Max Robotics — Global Service Robot Directory + US Certification

About

Antonio Berlanga is a researcher specializing in evolutionary computation, neural network-based robot control, and autonomous navigation systems. His work, concentrated primarily in the early 2000s, made meaningful contributions to the challenge of training neural network controllers for mobile robots operating in dynamic environments — a problem made difficult by the complexity of generating effective training data for real-world navigation scenarios. Berlanga's most significant contribution is the development of Uniform Coevolution, a general-purpose coevolutionary framework that simultaneously evolves neural network controllers and the test environments used to evaluate them. This competitive co-adaptation approach addressed a fundamental limitation in evolutionary robotics: the tendency for learned behaviors to overfit narrow training conditions. His most-cited work, "Neural networks robot controller trained with evolution strategies" (2003, 18 citations), demonstrated how evolution strategies could effectively replace traditional learning methods for weight optimization in robotic controllers. Across his body of work, Berlanga tackled obstacle avoidance, goal-directed navigation, and behavioral generalization using Braitenberg vehicles and classifier systems. His research bridged symbolic and subsymbolic approaches to robot learning, including automatic symbolic modeling of evolved skills. Collectively accumulating over 60 citations, his contributions laid useful groundwork for adaptive, self-learning autonomous systems.