A Student Scientific Research (NCKH) Project from VNU - University of Engineering and Technology (VNU-UET), Academic Year 2025-2026.
This repository contains the source code, hardware schematics, and simulation models for the research project: "Research, development, and simulation of a multi-robot system for optimizing transport throughput in warehouses."
As warehouse automation scales, deploying a massive fleet of autonomous mobile robots (AMRs) in a confined grid space introduces severe challenges regarding traffic coordination, collision avoidance, and deadlocks. While many solutions exist merely in software simulations due to high hardware deployment costs, this project builds a comprehensive, end-to-end multi-robot coordination system bridging theoretical simulation and a functional physical model.
- Advanced Path Planning & Task Allocation: Implements a hybrid algorithm combining Cooperative A (CA)** and an improved Particle Swarm Optimization (PSO). This significantly outperforms standard A* and Hungarian algorithms, ensuring deadlock-free routing and optimal load balancing.
- Custom Hardware Fleet: Includes the design and fabrication of an automated grid-map and 5 custom-built AMRs utilizing ESP32 microcontrollers. The robots utilize analog line-tracking (TCRT5000), absolute RFID positioning at intersections (RC522), and encoder-based PID control for precise movement.
- Low-Latency Multi-Agent Communication: Employs a hierarchical network architecture. A central server communicates via UART to an ESP32 Gateway, which then multi-casts packet-filtered instructions to the robot fleet using the ESP-NOW protocol, bypassing traditional Wi-Fi router limits and ensuring real-time (<1ms) responsiveness.
- Dual-Purpose Software Suite: Features a centralized Python (Tkinter) application that serves both as a virtual benchmarking simulator (to determine the saturation threshold of the warehouse) and an operation control system for the physical bots.
📦 nckh2026-uet-warehouse-robots
┣ 📂 Hardware/
┃ ┣ 📂 Gateway/ # ESP32 firmware for the communication bridge
┃ ┣ 📂 agents/ # ESP32 firmware for the 5 autonomous robots (NCKH1-5)
┃ ┗ 📂 server/ # Python server & UI for real-time hardware control
┣ 📂 Simulation/ # Python visual simulator and algorithm benchmarking tool
┣ 📂 pics/ # Project images, PCB designs, and charts
┗ 📂 video/ # Demonstration videos of the physical model operation
Through extensive rigorous benchmarking, the proposed CA* + improved PSO hybrid algorithm proved exceptional at maintaining continuous flow under high-density scenarios. Furthermore, the mathematical regression analysis established the physical saturation point of our specific experimental map at 20 robots/grid. Exceeding this threshold results in negative marginal throughput due to exponential increases in wait times and traffic congestion. The successful operation of the physical model validated the feasibility of mapping simulated algorithms to real-world edge devices.
Students:
- Phạm Việt Đan - K68-M-MT2
- Trần Minh Toàn - K68-M-MT1
- Trần Hùng Trường - K68-M-MT1
- Dương Văn Hải Đăng - K69-M-AT1
- Đặng Mạnh Toàn - K69-M-AT2
Instructor: TS. Phạm Duy Hưng ( VNU-UET)