An Implementation of Simultaneous Localization and Mapping Using Dynamic Field Theory

Abstract

Simultaneous Localization and Mapping (SLAM) algorithms are commonly used for robotic navigation and spatial awareness. A key challenge with SLAM is the large memory cost and associated computational overhead. In this study, we examine how to implement a lower computational cost version of SLAM by utilizing Dynamic Field Theory (DFT). This implementation performs key SLAM tasks with similar accuracy but with 1/5 of the memory cost of other common SLAM algorithms. Future work involves transitioning this algorithm to a physical platform with neuromorphic hardware for an energy efficient robotics solution.