Patrick MacAlpine
Papers
24
Total Citations
550
H-Index
13
About
Patrick MacAlpine is a pioneering researcher at the intersection of robotics, machine learning, and autonomous multi-agent systems, with a particular focus on humanoid robot locomotion and soccer-playing agents. Best known for his work with the UT Austin Villa RoboCup team, MacAlpine has made foundational contributions to how robots learn to walk, coordinate, and compete in dynamic environments. His award-winning research on omnidirectional humanoid walking and overlapping layered learning — a hierarchical machine learning paradigm that enables robots to master complex behaviors by building incrementally on simpler sub-skills — has shaped modern approaches to robot skill acquisition. A recurring theme in his work is bridging the gap between simulation and real-world performance, demonstrating how optimized parameters can transfer across both domains. His interdisciplinary 2018 study connecting infant motor development to AI robot learning (71 citations) exemplifies his creative reach beyond traditional robotics. Additional contributions to role assignment algorithms like SCRAM highlight his expertise in multi-robot coordination. Collectively, his body of work — spanning championship competition victories and widely cited publications — has significantly advanced the field of autonomous humanoid robotics and reinforcement learning.
Research Focus
Key Achievements
Top Papers
- 1Variety Wins: Soccer-Playing Robots and Infant Walking71 citations · 2018
- 2
- 3
- 4
- 5
- 6Overlapping layered learning36 citations · 2017
- 7
- 8
- 9
- 10