Martin C. Martin

Papers

Total Citations

H-Index

About

Martin C. Martin is a researcher whose work bridges artificial intelligence, robotics, and computer vision, with a particular focus on applying evolutionary and machine learning techniques to real-world robotic control. His most cited paper, "Genetic Programming for Robot Vision" (2002, 7 citations), pioneered the use of genetic programming to automatically generate the vision subsystem for a mobile robot's reactive obstacle avoidance system. By evolving algorithms that captured the spirit of traditional hand-written vision operators, Martin demonstrated how computational intelligence could automate the design of perception systems, reducing the need for manual engineering. In his follow-up work, "Controlling Cardea: Fast Policy Search in a High Dimensional Space" (2004, 4 citations), he introduced the essential dynamics algorithm—a novel policy search method for stochastic Markov decision processes with continuous state and action spaces. He applied this to control a complex 5-degree-of-freedom robot arm mounted on a Segway base, tackling the challenging dynamics of a moving platform. Though his citation counts are modest, Martin’s contributions are notable for their technical ambition, combining evolutionary computation, high-dimensional policy search, and real-time robotic control in ways that anticipated later advances in learned robotics and automated vision systems.