newton-physics/newton
Jul 2025 – Sep 2025
3 Months active
Languages Used
C++Python
Technical Skills
C++C++ (via Python bindings)Control SystemsPhysics SimulationPyTorchPython
jvonmuralt
PROFILE
Jvonmuralt
Jonas von Muralt contributed to the newton-physics/newton repository by developing and optimizing robotics simulation features over a three-month period. He enhanced the Anymal walking policy through PD control optimization and MuJoCo solver tuning, improving control responsiveness and simulation realism using C++ and Python. Jonas integrated USD-based robot models and demonstration scripts, streamlining validation workflows and enabling efficient onboarding of new robots. He also standardized spatial vector representations and introduced solver stability improvements, addressing edge-case bugs and reducing simulation jitter. His work demonstrated depth in physics simulation, code refactoring, and testing, resulting in a more robust and maintainable simulation platform.
Overall Statistics
Feature vs Bugs
83%Features
Repository Contributions
9Total
Bugs
1
Commits
9
Features
5
Lines of code
1,498
Activity Months3
Your Network
1370 people
Work History
September 2025
2 Commits • 1 Features
Sep 1, 2025September 2025 delivered two key enhancements to the Newton physics engine, strengthening correctness, stability, and maintainability. Implemented Spatial Vector Representation Standardization to align how linear and angular components are stored and processed across spatial vectors (twists and wrenches), improving correctness and reducing edge-case bugs. Fixed jitter in the XPBD solver by introducing a relaxation parameter into contact delta calculations, replacing a hardcoded value to stabilize and smooth simulations. These changes deliver clearer traceability through commits and pave the way for faster future feature delivery and easier maintenance.
2 Commits • 1 Features
Sep 1, 2025September 2025 delivered two key enhancements to the Newton physics engine, strengthening correctness, stability, and maintainability. Implemented Spatial Vector Representation Standardization to align how linear and angular components are stored and processed across spatial vectors (twists and wrenches), improving correctness and reducing edge-case bugs. Fixed jitter in the XPBD solver by introducing a relaxation parameter into contact delta calculations, replacing a hardcoded value to stabilize and smooth simulations. These changes deliver clearer traceability through commits and pave the way for faster future feature delivery and easier maintenance.
September 2025
August 2025
3 Commits • 2 Features
Aug 1, 2025In August 2025, delivered enhancements to Newton that strengthen USD-based workflow validation and test reliability. Key outputs include USD-based simulation demonstration scripts with integrated Anymal D and H1 robot models, and a robust Anymal reset testing framework across multiple contact types. The initiatives improve model validation speed, cross-device testing (CPU/CUDA), and code maintainability, enabling faster onboarding of new robots and USD workflows into Newton.
August 2025
3 Commits • 2 Features
Aug 1, 2025In August 2025, delivered enhancements to Newton that strengthen USD-based workflow validation and test reliability. Key outputs include USD-based simulation demonstration scripts with integrated Anymal D and H1 robot models, and a robust Anymal reset testing framework across multiple contact types. The initiatives improve model validation speed, cross-device testing (CPU/CUDA), and code maintainability, enabling faster onboarding of new robots and USD workflows into Newton.
July 2025
4 Commits • 2 Features
Jul 1, 2025July 2025 (Month: 2025-07) focus: advancing Anymal walking policy in the Newton physics stack through PD control optimization, MuJoCo solver tuning, and a validation/demonstration suite. The work improved control responsiveness, realism of contact physics, and provided a robust pipeline for validation and demonstration, enabling faster iteration and clearer demonstration of capabilities to stakeholders.
4 Commits • 2 Features
Jul 1, 2025July 2025 (Month: 2025-07) focus: advancing Anymal walking policy in the Newton physics stack through PD control optimization, MuJoCo solver tuning, and a validation/demonstration suite. The work improved control responsiveness, realism of contact physics, and provided a robust pipeline for validation and demonstration, enabling faster iteration and clearer demonstration of capabilities to stakeholders.
July 2025
Activity
Loading activity data...
Quality Metrics
Correctness91.2%
Maintainability90.0%
Architecture82.2%
Performance82.2%
AI Usage20.0%
Skills & Technologies
Programming Languages
C++Python
Technical Skills
C++C++ (via Python bindings)Code RefactoringControl SystemsGame DevelopmentNumerical MethodsOptimizationPhysics SimulationPyTorchPythonPython ScriptingReinforcement LearningRoboticsRobotics SimulationSoftware Development
Repositories Contributed To
Overview of all repositories you've contributed to across your timeline