Salt-Pepper Project

This is the Robotics Project (VIBOT 3rd semester, Heriot-Watt University) done by Paola Ardon([email protected]) Kaisar Kushibar([email protected]), and Songyou Peng ([email protected]).

Requirements

• ROS
• Python 2.7
• OpenCV
• ros-naoqi
• ros-teleop
• ORB SLAM 2

Demo

The demonstration of SLAM + Object recognition with Pepper robot can be found here.

How to run

First of all, start ROS and connect the host computer with Pepper:

$ roscore
$ roslaunch pepper_bringup pepper_full_py.launch nao_ip:="ROBOT_IP" roscore_ip:="HOST_IP"

Run JoyPepper with Autonomous life ON. This will allow to control the robot using the joystick, but Pepper behaves like a kid (if she gets distracted, she just stops listening you)

$ rosrun joy joy_node
$ rosrun joy_pepper joypepper.py

If you want to control Pepper with Joystick with Autonomous life OFF, add this line before:

$ roslaunch pepper_dcm_bringup pepper_bringup.launch robot_ip:="10.42.0.76" network_interface:=wlan0

Joystick configurations

• Base control:
  

  • left arrow keys
    • <- turn left, -> right
    • ^ move forward, v move backward

• Head control:

  • right buttons

• Sleep & WakeUp:

  • R1 & R2 buttons

Run ORB SLAM 2 - Monocular

$ rosrun ORB_SLAM2 Mono /Path/To/ORB_SLAM2/Vocabulary/ORBvoc.bin /Path/To/ORB_SLAM2/Examples/Monocular/TUM1.yaml

RUN ORB SLAM - RGB-D

$ rosrun ORB_SLAM2 RGBD /Path/To/ORB_SLAM2/Vocabulary/ORBvoc.bin /Path/To/ORB_SLAM2/Examples/RGB-D/pepperCameraSettings.yaml false (create a new map) / true (use a saved map)

Run object recognition

$ rosrun pepper_recog recog.py

Special thanks

We want to thank

• Raul Mur-Artal for his awesome ORB SLAM2
• Bence Magyar for his advice of using Joystick teleop
• José María Sola Durán for his object recognition code framework

Logbook

link