John Fischer
Papers
2
Total Citations
24
H-Index
2
About
John Fischer’s research lies at the intersection of robotics, neural networks, and real-time control systems, with a focus on enabling autonomous learning in resource-constrained environments. His most notable contributions include pioneering work on connectionist learning for physical robots, where he demonstrated that neural networks could learn complex motor tasks—such as trailer backing and inverted pendulum balancing—directly on board miniature robots with severely limited computing power and memory. In his 2002 paper on fast connectionist learning for trailer backing (14 citations), Fischer showed that an autonomous mini-robot could learn a non-linear control task in real time, despite the constraints of short battery life and minimal onboard processing. His earlier 1999 paper on a neural network pole balancer (10 citations) remains a landmark in real-time robot learning, proving that a physical robot could learn to stabilize an inverted pendulum without external computation. These works have influenced subsequent research in embedded machine learning and low-resource robotics, highlighting Fischer’s role in advancing practical, deployable intelligent systems.
Research Focus
Key Achievements
Top Papers
- 1Fast connectionist learning for trailer backing using a real robot14 citations · 2002
- 2