Hw/Sw co-design for fixed point EKF-SLAM

• Grant number: 12/03777-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): May 01, 2012
• Effective date (End): November 30, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science
• Principal Investigator: Vanderlei Bonato
• Grantee: Leandro de Souza Rosa
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated scholarship(s): 12/20224-8 - A fixed-point EKF-SLAM hardware for embedded robotics, BE.EP.IC

Abstract

Localization and mapping are two of the most important capabilities of autonomous mobile robots and an important method for dealing with these problems is the Extended Kalman Filter (EKF - Extended Kalman Filter), which simultaneously maps the environment, estimating a model for this and the robot's position. This problem is known in the literature as SLAM (simultaneous localization and mapping). In contrast, the EKF, is typically implemented using floating-point representations, possesses a higher-order computer ((n2)) where n is the number of elements on the map) so that, for the large map, requires a computer system with high computational power to handle the solution of the SLAM-time navigation (online). To contour these restrictions, this project aims to implement a hardware architecture and software for the EKF-SLAM with fixed-point arithmetic in hardware / FPGA (Field-Programmable Gate Array), allowing it to be embedded in mobile robots to take advantage of better computational resources. (AU)