Simultaneous Localization and Mapping Based Indoor Delivery Robot

Quek, Wei Chiang (2020) Simultaneous Localization and Mapping Based Indoor Delivery Robot. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text QuekWeiChiang_Full Text.pdf Restricted to Registered users only Download (2MB)

Abstract

Rise of industry 4.0 and Internet of Things have cause the surge in interest related to robot capabilities and is impact. Many company have begun implementing robot to automate their productivity such as Amazon an E Commerce company, they uses plenty of indoor delivery robot to move packages within the warehouse. Indoor delivery robot found today, only operate in a single floor and is expensive and time consuming to setup. It uses an old system such as line follow which makes it difficult to implement in the industry, thus SLAM based robot may be the key to solve the problem. Current SLAM algorithm used in SLAM based robot are mostly use in a single storey environment, this project propose to implement SLAM based indoor delivery robot using two sensor capable of mapping and navigating a multi-storey building filled with moving obstacle. In this project, a multi-storey SLAM robot was implemented to reduce the setup time capable of mapping and navigating a multi-storey environment, and able to avoid multiple unregister obstacle within the environment as compared to the traditional approach. 2D-LiDAR sensors was used to map the whole building while perform obstacle avoidance during navigation. Furthermore, path planning algorithm was designed to work with multi-storey building capable of navigating an environment filled with obstacle not found within the map and is require to handle sudden appearance of an obstacle in front of the robot. The whole experiment was conducted under a simulated environment resemble a two storey hotel with mapped obstacle and un-map obstacle. The SLAM based indoor delivery robot is able to successfully mapped the environment of the two storey building using g-mapping and filtering for the robot to use during navigation. The robot is able to navigate to the objective without hitting the obstacle, even if the obstacle is suddenly put in front of the robot. The robot obstacle avoidance feature is able to stop the robot before it hit the obstacle. The program is able to publish the desire coordinate for the navigation system to guide the robot to its path. This allow the robot to save all the path and add more easily into the program. In conclusion, SLAM based Indoor Delivery robot is far superior than traditional robot and is able to reduce setup cost, speed up robot setup time, and operate in human environment.