GPU-Accelerated Trajectory Optimization for Robotics

This is a batched version of the trajectory optimization solver from the paper "MPCGPU: Real-Time Nonlinear Model Predictive Control through Preconditioned Conjugate Gradient on the GPU".

Installation

uv is used as a Python package and project manager.

git clone https://github.com/A2R-Lab/GATO.git
cd GATO
git submodule update --init --recursive

#uv
curl -LsSf https://astral.sh/uv/install.sh | sh
uv sync
source .venv/bin/activate

#docker
docker-compose up -d
docker-compose exec dev bash

make build #builds examples and bindings

Bindings only

make build-bindings

Requirements

GATO works with:

• Ubuntu 22.04
• CUDA v12.2
• C++17
• Python 3.10.12
• Docker 26.1.3

Usage

See batch_sqp.cu for an example of a batch solve in C++/CUDA, and batch_sqp.py for an example using Python bindings. Examples of MPC with GATO are in indy7-mpc/

Nomenclature

TODO

Related

• The open-source MPCGPU solver
• GRiD, a GPU-accelerated library for computing rigid body dynamics with analytical gradients