Abstract:
During the previous decades, a tremendous growth in the field ofrobotics automation
has been made. Daily activities involve interaction with smart machines that present
a certain level of autonomy. Robots ensure that a task can be done more accurately
and efficiently. They can perform the repetitive task without any difficulty. In this
autonomous robot, finding object in an unexplored area is a major mission in this work,
for advance work the robot will pick the required object. A self-customized mobile
robot having low-cost and low-power equipment’s will be utilized in this project. The
Robot Operating System or ROS is being executed with the help of Jetson Nano.
Mapping is established using Rplidar . A mobile robot structure planning for object
detection and navigation is being developed in this project. This work proposes a pose
planning method to observe a target object using 2D information from Rplidar. The
proposed planning method is utilized under the ROS environment; the ROS
Navigation stacks nodes help to process sensor information. The mobile robot
navigation is handled by the ROS Navigation stacks. 2D ROS navigation stack takes
information of odometry, sensors and velocity of wheels to send to mobile robot. To
be able to work on navigation stack the robot must be running ROS which has tf
transform tree to place and publish sensor data using the correct ROS messages
command. ROS navigation stack is also needed to configure robot pose, shape and
structure to perform the high level tasks. The object detection model is trained through
python. The camera will find where in the frame object is located by extracting its
bounding boxes. Then it follows the goal point ofthe map generated by SLAM. In the
end, robot moves towards the object and pick it up with the help ofrobotic arm
