Max Robotics — Global Service Robot Directory + US Certification

About

David DeMers has made foundational contributions to the field of robotic manipulation, particularly in solving the complex inverse kinematics problem for redundant and dexterous manipulators. His research focuses on learning global properties of robot kinematic mappings, addressing the fundamental challenges of ill-posedness and multiple solution branches that arise when mapping end-effector positions to joint configurations. DeMers pioneered the use of unsupervised learning and clustering techniques to partition configuration spaces, enabling robots to learn global direct inverse kinematics from sample data alone—a bootstrap method that bypasses traditional analytical solutions. His most influential work, "Inverse Kinematics of Dextrous Manipulators" (1997), has garnered 31 citations, while his earlier 1991 paper on learning global direct inverse kinematics remains a cornerstone with 25 citations. DeMers also advanced the field by introducing regularization strategies for redundant manipulators, tackling both global ill-posedness and excess degrees of freedom. His work on exploiting topological properties of kinematic mappings for neural network-based control (1993, 2002) has shaped modern approaches to configuration control. Through these contributions, DeMers has established himself as a key figure in bridging robotics, machine learning, and topology, offering elegant solutions to one of robotics' most persistent challenges.