Max Robotics — Global Service Robot Directory + US Certification

About

M. Howarth is a pioneering researcher in intelligent robotic assembly, specializing in autonomous systems that learn and adapt in unstructured environments. Their work centers on merging neural network architectures, particularly Adaptive Resonance Theory (ART), with reinforcement learning to enable robots to acquire manipulative skills online. Howarth’s most significant contribution is the development of a novel neural network controller (NNC) that allows self-adapting robots to learn complex assembly tasks without explicit programming, as detailed in their highly cited 2002 paper (16 citations). They also introduced a geometrically validated approach to robotic assembly (2003, 15 citations), demonstrating how real-time force and torque data can guide component mating. Howarth’s foundational research on task-level programming (1998, 11 citations) and contact localization (2002, 6 citations) further advanced flexible, nonlinear automation. With over 70 total citations, their work has shaped modern approaches to autonomous assembly, offering practical solutions for reducing uncertainty in mechanical component interaction. Howarth’s ART-R algorithm (2003) remains a notable achievement, merging fast, stable learning with reinforcement learning for state representation.