Genesis was created by a consortium of 19 universities and companies, including renowned institutions such as MIT, Imperial College London, Peking University, and technology companies such as Nvidia and the MIT-IBM Watson AI Lab. The simulator allows robots to train in virtual environments and achieve learning speeds that would be unattainable in the real world. According to the developers, Genesis robots learn up to 430,000 times faster than conventional methods.
This enormous acceleration is made possible by several innovative functions. One of them is the ability to create 3D environments using AI and text prompts. Researchers and developers can thus quickly and easily generate complex scenarios, such as factory halls, apartments or urban environments, by giving the system textual instructions. Genesis' underlying physics engine was developed from the ground up and is capable of simulating a wide range of materials and physical phenomena, from the collision of objects to the interaction of liquids and gases.
Another advantage of Genesis is the high speed it can achieve without the need for supercomputers. On a standard computer Standard Nvidia GeForce RTX 4090 graphics card up to 43 million frames per second (FPS) are possible in some simulations. This is possible thanks to the combination of efficient code and the simulator's cross-platform compatibility. The physics simulator works with Nvidia and AMD graphics cards, conventional CPUs and Apple's Metal, making it accessible to a wide range of users. One hour of computing in the simulator is equivalent to ten years of real-world training, which can significantly shorten robot development cycles.
The project is open source and has been released under the Apache License, version 2.0. This gives developers a lot of freedom in using and modifying the code. Source code and other information are available at on GitHub.
Genesis has a wide range of applications. In industry, the simulator could be used to optimize production lines and robot movements, including for tasks such as welding, painting or assembly. As part of the development of domestic robots, Genesis makes it possible to test navigation algorithms in virtual homes and to simulate interactions with objects and people. In research and development, Genesis can be used to study new robot designs, new control methods, and to conduct experiments in safe virtual environments. Genesis also offers potential in the field of disaster response, for example to train robots for use in dangerous environments, such as searching for buried victims.
The impressive results of the simulations can be viewed at on the project page. Videos there show a four-legged robot that, with the help of Genesis, learned to crawl, run at a gallop, jump on two legs and even do a backflip. The neural network that controls the robot was likely trained in just a few hours. The RoboGen tool, which runs in Genesis, was used for this purpose.
Genesis may also be of interest to the film industry and game developers. The simulator can easily create animations of liquids and gases that would otherwise be difficult to program. Genesis can also simulate complex motion sequences in character design, making character animations more realistic.
