Secur-a-Bot MOVO Service Robot Stack

Watch the full demo showing the complete robot stack in action:

https://github.com/Kushion32/security_robot/raw/master/media/fulldemo.mov

Please find the rest of the demonstration videos and photos sprinkled throughout the features sections below.

The project implemented below includes a full ROS-based software stack for a MOVO service robot, featuring:

• Navigation to waypoints with UI integration
• Person-following safety behavior using leg detection
• Automatic head/screen flipping interactions
• Real-time person counting using YOLOv8 on RealSense camera feed
• Face authentication for unauthorized user detection
• Easy map management for SLAM and room labeling
• Operator UI (movo-nav-ui) and attendee UI (secure-a-bot-ui/robot-ui)

Overview

This repository combines ROS navigation, interaction behaviors, and two web UIs:

• movo-nav-ui: operator-facing dashboard for navigation and map tools
• secure-a-bot-ui/robot-ui: attendee-facing touchscreen UI for selecting destinations
• src/kinova-movo/movo_nav: waypoint backend (goto_points.py) and nav integration
• src/kinova-movo/movo_people: people detection, person-following monitor, head/screen flip nodes

Features

1. Waypoint Navigation

• Send robot to named waypoints from UI or terminal
• Return robot to home after user confirmation
• Publishes UI-ready status and arrival topics:
  • /ui_nav_ready
  • /ui_waypoints_list
  • /ui_robot_arrived

2. Person-Following Safety

• Uses leg detections and monitor logic to ensure user remains with robot
• Robot pauses if person is lost and can resume when person returns
• Key topics:
  • /leg_tracker_measurements
  • /movo/person_following_status

3. Head/Screen Interaction

• Auto mode: detect the side a person is standing on from depth, rotate head/tablet side accordingly
• Supports return-to-front behavior when robot navigation trip is complete and user sends robot back
• Interfaces:
  • Service mode: /movo/flip_to_back_screen, /movo/flip_to_front_screen
  • Auto mode topics: /movo/auto_screen_flip/command, /movo/auto_screen_flip/state

4. Person Counting

• Uses YOLOv8 object detection on RealSense color camera stream
• Continuously counts number of people in the camera's field of view
• Publishes count to /people/count (std_msgs/Int32)

5. Face Authentication (Unauthorized User Detection)

• Validates faces against authorized profiles using face_recognition
• Identifies whether a detected person is an authorized user or unauthorized
• Publishes status to /face_auth/status (std_msgs/String)
• Publishes annotated camera feed with bounding boxes to /face_auth/overlay
• Authorized face encodings are loaded from a local directory (e.g., config/known_faces/) and should be pre-populated with images of authorized users (e.g., me.jpg)

6. Map Manager (SLAM & Room Labeling)

• Handles dynamic SLAM map saving and loading during runtime
• Manages room labels and mapping statuses dynamically
• Interface: JSON command inputs on /map_manager/command
• Publishes JSON status on /map_manager/status

7. Operator UI (MOVO UI)

• ROS-connected operator dashboard
• Waypoint/nav control and map utilities
• Path: movo-nav-ui

8. Attendee UI (Secure-a-Bot UI)

• Touch-friendly destination selection for attendees
• Hardcoded destination buttons mapped to waypoint names (e.g. washroom, elevator, stairs)
• Path: secure-a-bot-ui/robot-ui

Repository Layout

catkin_ws/
	src/kinova-movo/
		movo_nav/
		movo_people/
		movo_demos/
		...
	movo-nav-ui/
	secure-a-bot-ui/robot-ui/
	COMMANDS.md

Prerequisites

• Ubuntu + ROS Noetic
• Catkin workspace configured at ~/catkin_ws
• MOVO robot network access (movo1/movo2 reachable)
• Node.js and npm for both UIs

Robot Bring-Up

1. Power on MOVO main system (movo1).
2. Power on secondary computer (movo2) from the back panel.
3. Verify connectivity:

ping movo1
ping movo2

4. Source ROS workspace:

cd ~/catkin_ws
source devel/setup.bash

5. Follow the rest of the setup instructions in 'src/kinova-movo/README.md' for extra networking setup

Recommended ROS Environment

Set ROS networking in each terminal (update values for your network): The below works when connected to 'movo' networking projected by the robot

export ROS_MASTER_URI=http://10.66.171.2:11311
export ROS_IP=10.66.171.254
export ROS_HOSTNAME=10.66.171.254
source ~/catkin_ws/devel/setup.bash

Notes:

• Use the interface/IP reachable by the robot network.
• If running over wired tether, prefer the robot subnet IP (for example 10.66.x.x) instead of hotspot-only IPs.

