LIMITATIONS OF DERIVATIVE PATH IN NEUROMORPHIC PID CONTROLLERS AND ADAPTIVE SOLUTIONS
DOI:
https://doi.org/10.63034/esr-726Keywords:
Neuromorphic control, Spiking Neural Networks, PID controller, Input-Weighted Threshold Adaptation (IWTA), Adaptive control, Derivative limitationAbstract
This work investigates a notable limitation in neuromorphic PID controllers: the derivative path using multiple time constants. Current methods provide accurate derivative estimation only within a limited frequency range, leading to overestimation of low-frequency inputs and underestimation of high-frequency inputs. We propose an adaptive approach, leveraging Input-Weighted Threshold Adaptation (IWTA), to dynamically adjust the spiking thresholds of derivative neurons. This mechanism aims to improve derivative tracking across a broader frequency range, enhancing control precision and stability in neuromorphic controllers, particularly for dynamic systems such as quadrotors.
References
Clawson, J., Kary, B., & Orellana, J. (2022). Learning a Linear-Quadratic Regulator with Reward-Modulated Spike-Timing Dependent Plasticity for Flapping Wing Flight. IEEE Transactions on Neural Networks and Learning Systems.
Thakur, C. S., et al. (2018). Large-scale neuromorphic spiking array processors: A quest to mimic the brain. Frontiers in Neuroscience, 12, 891.
Pfeiffer, M., & Pfeil, T. (2018). Deep learning with spiking neurons: Opportunities and challenges. Frontiers in Neuroscience, 12, 774.
Sandamirskaya, Y. (2014). Dynamic neural fields as a step toward cognitive neuromorphic architectures. Frontiers in Neuroscience, 7, 276.
Eliasmith, C., & Anderson, C. (2003). Neural Engineering: Computation, Representation, and Dynamics in Neurobiological Systems. MIT Press.
Eliasmith, C., et al. (2012). A large-scale model of the functioning brain. Science, 338(6111), 1202–1205.
Tavanaei, A., et al. (2019). Deep learning in spiking neural networks. Neural Networks, 111, 47–63.
Chan, H., van Schaik, A., et al. (2021). Neuromorphic PID Control Using Spiking Neurons. In IEEE International Conference on Robotics and Automation (ICRA).
Hunsberger, E., & Eliasmith, C. (2015). Spiking deep networks with LIF neurons. arXiv:1510.08829.
Liu, H., & Wang, X. (2020). Adaptive threshold mechanisms improve spiking neural network stability. Neural Computation, 32(9), 1821–1845.
Davies, M., et al. (2018). Loihi: A neuromorphic manycore processor with on-chip learning. IEEE Micro, 38(1), 82–99.
Sandler, M., & Blum, J. (2019). Drone flight stabilization using neuromorphic control. Robotics and Autonomous Systems, 118, 1–12.
Schmuker, M., Pfeil, T., & Nawrot, M. (2014). A neuromorphic network for real-time odor classification. PNAS, 111(7), 2548–2553.
van Albada, S. J., et al. (2018). Performance comparison of the neuromorphic SpiNNaker platform and the NEST simulator. Frontiers in Neuroscience, 12, 291.
Yang, Y., & Chen, T. (2022). Improving derivative estimation in spiking neural networks using multitimescale synaptic dynamics. Neural Processing Letters, 54, 1299–1318.
Downloads
Published
How to Cite
Issue
Section
Categories
License
Copyright (c) 2026 Akzhol Adil
This work is licensed under a Creative Commons Attribution 4.0 International License.
