Max Robotics — Global Service Robot Directory + US Certification

About

Amirhossein Taghvaei is a researcher whose work sits at the intersection of nonlinear filtering, stochastic control, and bio-inspired robotics. His primary contributions lie in advancing the theory and application of the feedback particle filter (FPF), particularly for systems evolving on complex geometric structures. In his highly cited 2017 paper (36 citations), he generalized the FPF to connected Riemannian manifolds and matrix Lie groups, solving a critical problem for continuous-time nonlinear filtering in domains like attitude estimation. This work, alongside his 2016 paper on matrix Lie groups, provides rigorous mathematical foundations for filtering on non-Euclidean spaces, enabling more accurate state estimation in robotics and aerospace. Beyond filtering, Taghvaei has made notable strides in reinforcement learning for locomotion. His 2019 work on Q-learning for partially observable Markov decision processes (POMDPs) introduces a framework for learning optimal periodic gaits in coupled rigid-body systems, with applications to snake robots and other bio-inspired platforms. Extending this in 2020, he developed a central pattern generator (CPG)-type architecture for sensorimotor control, bridging biological principles with machine learning. With a total of over 50 citations across these key papers, Taghvaei’s research is shaping how robots perceive and move through the world, offering elegant mathematical solutions to real-world control challenges.