B.G. Horne

About

B.G. Horne is a pioneering researcher in the intersection of neural networks and robotics, whose work in the late 20th and early 21st centuries laid foundational groundwork for intelligent control systems. Horne’s primary research areas include neural network-based control, robotic manipulation, and connectionist learning paradigms. Their most notable contribution is a 1990 study (11 citations) that demonstrated the first application of a multilayer perceptron (MLP) for position control of a two-link robot, addressing critical issues such as network architecture—layer count and node density—and computational efficiency. This work provided early evidence that neurocontrollers could handle complex robotic dynamics, influencing subsequent research in adaptive control. Horne further advanced the field with a 2003 study (6 citations) introducing a memory-based connection network for robotic gripper control, which learned relationships between control effort and state changes. Though achieving only partial success in implementing step responses, this work explored velocity-based learning, offering a novel paradigm for adaptive manipulation. While citation counts are modest, Horne’s contributions are significant for their early integration of neural networks into robotics, inspiring later advances in intelligent automation and serving as a reference for researchers exploring neural control architectures.

Research Focus

Key Achievements

Top Papers

1. 1
   A neural network-based controller for a two-link robot

   11 citations · 1990
2. 2
   A connection network for robotic gripper control

   6 citations · 2003

Key Collaborators

Contact & Links