Intelligent navigation of mobile robots in an aircraft inspection system

• Main Author: Lim, Samuel Jun Hao
• Other Authors: Soong Boon Hee
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/139590
• Institution: Nanyang Technological University

For a mobile robot to be able to help in detecting the defects of the lower parts of the aircraft, the mobile robot must be able to move independently and safely. In order to allow mobile robots to move around independently, it is essential for a them to know its location well. Mobile robots will tend to get lost easily if it does not able to localise itself in an environment. This would result them to be unable to tell which direction it is heading to. Therefore, localization plays a very important role in implementing on mobile robots. It is a process to know exactly where a mobile robot will be located with respect to its environment. There are various methods that have been presented over the years such as Bayesian localization algorithms which consist of Kalman Filter methods, Markov localization and Monte Carlo Algorithms. Each implementation method has its’ own advantages and disadvantages. In addition, ROS will be used in this project to control the robot because there will be an additional drone that will be integrated along with the mobile robot for aircraft inspection. Thus, it will be easier to use ROS for the communication of the robot with the drone. Selecting a suitable LiDAR Sensor to integrate with the mobile robot is an important step in localization. LiDAR Sensor acts as an “eye” to the mobile robot. Having a good LiDAR Sensor allow the mobile robot to detect the surrounding of the environment in Point Cloud Data. Before trying out the LiDAR Sensor on the physical robot, it is important the check the performance of the LiDAR Sensor on simulation first. This involved in setting up the airplane environment, creating of models and robot in simulation called Gazebo. There are many types of LiDAR Sensor in the market and we have chosen Hokuyo LiDAR Sensor initially. However, due to the drawbacks of the sensor, we have decided to change it to RSBpearl LiDAR Sensor. These point clouds data that are detected by the LiDAR Sensor will then save into a bagfile whereby an algorithm will be created to check for obstacles. It is important for the robot and the drone to be able to detect obstacles that were blocked in their path when it is moving to prevent collision. This algorithm will result the terminal to prompt a message to indicate whether is the robot and drone are cleared to moved. Thus, completing the frontend testing.