FPGA-implemented corner feature extracting simultaneous localization and mapping

Abstract

In this paper, a novel corner feature extracting simultaneous localization and mapping (CFESLAM) algorithm is proposed, which employs a mechanism of corner feature extraction that regards only corners as the landmarks to ease the computational burden. To further improve the overall computational efficiency, the proposed CFESLAM is implemented by hardware on a FPGA platform. By doing so, the proposed approach is capable of providing a robust, reliable, and low-cost real-time SLAM system. Experimental results have confirmed that notable improvement is made by using the proposed approach.