Max Robotics — Global Service Robot Directory + US Certification

About

Pedram Agand is a researcher whose work sits at the intersection of robotics, control theory, and adaptive machine learning. His primary contributions lie in developing stable, real-time neural network architectures for robot dynamics identification and control. Agand’s most cited work, “Adaptive recurrent neural network with Lyapunov stability learning rules for robot dynamic terms identification” (44 citations), introduces a groundbreaking approach that uses continuous Lyapunov functions to guarantee uniform ultimate boundedness (UUB) stability during online learning—a critical advancement over traditional batch algorithms that struggle with model mismatches and disturbances. This theme of stability-guaranteed adaptation is further explored in his 2019 paper on adaptive model learning (9 citations). Agand has also made notable contributions to teleoperation under uncertainty, proposing decentralized robust control for needle insertion with communication delays (15 citations). His practical engineering impact is evident in his work on vision-based kinematic calibration of spherical robots (8 citations), which offers an accurate, low-cost method using a single camera. Additionally, his 2016 paper on transparent, flexible neural networks (7 citations) introduces a novel multilayer perceptron with adaptive activation functions, enabling data-driven identification of robot dynamics from end-effector measurements. Through these works, Agand has established himself as a key figure in creating theoretically rigorous, practically deployable AI for robotic systems.