A condensed-space interior-point method factorized on GPU/SIMD hardware solves constrained LQ-MPC problems an order of magnitude faster than CPU when the number of inputs is small and the horizon moderate.
Model predictive control for autonomous navigation using embedded graphics processing unit
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2022 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Exploiting GPU/SIMD Architectures for Solving Linear-Quadratic MPC Problems
A condensed-space interior-point method factorized on GPU/SIMD hardware solves constrained LQ-MPC problems an order of magnitude faster than CPU when the number of inputs is small and the horizon moderate.