Design and implementation of embedded multi-sensor outdoor robot localization system using FPGA

Abstract

Recently, there has been a significant increase in the interest for robots. However, manually navigating a robot requires a skilled pilot who has high constant concentration levels for sustained periods. Therefore, strong scientific interest has emerged in terms of developing solutions that allow a robot to navigate autonomously without needing constant human supervision. This is useful for a variety of potential applications ranging from surveillance and reconnaissance purposes, aerial filming, remote visual inspection of industrial sites and military applications as well. Unlike humans, a robot does not have the sensing capacity and the ability to explore its environment and determine location. Therefore, sensors such as Infrared, ultrasonic, accelerometer and GPS need to be integrated on the robot to improve its environmental perception. The robot then needs be able to determine its destination and find shortest path to get there. Dijkstra algorithm is implemented to examine all related nodes along the robot’s path towards destination. The robot is marking nodes as visited and unvisited to decide which path is proper to be followed. Furthermore, the robot has the capacity to detect obstacles within path toward destination. FPGA DE0 Nano board is used because of its portability and extensive computational power that overcome the delay and power consumption in the project. There are three modules in a path planning approach - the first generates a route to the desired location, the second detects if there are obstacles, and the third is for collision avoidance. The GPS sensor was utilised to determine the positions for the current and desired locations. DE0-Nano sent proper signals to the L293D to control robot’s motors according to sensor signals. The result revealed that the output reading voltage of the IR sensor exhibited high reflectivity for white coloured objects compared to other colours like black and blue. The IR and ultrasonic sensors already calibrated to detect obstacles and keep the distance between the robot and obstacles from 30 cm to 50 cm. The DE0 Nano achieved a maximum operating frequency up to 1.3 GHz and total logic elements were 6,032. This means that the robot’s platform’s frequency requirements were achieved at 1.3 GHz.