The Simulatious Localisation And Mapping (SLAM) algorithm is described as an indispensive backbone for most autonomous robotic applications. After 30-years of SLAM research dominated by Extended Kalman Filtering with well-known consistency problems and lack of global stability results, a more solid Geometric Nonlinear SLAM Observer with global convergence has been recently proposed. Despite of this, there still is a significant knowledge gap in the lack of stability guaranteed nonlinear control system design for this complex vision-based dynamic robot control.Our objective is to introduce a paradigm change to autonomous system development by bringing our novel stability-guaranteed nonlinear subsystem-based control theory to solve this open problem addressed in high impact real-world scenarios. Moreover, we address MMLRA whole-body motion estimation problem for enhancing their vision-based controllability, and their seamless integration into future autonomous digitalized worksites.

2021

2023