Build

cd ~/catkin_ws
catkin build
source devel/setup.bash

Core Runtime Commands

Once you've gone through the launch commands on each of movo1 and movo2 as detailed in 'src/kinova-movo/README.md', you can run the following core commands for each feature:

Navigation + Map

Run navigation stack:

roslaunch movo_demos map_nav.launch sim:=false local:=true map_file:=movo_map

Open RViz navigation view:

roslaunch movo_viz view_robot.launch function:=map_nav

Save map:

rosrun map_server map_saver -f ~/movo_map

Waypoint Backend

Start waypoint navigator (UI mode):

rosrun movo_nav goto_points.py _interactive:=false _ui_mode:=true

Other modes:

# No UI confirmation required
rosrun movo_nav goto_points.py _interactive:=false _ui_mode:=false

# Interactive terminal mode
rosrun movo_nav goto_points.py

# Custom auto-return delay
rosrun movo_nav goto_points.py _ui_mode:=false _auto_return_delay:=10.0

Person-Following + Leg Detection

Start person following system, can be used when navigating with waypoints:

roslaunch movo_people person_following_navigation.launch

Verify detections/status:

rostopic echo /leg_tracker_measurements
rostopic echo /movo/person_following_status

Head/Screen Flip

Manual service-based node:

rosrun movo_people screen_flip_node.py
rosservice call /movo/flip_to_back_screen
rosservice call /movo/flip_to_front_screen

Automatic depth-based node:

roslaunch movo_people auto_screen_flip.launch trigger_distance_m:=0.40 clear_distance_m:=0.60

Monitor/send commands:

rostopic echo /movo/auto_screen_flip/state
rostopic pub -1 /movo/auto_screen_flip/command std_msgs/String "data: 'home'"

Person Counting

Run the YOLOv8 person counting node (connects to /camera/color/image_raw):

rosrun movo_people realsense_person_counter.py

Monitor count:

rostopic echo /people/count

Face Authentication (Unauthorized User Detection)

Run the face authentication node (requires pre-configured known face image me.jpg in config):

roslaunch movo_nav face_authentication.launch

Monitor status and view video feed:

rostopic echo /face_auth/status
rosrun image_view image_view image:=/face_auth/overlay

Map Manager (SLAM & Saved Maps)

Start the runtime mapping session and label manager Node:

rosrun movo_nav map_manager.py

Send a command manually to list saved maps or start mapping:

rostopic pub -1 /map_manager/command std_msgs/String "data: '{\"action\": \"start_mapping\"}'"
rostopic pub -1 /map_manager/command std_msgs/String "data: '{\"action\": \"list_maps\"}'"

Operator UI Setup (movo-nav-ui)

cd ~/catkin_ws/movo-nav-ui
npm install
npm start

Default local URL: http://localhost:3000

Attendee UI Setup (secure-a-bot-ui/robot-ui)

cd ~/catkin_ws/secure-a-bot-ui/robot-ui
npm install
npm run dev

If needed, set ROSBridge URL in UI settings or via env before launch:

export VITE_ROSBRIDGE_URL=ws://<rosbridge_host>:9090

ROS Topics and Commands Used by UIs

Subscribed by UI

• /ui_nav_ready (std_msgs/Bool)
• /ui_waypoints_list (std_msgs/String, JSON array)
• /ui_robot_arrived (std_msgs/String)

Published by UI

• /ui_navigation_command (std_msgs/String)
  • go:<waypoint_name>
  • confirm:<waypoint_name>

Example:

rostopic pub -1 /ui_navigation_command std_msgs/String "data: 'go:washroom'"
rostopic pub -1 /ui_navigation_command std_msgs/String "data: 'confirm:washroom'"

Demo Media

Add your recorded media here.

Navigation Demo

• Notes: Navigation testing

Person Following + Leg Detection Demo

• Notes: Shows the safety behavior using leg detection

Head Turn / Screen Flip Demo

https://github.com/Kushion32/security_robot/raw/master/media/headmoving.MOV

Person Counting Demo

• Notes: Evaluates YOLOv8 counts

Face Authentication (Unauthorized User Detection) Demo

https://github.com/Kushion32/security_robot/raw/master/media/facedetection.mp4

Map Manager (Live SLAM & Room Labels) Demo

• Notes: Live SLAM demonstrations

Operator UI Demo

Attendee UI Demo

Troubleshooting

UI connects but robot does not move

• Check goto_points.py is running.
• Check nav readiness:

rostopic echo /ui_nav_ready

• Confirm action server exists:

rostopic list | grep move_base

Waypoint unknown errors

• Confirm waypoint file contains the exact key name:

cat ~/catkin_ws/src/kinova-movo/movo_nav/movo_waypoints.yaml

• Confirm UI command:

rostopic echo /ui_navigation_command

Robot starts at wrong map location (orange AMCL marker off-map)

• In RViz, use 2D Pose Estimate to set initial pose.
• Rotate robot slowly in place to converge AMCL.

Head does not return to front after attendee confirmation

• Ensure either of these is active:
  • screen_flip_node.py (service path)
  • auto_screen_flip_node.py (topic command path)

Acknowledgements

• Kinova MOVO ROS packages and navigation stack
• ROS Noetic ecosystem packages (move_base, amcl, map_server, etc.)