Autonomous Warehouse Management System

Team CON-SOL-E - Flipkart Grid 3.0

Project Description

This project focused on designing an autonomous warehouse management system leveraging Autonomous Mobile Robots (AMRs) to streamline parcel transportation. The system seamlessly moved parcels from two Induct Stations to nine destination chutes with minimal human intervention.

Key features of the system included:

Deployment of three AMRs to autonomously transport parcels.
Integration of a central camera positioned above the arena for system-wide oversight and coordination.
Two cameras mounted at the Induct Stations for scanning parcel QR codes, ensuring accurate tracking and destination assignment.
Implementation of advanced computer vision algorithms to enable:
- Dynamic calculation of the shortest path for each AMR to its assigned destination chute.
- Real-time collision avoidance, maintaining efficiency and safety within the operational environment.

This project demonstrated a robust approach to automating warehouse operations, highlighting the potential of AMRs and computer vision in modern logistics.

TEAM_CONSOLE_FLIPKART_GRID_3_ROUND_2_LIVE_VIDEO.mp4

Warehouse Management System

Parcels are efficiently transported from the two loading stations to their designated chutes with precision and collision-free operations, ensuring seamless and reliable delivery.

TEAM_CONSOLE_FIPKART_GRID_3_ROUND_2_IP_VIDEO.mp4

A live video feed from the Induct Station captures parcels being loaded onto the AMRs, showcasing the system's operation in real time. The feed also provides a comprehensive view of the arena, enabling observers to monitor the preparation and transport of parcels seamlessly.

Project Insights

Arena Description

Grid-Based Arena for Autonomous Parcel Delivery

The system operated within a grid-based arena featuring nine destination chutes and two Induct Stations, all under the watchful eye of a central camera providing real-time oversight.

Key innovations included:

Autonomous Mobile Vehicles (AMVs) equipped with ARUCO markers for precise localization and mapping within the arena.
Seamless, autonomous transportation of parcels from the Induct Stations to the designated destination chutes, eliminating the need for operator intervention.
Each parcel was labeled with a QR code, scanned upon placement on the AMV, ensuring accurate tracking and delivery throughout the process.

This setup showcased the potential of AMVs for efficient and autonomous parcel handling in a controlled, monitored environment.

Arena Specifications

Overall Dimensions: The arena measures 7 x 7 ft, excluding the Induct Points.
Grid Layout: Each grid block is 6 x 6 in, marked with 1 cm tape for precise navigation.
Induct Points: Represented as purple boxes, serving as the starting points for parcel transport.
Destination Chutes: Depicted as yellow boxes, each surrounded by a 1-inch-high wall to contain parcels securely.
Boundary Walls: The entire arena, including the grid and Induct Points, is enclosed by a 1-inch-high wall for structure and containment.

This well-structured arena design ensures efficient operation and clear segmentation for system components.

Robot Specifications

Dimensions: Each robot measures 6 x 6 inches, optimized for navigation within the grid-based arena.
Payload Capacity: Equipped with a tray on top designed to carry parcels approximately 20 x 20 x 20 mm in size.
Item Delivery Mechanism: The tray features a flipping mechanism to deposit parcels into the destination chutes efficiently.
Sensor-Free Design: Robots rely entirely on external systems for navigation and object detection, with no sensors mounted directly on the robot.

This design prioritizes simplicity and external coordination for seamless operation within the system.

Algorithm

The system's operation is structured around three key processes:

Parcel Scanning and Destination Allocation
- Parcels are scanned to extract their QR code information, enabling accurate identification of their designated destination chute.
Drop Point Localization
- The system identifies the precise drop points near the destination chutes to ensure efficient and accurate parcel delivery.
Robot Orientation Assessment
- This process evaluates the robot's orientation, ensuring it correctly aligns and parks back at the Induct Station after completing the delivery task.

Overhead Camera: A single camera is designated to monitor and provide feedback for the navigation of robots within the arena.
Induct Station Cameras: Two cameras are positioned at the Induct Stations to efficiently scan parcel QR codes for destination allocation.
Central Unit: All three cameras are connected to a PC, which acts as the Central Unit of the system, coordinating operations and processing data.
Communication Infrastructure: A local server facilitates communication via MQTT, ensuring seamless data transfer and integration across the system components.

Process Flow

Real-Time Parcel Delivery Dashboard

The system features a webpage interface that provides real-time updates on parcel delivery operations, ensuring transparency and efficient monitoring.

Achievements

Page updated

Google Sites

Report abuse