Integration of mapping and neural network for mobile robot with laser distance range sensor (LDS)

Kamaruzaman@Kamaruldin, Nur Fatini (2021) Integration of mapping and neural network for mobile robot with laser distance range sensor (LDS). Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Integration of mapping and neural network for mobile robot with laser distance range sensor (LDS).pdf - Submitted Version Download (1MB)

Abstract

Mobile robots are becoming increasingly popular. They have already moved out of laboratories. For instance, they are widely used not only in industry but also at home. The ability to map the surrounding environment is one of the most common tasks in autonomous navigation for a mobile robot. However, the problem is the accuracy of the build map depends on sensor measurements. Thus, noisy sensor measurements would produce a grid map that prone to error when using low-cost sensors. Therefore, machine learning algorithm integration using a neural network approach has been introduced to improve the quality of the map created and to evaluate the accuracy of the environmental grid map using a low-cost laser distance sensor (LDS). In this work, the neural network and the occupancy grid map algorithm will be experimented using the Turtlebot3 mobile robot. The process of conducting project is to collect sensor measurements, train the neural network, and test a neural network to build an occupancy grid map. Throughout the project, the software that will be used are Ubuntu, ROS, Gazebo, RViz while the hardware is Turtlebot3 and LDS sensor. At the end of the project, a mobile robot able to build a high-accuracy environmental map.

Actions (login required)

View Item