Implementation of a Person Following Robot in ROS-gazebo platform

This research work involves the implementation of a person follower robot in ROS_Gazebo platform As part of this research work, we designed a mobile robot (for simulation)that can follow human beings. It assists people carrying heavyweights to their desired locations by following them. The entire simulation work is carried out with the help of ROS2 (Robot Operating System Version2). We used Gazebo and RVIZ as simulation tools to verify the outputs. MakeHuman and Blender software's assisted in creating an animated person and verify the rendering in the gazebo. The mobile robot created makes use of a Depth camera (Kinect) to detect a person and Lidars to calculate the distance between the robot and person. YOLOv3 algorithm has been used for image recognition. With this mobile robot, we simulated with multiple persons in gazebo software to make sure the robot is following only one particular person. The Kinect sensor uses object detection algorithms to identify and follow a particular person.

I. Introduction

Robotics is a vast area of research, which holds knowledge from many different streams. Robotics is extensively used in manufacturing industries to accomplish some tasks which are hazardous to mankind. Robotics is helping mankind in every aspect to accomplish their tasks with ease; it replicates humans with its machine intelligence. “Follow-me (Human Following) Robot” is one such kind of robot, it assists a human by carrying heavy loads and follow him/her to their desired location. The created mobile robot has a Kinect depth camera sensor to detect a specific person and follow him; it uses an Image recognition alogirithm to accomplish the person detection task.