Search papers, labs, and topics across Lattice.
This paper presents CUDA MPC, a GPU-native framework for Model Predictive Control (MPC) that significantly enhances real-time optimization capabilities for systems with fast dynamics and high-dimensional models. By co-designing the optimization algorithm, execution model, and memory architecture specifically for CUDA hardware, CUDA MPC minimizes latency and maximizes efficiency through a parallel-in-horizon ADMM approach and a fused CUDA kernel. The framework demonstrates remarkable performance, achieving real-time rates with long horizons and outperforming existing solvers by up to 965 times in speed, even in complex scenarios like centralized multi-agent coordination.
CUDA MPC achieves real-time optimization for high-dimensional and fast-dynamic systems, outperforming traditional solvers by up to 965 times.
Model Predictive Control (MPC) delivers constraint-aware control, but its reliance on online optimization limits its use on systems with fast dynamics, high-dimensional models, or long horizons. Existing GPU implementations typically treat the device as a linear-algebra accelerator, leaving the optimization loop dependent on repeated kernel launches and high-latency memory transfers. This paper introduces CUDA MPC, a GPU-native MPC framework that co-designs the optimization algorithm, execution model, and memory architecture for CUDA hardware. CUDA MPC pairs a parallel-in-horizon alternating direction method of multipliers (ADMM) splitting with a fused CUDA kernel that runs the entire iterative solve on the device. Intermediate optimization variables stay in low-latency, on-chip shared memory, and a localized atomic-flag protocol synchronizes only adjacent horizon blocks, minimizing host intervention, kernel-dispatch overhead, and global-memory traffic. Across six nonlinear robotics benchmarks spanning increasing state dimension and constraint density, CUDA MPC sustains real-time rates at horizons one to two orders of magnitude longer than CPU solvers: it solves an optimization-based collision-avoidance parking problem with 100 s of lookahead within a 0.1 s sampling interval, and is the only solver evaluated that achieves both real-time execution and collision-free coordination for a centralized 10-agent swarm, where acados and CasADi return no feasible solution and require 3.5 s and 4.5 s per solve. Against tensor-framework implementations of the same ADMM splitting, the fused kernel is up to $965\times$ faster.
