Christiaan J. Pretorius
Papers
11
Total Citations
159
H-Index
9
About
Christiaan J. Pretorius is a pioneering researcher in Evolutionary Robotics (ER), whose career has been defined by a singular and transformative question: can artificial neural networks replace conventional physics-based simulators in robotic evolution? Beginning with foundational work in 2009, Pretorius systematically explored the viability of neural network-based simulators, arguing that their noise-tolerance and generalization capabilities make them well-suited alternatives to the complex, resource-intensive models traditionally employed in ER. His most cited work, "Simulating Robots Without Conventional Physics" (2012, 29 citations), crystallized this vision and set the stage for a productive research program spanning diverse robotic platforms — from differentially-steered robots and snake-like locomotion systems to hexapods and inverted pendulums. A particularly notable contribution is his concurrent co-evolution framework, in which both the controller and simulator neural networks develop simultaneously, reducing dependence on pre-built physical models. With cumulative citations exceeding 150 across his key publications, Pretorius has established himself as a consistent voice in demonstrating that data-driven simulation offers a credible, accessible pathway for advancing robotic evolution research. His work remains especially relevant as the field grapples with the sim-to-reality transfer problem.
Research Focus
Key Achievements
Top Papers
- 1Simulating Robots Without Conventional Physics: A Neural Network Approach29 citations · 2012
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10Neural Networks for Mobile Robot Inverse Kinematics7 citations · 2018