Neuromorphic Sensing Data Transmissions and Soft Actuators

Neuromorphic architectures can enable significant power efficiencies in machine learning tasks. The use of soft, organic materials can allow for integration of sensing, actuation and data transmission onto neuromorphic platforms. In this talk, I will share our work on developing asynchronous data transmission systems¹ for neuromorphic in combination with soft flexible electronic materials and devices². Using such systems, we can generate unique datasets containing high speed sensory data, for e.g., a MNIST like tactile dataset³ and apply them to use cases in robotics^4,5.<br/><br/><br/><b>References </b><br/><br/>1. van de Burgt, Y., Melianas, A., Keene, S.T. <i>et al.</i> Organic electronics for neuromorphic computing. <i>Nat Electron</i> <b>1</b>, 386–397 (2018). https://doi.org/10.1038/s41928-018-0103-3<br/>2. Lee, W. W., Tan, Y. J., Yao, H., Li, S., See, H. H., Hon, M., Ng, K. A., Xiong, B., Ho, J. S., & Tee, B. C. K. (2019). A neuro-inspired artificial peripheral nervous system for scalable electronic skins. <i>Science Robotics</i>, <i>4</i>(32), eaax2198. https://doi.org/10.1126/scirobotics.aax2198<br/>3. Yao, H., Yang, W., Cheng, W., Tan, Y. J., See, H. H., Li, S., Ali, H. P. A., Lim, B. Z. H., Liu, Z., & Tee, B. C. K. (2020). Near–hysteresis-free soft tactile electronic skins for wearables and reliable machine learning. <i>Proceedings of the National Academy of Sciences</i>, 202010989. https://doi.org/10.1073/pnas.2010989117<br/>4. ST-MNIST -- The Spiking Tactile MNIST Neuromorphic Dataset, 2020, https://arxiv.org/abs/2005.04319<br/>5. Taunyazov, T., Sng, W., See, H. H., & Lim, B. (2020). Event-Driven Visual-Tactile Sensing and Learning for Robots. <i>Robotics: Science and Systems</i>.