous localization and mapping,” in IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA ’04. 2004, Apr. 2004, vol. 1, pp. 403-408 Vol.1. doi: 10.1109/ROBOT.2004.1307183.
[33] T. Kyriacou, “An Implementation of a Biologically Inspired Model of Head Direction Cells on a Robot,” in Towards Autonomous Robotic Systems, Berlin, Heidelberg, 2011, pp. 66–77. doi: 10.1007/978-3-642-23232-9_7.
[34] Y. Zhou and D. Wu, “Biologically inspired model of path integration based on head direction cells and grid cells,” Front. Inf. Technol. Electron. Eng., vol. 17, no. 5, pp. 435–448, May 2016, doi: 10.1631/FITEE.1500364.
[35] T. Degris, L. Lacheze, C. Boucheny, and A. Arleo, “A Spiking Neuron Model of Head-Direction Cells for Robot Orientation”.
[36] P. Stratton, M. Milford, J. Wiles, and G. Wyeth, “Automatic calibration of a spiking head-direction network for representing robot orientation,” in Proceedings of the 2009 Australasian Conference on Robotics and Automation, S. Scheding, Ed. Australia: Australian Robotics & Automation Association, 2009, pp. 1–8. Accessed: Jul. 14, 2021. [Online]. Available: http://www.araa.asn.au/acra/acra2009/
[37] R. Kreiser, M. Cartiglia, J. N. P. Martel, J. Conradt, and Y. Sandamirskaya, “A Neuromorphic Approach to Path Integration: A Head-Direction Spiking Neural Network with Vision-driven Reset,” in 2018 IEEE International Symposium on Circuits and Systems (ISCAS), May 2018, pp. 1–5. doi: 10.1109/ISCAS.2018.8351509.
[38] N. Qiao et al., “A reconfigurable on-line learning spiking neuromorphic processor comprising 256 neurons and 128K synapses,” Front. Neurosci., vol. 9, 2015, doi: 10.3389/fnins.2015.00141.
[39] Y. Sandamirskaya and G. Schöner, “An embodied account of serial order: How instabilities drive sequence generation,” Neural Netw., vol. 23, no. 10, pp. 1164–1179, Dec. 2010, doi: 10.1016/j.neunet.2010.07.012.
[40] S. Amari, “Dynamics of pattern formation in lateral-inhibition type neural fields,” Biol. Cybern., vol. 27, no. 2, pp. 77–87, Jun. 1977, doi: 10.1007/BF00337259.
[41] R. Kreiser, D. Aathmani, N. Qiao, G. Indiveri, and Y. Sandamirskaya, “Organizing Sequential Memory in a Neuromorphic Device Using Dynamic Neural Fields,” Front. Neurosci., vol. 12, 2018, doi: 10.3389/fnins.2018.00717.
[42] Y. Huang-Yu, H.-P. Huang, Y.-C. Huang, and C.-C. Lo, “Flyintel – a Platform for Robot Navigation based on a Brain-Inspired Spiking Neural Network,” in 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS), Mar. 2019, pp. 219–220. doi: 10.1109/AICAS.2019.8771614.
[43] Y.-C. Huang et al., “A Single-Cell Level and Connectome-Derived Computational Model of the Drosophila Brain,” Front. Neuroinformatics, vol. 12, 2019, doi: 10.3389/fninf.2018.00099.
[44] F. K. Hsieh, “Computational neuroethology － Developing a virtual worm system and its energy homeostasis mechanism,” Master Thesis, National Tsing Hua University, 2016. [Online]. Available: https://etd.lib.nctu.edu.tw/cgi-bin/gs32/hugsweb.cgi?o=dnthucdr&s=id=%22GH02102080510%22.&searchmode=basic
[45] R. Han, T.-M. Wei, S.-C. Tseng, and C.-C. Lo, “Characterizing approach behavior of Drosophila melanogaster in Buridan’s paradigm,” PLOS ONE, vol. 16, no. 1, p. e0245990, Jan. 2021, doi: 10.1371/journal.pone.0245990.
[46] L. Meier, D. Honegger, and M. Pollefeys, “PX4: A node-based multithreaded open source robotics framework for deeply embedded platforms,” in 2015 IEEE International Conference on Robotics and Automation (ICRA), May 2015, pp. 6235–6240. doi: 10.1109/ICRA.2015.7140074.
[47] F. Furrer, M. Burri, M. Achtelik, and R. Siegwart, “RotorS—A Modular Gazebo MAV Simulator Framework,” in Robot Operating System (ROS): The Complete Reference (Volume 1), A. Koubaa, Ed. Cham: Springer International Publishing, 2016, pp. 595–625. doi: 10.1007/978-3-319-26054-9_23.
[48] M. Quigley et al., “ROS: an open-source Robot Operating System,” p. 6.
[49] N. Koenig and A. Howard, “Design and use paradigms for Gazebo, an open-source multi-robot simulator,” in 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE Cat. No.04CH37566), Sep. 2004, vol. 3, pp. 2149–2154 vol.3. doi: 10.1109/IROS.2004.1389727.
[50] S. Heinze, “Unraveling the neural basis of insect navigation,” Curr. Opin. Insect Sci., vol. 24, pp. 58–67, Dec. 2017, doi: 10.1016/j.cois.2017.09.001.
[51] K. Pfeiffer and U. Homberg, “Organization and Functional Roles of the Central Complex in the Insect Brain,” Annu. Rev. Entomol., vol. 59, no. 1, pp. 165–184, 2014, doi: 10.1146/annurev-ento-011613-162031.
[52] B. Webb and A. Wystrach, “Neural mechanisms of insect navigation,” Curr. Opin. Insect Sci., vol. 15, pp. 27–39, Jun. 2016, doi: 10.1016/j.cois.2016.02.011.
[53] J. D. Seelig and V. Jayaraman, “Feature detection and orientation tuning in the Drosophila central complex,” Nature, vol. 503, no. 7475, Art. no. 7475, Nov. 2013, doi: 10.1038/nature12601.
[54] J. D. Seelig and V. Jayaraman, “Neural dynamics for landmark orientation and angular path integration,” Nature, vol. 521, no. 7551, Art. no. 7551, May 2015, doi: 10.1038/nature14446.
[55] M. Collett, L. Chittka, and T. S. Collett, “Spatial Memory in Insect Navigation,” Curr. Biol., vol. 23, no. 17, pp. R789–R800, Sep. 2013, doi: 10.1016/j.cub.2013.07.020.
[56] B. Liu, A. J. White, and C.-C. Lo, “Augmenting Flexibility: Mutual Inhibition Between Inhibitory Neurons Expands Functional Diversity,” bioRxiv, p. 2020.11.08.371179, Apr. 2021, doi: 10.1101/2020.11.08.371179.