C. Kozakiewicz

Papers

Total Citations

H-Index

About

C. Kozakiewicz is a pioneering researcher in the field of robotics, with a primary focus on neural network applications for robot motion control and calibration. Their work bridges artificial intelligence and mechanical engineering, addressing fundamental challenges in robotic manipulation and navigation. Kozakiewicz’s most influential contribution is a neural network approach to path planning for two-dimensional robot motion, which uses camera image feedback loops to generate collision-free paths in environments with polygonal obstacles (19 citations). This work demonstrates how neural networks can enable real-time, adaptive obstacle avoidance—a critical capability for autonomous robots. Their second major contribution is the development of a partitioned neural network architecture for inverse kinematic calculations in six-degree-of-freedom manipulators (18 citations). This innovative design uses a preprocessing layer and dedicated partition modules to achieve high learning accuracy, offering a parallel computing solution to complex robotic control problems. Kozakiewicz also contributed to robot calibration, developing a software-based method to correct positioning errors in direct drive SCARA robots caused by static arm deflection (5 citations). By constructing stiffness models to calculate joint angle corrections, this work improved precision in assembly tasks. Kozakiewicz’s research remains foundational for students and engineers working at the intersection of neural networks and robotics.