Skip to content

Research and Development in Neuromorphic AI for an Anthropomorphic Android to Achieve a Human-centered AI Society

Brief description

The goal of this project is about the research and development in neuromorphic AI for an anthropomorphic android. The project is based on an innovative approach to neuromorphic AI control by focusing on developing an ultra-low-power neuromorphic AI chip, which will be prototyped initially on an FPGA and later transitioned into an ASIC. The originality of this work lies in its emphasis on achieving energy efficiency and computational reliability—two critical features necessary for advancing modern AI systems. The proposed chip will incorporate novel design strategies to address these challenges, ensuring its suitability for integration into anthropomorphic AIzuHand and assistance anthropomorphic robots. By providing the robots with enhanced intelligence and operational efficiency, this work aims to push the boundaries of autonomous systems. The goal is to create a groundbreaking prototype that showcases cutting-edge technology and sets new benchmarks for performance and sustainability in neuromorphic AI applications. We plan to create new customer value and also a new social value with a view to social implementation.

Results

Patents

  1. Khanh N. Dang, Ryoji Kobayashi[^1], Yuga Hanyu[^1], Zhishang Wang, A. Ben Abdallah, "Candidate Operator Narrowing Method", Japan patent, (filed patent).
  2. Khanh N. Dang, Ryoji Kobayashi[^1], Yuga Hanyu[^1], Zhishang Wang, A. Ben Abdallah, "Method for calculating the neural network processor and weight data redownload cycle", Japan patent, (filed patent).
  3. Khanh N. Dang, Ryoji Kobayashi[^1], Yuga Hanyu[^1], Zhishang Wang, A. Ben Abdallah, "A neural network processor and a method for selecting neurons to be pruned", Japan patent, (filed patent).

Papers

  1. Van-Vu Luyen, Thanh-Dat Nguyen, Quang-Thai Pham, Duy-Anh Nguyen, Khanh N. Dang, Van-Hai Pham, Dao Thanh Toan, "Energy-Efficient Izhikevich Neuron Design Using Approximate CORDIC-Based Multipliers for Low-Power Neuromorphic Hardware", IEEE Access, vol. 14, pp. 22146-22161, 2026. [DOI: 10.1109/ACCESS.2026.3662681]/[PDF]/[Source Code]
  2. Ryoji Kobayashi, Ngo-Doanh Nguyen, Abderazek Ben Abdallah, Nguyen Anh Vu Doan, and Khanh N. Dang, "ApproxiMorph: Energy-efficient Neuromorphic System with Layer-wise Approximation of Spiking Neural Networks and 3D-Stacked SRAM", IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 45, no. 3, pp. 1182-1196, March 2026. [DOI: 10.1109/TCAD.2025.3597251]/[PDF].
  3. Hanyu Yuga, Subbaiah Ravi Hariprakash, Abderazek Ben Abdallah, Zhishang Wang, and Khanh N. Dang, "GreenMorph: Sustainable Neuromorphic Computing through Energy-Harvesting and Energy-Driven Online STDP Learning", 2026 IEEE International Symposium on Circuits and Systems (ISCAS), Lecture Presentation, May 24-27, 2026.
  4. Satvik Ganesh, Hanyu Yuga, Zhishang Wang, and Khanh N. Dang, "FSPC: Spike Compression Through Correlated-AER Merging In Spiking Neural Networks", 2025 IEEE 18th International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC), Dec. 15-18, 2025.
  5. Hanyu Yuga, Zhishang Wang, and Khanh N. Dang, "Evolutionary Algorithm for STDP-based Spiking Neural Network Model Compression", 2025 IEEE 18th International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC), Dec. 15-18, 2025.