From time to time, lab members release software associated with our research. Some of these are concerted efforts by groups within the lab to develop and release open-source software proactively, and others are ‘one-offs’ – where a member of the lab happens to develop something that they think others may find useful and releases it. Some of these efforts are tied to one or more papers we’ve published (in which case, they are linked on the publications page as well), and in some cases, in collaborations with other labs. This is not an exhaustive list.
- Language-World is a language and robotics benchmark.
- QuadSwarm is a codebase for training reinforcement learning policies for quadrotor swarms.
- Crazyswarm allows you to fly a swarm of Bitcraze Crazyflie quadcopters in tight, synchronized formations.
- Tiny Differentiable Simulator is a header-only C++ and CUDA physics library for reinforcement learning and robotics with zero dependencies.
- Sample Factory is one of the fastest RL libraries focused on very efficient synchronous and asynchronous implementations of policy gradients (PPO).
- Garage is a toolkit for reproducible reinforcement learning research.
- GranularGym is designed for high-performance realtime simulation for robotic tasks with granular materials (hundreds of thousands of particles interacting (one-way) with rigid bodies of arbitrary geometry).
- Interested in 1khz Franka robot control in pure Julia? Try Franka.jl
- Matlab scripts for plotting data from the EcoMapper AUV (OceanServer Iver2)
- Gazebo, one of the most successful robot simulators, had its beginnings at USC.
- Stage, an early and prolific 2D robot simulator, was started at USC.
