This is the editors' words from Haoran and Yuyang. They want to explain their ideas of creating this website. If you find what you want here, or just simply like it, please consider giving a star to its repo on GitHub!
At the intersection of robotics and computer vision, physics simulators serve as crucial tools for validating algorithms and testing models before they're deployed in real-world robots. They cater to a wide range of functions, from rendering photorealistic perceptions with ground truth segmentation masks, to spawning parallel environments with GPU for efficient data sampling in Reinforcement Learning. Consequently, an array of physics simulators has emerged, each with its unique set of offerings. However, several challenges persist:
- Some of the required functions are found missing on the half way of development with an initially chosen simualtor.
- Simulation results are inconsistent across different simulators, adding another layer of complexity.
- Some simulators suffer from poor documentation, leaving users in the dark about how to implement the desired functions using the provided APIs.
- Troubleshooting can be a nightmare, with solutions either hard to come by, incomplete or entirely absent from online resources.
- The list goes on and on and on and on...
Don't stop believing
Addressing these hurdles for the benefit of the wider community, particularly researchers who could better spend their time on more meaningful pursuits, we've created Simulately. Our mission extends beyond creating a repository for our experiences in research and development. We aim to foster a collaborative platform where learnings and insights are shared, and where others are encouraged to do the same. Welcome to Simulately, your comprehensive guide to navigating the world of robotics simulators.