Abstract :
Summary form only given. This paper presents a Hardware/Software Co-design of a probabilistic algorithm such as Kalman Filter (KF). In this case we present a solution for calculating the path length, which can be used for navigation techniques used in mobile robotics. The KF algorithm has been implemented and run on an Altera Ciclone II FPGA with a Nios II processor. To achieve this, first of all we had to do the modeling of our mobile system applying concepts of odometry, considering the readings of the encoders on our mobile platform Pioneer 3AT (P3AT). Then we proceeded to develop the probabilistic algorithm of the Kalman Filter from our system model using as additional measure the information of path length provided by the internal robot system. This probabilistic algorithm is based on the concept of sensor fusion, allowing us to get the better estimate of the value of the path length. At the end, we show a comparison of the algorithm performance developed in software and hardware independently.
Keywords :
Kalman filters; distance measurement; field programmable gate arrays; hardware-software codesign; mobile robots; path planning; probability; reconfigurable architectures; sensor fusion; Altera Ciclone II FPGA; KF algorithm; Kalman filter; Nios II processor; P3AT; hardware-software codesign; internal robot system; mobile platform Pioneer 3AT; mobile robotics; mobile system; navigation techniques; odometry; path length; probabilistic algorithm; reconfigurable architectures; sensor fusion; Algorithm design and analysis; Hardware; Mobile communication; Probabilistic logic; Robots; Software; Software algorithms;
