Robot Simulation Environment is a robotics project focused on creating virtual environments where robotic systems can be designed, tested, and improved before deploying on real hardware.
The goal of this repository is to safely test robot movement, sensors, control algorithms, and AI models through simulation.
- Build virtual robot testing environments
- Reduce hardware testing risks
- Test algorithms before real deployment
- Create simulation tools for robotics research
Virtual representation of robots.
Features:
- Robot structure modeling
- Joint configuration
- Physical properties
- Movement constraints
Simulates real-world behavior.
Features:
- Gravity
- Collision detection
- Forces
- Motion dynamics
Simulated sensor environment.
Features:
- Camera simulation
- Distance sensors
- IMU simulation
- Encoder feedback
Test robot controllers safely.
Features:
- Motor command testing
- Motion validation
- PID tuning
- Error analysis
Train intelligent robots.
Features:
- Reinforcement Learning environments
- Reward systems
- Autonomous task learning
Robot Algorithm
|
Simulation Engine
|
Virtual Robot
|
Physics System
|
Performance Feedback
- Robotic Arm Simulation
- Autonomous Robots
- AI Robot Training
- Digital Twin Systems
- Reinforcement Learning
- Robot Model Loader
- Joint Simulation
- Movement Testing
- Collision System
- Sensor Simulation
- Environment Interaction
- RL Training Environment
- Autonomous Testing
- Digital Twin System
To create a safe virtual robotics laboratory where intelligent machines can learn before entering the real world.
A dependency-free, offline, CPU-only planar 2D robotic arm foundation has been implemented on main. It is not yet the later 3D or physics simulator. Python 3.10+ with Tkinter is required.
Windows: clone/download this repository and double-click launch_robot_simulator.bat.
From a terminal, from the repository folder:
python -m robot_sim
python -m robot_sim --model examples/six_joint.json
python -m unittest discover -s tests -v
python tests/gui_smoke.pyThe GUI supports joint sliders, live end-effector coordinates, pause/resume, step and reset. It accepts 2–6 planar revolute joints configured in JSON. The mathematics, simulation loop and robot model use only the Python standard library; no pip install, GPU or paid API is required.
Verification (2026-10-10): 43 unit tests passed, and both 2- and 6-joint GUI smoke tests passed using a real Tk window on a Linux virtual display. These are local tests; native Windows confirmation and user acceptance are still pending. See M1–M4 verification. Do not mark the milestone checkboxes complete until the agreed verification gates pass.
Start on Windows: double-click launch_advanced_simulator.bat in the downloaded repository. No pip packages, paid APIs or GPU needed.
Start from a terminal: python -m robot_sim --advanced
Six-joint example: python -m robot_sim --advanced --model examples/six_joint.json
Unit tests: python -m unittest discover -s tests -v
Real Tk GUI smoke: python tests/advanced_gui_smoke.py
The lab adds orbit/zoom/pan of a CPU-rendered 3D scene, XY target IK + trajectory planning, approximate gravity/torque physics, circle/box and self-collision stops, and bounded virtual PID motors. The control flow runs together through RobotLab, not isolated mock screens. The same GUI allows gravity and obstacle toggles. The arm's joint mechanics remain planar (XY); this version is not full spatial 3D kinematics or an industrial-grade coupled physics/contact simulator.
Evidence (2026-10-10): 100/100 offline unit tests passed; 2/2 real-window GUI smoke cases passed on Linux/Xvfb; the advanced GUI entrypoint launched. Source code on main was hash-verified against tested files. Windows and joint user/assistant acceptance have not happened, so no milestone has been fully ticked. See M5–M10 verification.
Windows: double-click launch_full_simulator.bat in a downloaded/cloned copy of this repository.
One-click local verification on Windows: double-click verify_on_windows.bat. This runs the suites locally and writes actual logs; it does not start hosted CI or a paid service.
# Integrated 3D visualization of planar arm + world + sensors + gripper:
python -m robot_sim --full
# Load a persisted 2D scene:
python -m robot_sim --full --scene examples/pick_and_place_scene.json
# Deterministic scripted pick-and-place evaluation (headless):
python -m robot_sim --evaluate --trials 5
# Offline unit/integration tests:
python -m unittest discover -s tests -v
# Machine-measured headless performance:
python scripts/benchmark.py --steps 2000
# Write actual command output and machine info to evidence/local/<timestamp>:
python scripts/verify.py --guiThe unified GUI shows 3D camera views of a planar robot arm, virtual scene objects, joint-encoder and distance-sensor readings, grip/release, path planning and an optional scripted demo. It works offline using the Python standard library and Tkinter.
Important status: M11–M15 source code and test scripts are committed to main, but new tests/5-trial evaluation/native Windows interaction have not been executed/verified here. M14's real partner-repository adapter is blocked until the other repos expose an actual callable interface; see Motion Adapter Contract. M16 tooling exists, but final acceptance is not complete. See M11–M16 status and required evidence.
The implementation plan, code tasks, real-run acceptance criteria, and required verification evidence are in Code Milestones. All milestones remain unverified until implemented, tested, run, and reviewed.