Max Robotics — Global Service Robot Directory + US Certification

About

Nathan McDonald’s research lies at the intersection of machine learning, robotics, and neuromorphic computing, with a focus on architectures that move beyond conventional deep neural networks. His work champions **hyperdimensional computing (HDC)** and **modular, hierarchical learning** as pathways to more flexible, online, and biologically plausible intelligence. In his most cited paper (2021, 4 citations), McDonald explored HDC for robotics, demonstrating its advantages in transfer learning, sensor fusion, and network topology—offering a lightweight alternative to monolithic ANNs for real-time adaptation. He further advanced this vision with a 2023 study integrating complex-valued HDC with modular neural networks, tackling “on the fly” learning challenges that traditional deep learning struggles to address. His 2024 paper introduced a modular, hierarchical ML framework for sequential goal completion in robotics, such as navigating a maze to pick up a key and unlock a chest. Earlier, McDonald demonstrated hardware-based ANNs on constrained platforms (2014), using a zero-instruction-set-computer chip for efficient video change detection. Though his citation counts are modest, McDonald’s work is pioneering in its push toward scalable, low-power, and adaptive intelligence—key for next-generation autonomous systems.