Back to Projects
RoboticsPublic

Autonomous Retail Intelligence Robot (ARIR)

Autonomous retail robot for inventory tracking, customer assistance, and ERP synchronization using ROS, YOLOv5, and DeepSORT.

PythonROS NoeticYOLOv5DeepSORTOpenCVGazeboOdoo ERPSLAMLinux
View repository

Detection

YOLOv5 + DeepSORT

Training

1000 epochs

Sync

live Odoo ERP updates

Project Overview

MSc robotics dissertation project integrating navigation, perception, and ERP data sync. The robot patrols shelves, tracks products in real time, and answers customer location requests.

Challenge

Retail environments require accurate product counting, continuous navigation, and reliable ERP updates under occlusion and motion.

Solution

Combined ROS navigation, YOLOv5 detection, DeepSORT tracking, and a debouncing strategy to prevent over-counting while synchronizing records to Odoo ERP.

Results

Validated autonomous navigation, robust shelf perception, and real-time inventory synchronization in Gazebo simulation with modular architecture suitable for extension.

Media Gallery

Retail simulation with real-time detection overlays
Retail simulation with real-time detection overlays
Customer location request and response workflow
Customer location request and response workflow
SLAM map and multi-camera visualization
SLAM map and multi-camera visualization
YOLOv5 model testing output
YOLOv5 model testing output
Training metrics and convergence curves
Training metrics and convergence curves
System architecture and data-flow methodology
System architecture and data-flow methodology
Autonomous Retail Intelligence Robot (ARIR) | Nasir Nasir-Ameen