IsaacSim-Nav: High-Fidelity Digital Twin Platform for Autonomous Navigation
This digital TestBed is an advanced simulation environment built on NVIDIA Isaac Sim (Omniverse), specifically tailored for the development and benchmarking of autonomous mobile robots (AMRs). It provides a high-fidelity “Digital Twin” infrastructure where users can test navigation stacks (such as ROS 2 Nav2) in photorealistic environments with accurate physics. The platform enables complex scenario testing, including multi-robot simulation infrastructure for coordination research, dynamic obstacle avoidance, and sensor fusion (LiDAR, Depth Cameras, IMU) without the risk of hardware damage. Its key differentiator is the high-fidelity sensor simulation and the ability to easily bridge the “Sim-to-Real” gap. It serves as a primary tool for researchers and students to validate algorithms in varied industrial and office-like 3D scenes before physical deployment.
| Name of Principal Investigator: | Kosta Jovanović |
| Position / institutional role: | Associate Professor |
| Info Email: | kostaj@etf.bg.ac.rs |
| ORCID persistent identifier (PID): | 0000-0002-9029-4465 |
| Name of Host Organization | University of Belgrade – School of Electrical Engineering (ETF) |
| Department or Lab | Department of Signals and Systems, ETF Robotics lab |
| Name of Building | Palace of Science |
| Physical Address | Kralja Milana 11, 11000 Beograd, Serbia |
| Website Links | https://robot.etf.bg.ac.rs |
| Institutional contact name | Prof. Kosta Jovanović |
| Institutional contact email | kostaj@etf.rs |
| Non-academic stakeholders |
| Industrial Partners; Startups; SMEs |
| Academic stakeholders |
| Undergraduate students; PhD students; MSc students; Researchers |
| Other types of stakeholders |
| Application case: | Short description: |
| Warehouse AMR Navigation | Testing an Autonomous Mobile Robot (AMR) in a cluttered warehouse scene to evaluate path planning robustness. |
| Sim-to-Real Validation | Comparing the behavior of a digital twin (e.g., TurtleBot or Nova Carter) against its physical counterpart at ETF Lab. |
| Socially-Aware Navigation | Simulating robot movement in corridors with dynamic human-like characters (NPCs) to test avoidance logic. |
list of hardware components with their brief descriptions:
– GPU-accelerated workstation for real-time rendering
– NVIDIA RTX 3080/4090 or A-series GPU (min. 8GB VRAM)
– Multi-core CPU (Intel i7/Xeon or AMD Ryzen 7)
– Minimum 32GB System RAM
list of software components with their brief descriptions:
– NVIDIA Isaac Sim (Omniverse platform)
– ROS 2 Humble/Foxy Middleware
– Python 3.10+ for scripting and AI logic