Max Robotics — Global Service Robot Directory + US Certification

About

Antony Waldock’s research lies at the intersection of reinforcement learning, fuzzy systems, and autonomous robot navigation. His most influential work, “Fuzzy Q-Learning with an adaptive representation” (2008, 15 citations), introduced a novel method for extending Q-learning to large or continuous state spaces, enabling robots to learn optimal control policies from high-dimensional sensor data. This foundational contribution has been applied across domains from data mining to robot control. Waldock further advanced hierarchical fuzzy systems, developing adaptive rule-based controllers that allow robots to learn complex behaviors through layered decision-making, as demonstrated in his 2015 paper on learning robot controllers (8 citations). His information-theoretic approach to hierarchical fuzzy rule systems (2006) provided a principled method for automatically structuring rule bases, reducing manual tuning. In the realm of long-term autonomous navigation, Waldock proposed OFLAAM (Optical Flow Localisation and Appearance Mapping, 2015, 3 citations), a lightweight visual navigation system designed specifically for Micro Air Vehicles. By combining downward-facing optical flow, IMU data, and forward-facing monocular cameras, OFLAAM enables robust, drift-free navigation without relying on expensive sensors. Waldock’s work bridges theoretical reinforcement learning with practical, deployable robotic systems, making him a key figure in adaptive, learning-based autonomous navigation.