Abstract: Quadrotor navigation in unknown environments is critical for practical missions such as search-and-rescue. Solving it requires addressing three key challenges: path-planning in non-convex ...

Abstract: Trajectory-sampling-based techniques, such as model predictive path integral control (MPPI), are well suited to solve complex optimization problems with numerous weighted cost terms and ...

Pure CUDA implementation for maximum performance Python bindings via nanobind JIT compilation support for custom dynamics Multiple controller variants Optimized CUDA kernels for parallel sampling This ...

This is an implementation of a Model Predictive Path Integral (MPPI) controller which uses IsaacGym as a dynamic model for the rollouts. We provide several examples of what tasks you can solve with it ...