The ARMI Laboratory is a research and innovation center in autonomous robots and machine intelligence at the Mechatronics Engineer Program (MEC), Cluster 3, Vietnamese German University (VGU), Vietnam.
ARMI Lab is dedicated to advancing autonomous robotics, digital twin technologies, and machine learning to solve complex, real-world industrial challenges. Our mission is to develop intelligent robotic systems that not only perceive and understand their environment, but also adapt, learn, and optimize their behavior autonomously in an industrial context.
Manipulation Research: Develop and optimize algorithms for Imitation Learning (ACT, Diffusion), advanced motion planning (RRT, trajectory optimization), inverse kinematics, and AI-driven perception including SLAM and multimodal sensor fusion.
Machine Learning & AI Models: Build and deploy machine learning models for predictive maintenance, anomaly detection, object recognition, pose estimation, and task automation. Utilize modern AI pipelines including deep learning, transformers, reinforcement learning, and foundation models tailored for robotics applications.
Simulation & Digital Twin: Configure high-fidelity Digital Twin environments to validate robotic behaviors, system stability, control robustness, and full mission workflows from simulation to deployment.
System Design: Architect robotics software systems using middleware like ROS/ROS2, integrated with simulation tools such as Gazebo, IsaacSim, and high-fidelity Digital Twin pipelines.
Mobile Robotics: Develop autonomous navigation stacks, including mapping, localization, obstacle avoidance, and path planning for AMRs, AGVs, and outdoor UGV platforms. Integrate sensor suites such as LiDAR, GPS, IMU, and stereo/depth cameras to achieve robust autonomy in dynamic industrial environments.
Integration: Interface specialized hardware with industrial robots UR and key components (LiDAR, depth cameras, actuators, embedded systems) via middleware such as ROS/ROS 2 and real-time communication bridges.
This paper presents a real-time machine health monitoring digital twin (DT) framework for robot machine-tending based on ISO 23247. The system integrates OPC UA and ROS for seamless communication between physical and cyber layers.
Designs a reactive optimization-based controller that avoids static and dynamic obstacles while driving the robot toward a target pose.
A ROS-based autonomous mobile robot for education and research with 2D SLAM, navigation, and outdoor capabilities.
A novel V-frame octocopter with FEM validation and Simulink PID tuning using Ziegler-Nichols method.
A novel saturation-function-based optimal controller for upper-extremity exoskeletons, validated in simulation.
14DOF humanoid upper body robot (6-DOF head + 2× 4-DOF arms), integrated with a depth camera. Designed for 3D printing or laser cutting — for research & education.
ROS-based service robot using 2D Lidar, embedded computer, and multiple sensors. Supports indoor SLAM and autonomous navigation for restaurants, hotels, and hospitals.
Award-winning charger docking and locking system for electric motorbikes — developed for the Bosch Green Challenge. Recognized for innovation in industrial design.
A six-legged, 23-DOF biomimetic robot with 3 DOF per leg. Uses servo motors and a lightweight frame — ideal for STEM education and locomotion research.
A remote-controlled 19-DOF humanoid built for mechanical engineering and robotics education. Designed to be easy to manufacture via 3D printing or laser cutting.
Core development in low-level and high-level languages for robotics and AI.
The ARMI V-Research Model is a modern adaptation of the classical V-Model, redesigned for autonomous robotics and machine intelligence workflows. It aligns verification (left side) with validation (right side), ensuring that every research activity—from requirements, system design, perception algorithms, ML model pipelines, to field testing—is tested against measurable criteria. This model guides ARMI Lab in developing safe, reliable, and high-performance robotic systems using digital twins, simulation-driven development, ML lifecycle management, and continuous sim-to-real improvement.
The model emphasizes four key principles:
This V-Research framework ensures that ARMI Lab produces scalable robotic technologies with strong scientific rigor, repeatability, and industrial-grade robustness — from early concept development to deployment in real autonomous robots.
