FGLQR: Factor Graph Accelerator of LQR Control for Autonomous Machines

TL;DR

FGLQR is a specialized hardware accelerator that significantly speeds up and reduces energy consumption of LQR control computations in autonomous machines by leveraging factor graph abstractions.

Contribution

The paper introduces FGLQR, a novel hardware accelerator that transforms LQR control into a factor graph problem for efficient real-time processing in autonomous systems.

Findings

Achieves 10.2x speedup over CPU implementation.

Reduces energy consumption by 32.9x.

Maintains accuracy with the factor graph approach.

Abstract

Factor graph represents the factorization of a probability distribution function and serves as an effective abstraction in various autonomous machine computing tasks. Control is one of the core applications in autonomous machine computing stacks. Among all control algorithms, Linear Quadratic Regulator (LQR) offers one of the best trade-offs between efficiency and accuracy. However, due to the inherent iterative process and extensive computation, it is a challenging task for the autonomous systems with real-time limits and energy constrained. In this paper, we present FGLQR, an accelerator of LQR control for autonomous machines using the abstraction of a factor graph. By transforming the dynamic equation constraints into least squares constraints, the factor graph solving process is more hardware friendly and accelerated with almost no loss in accuracy. With a domain specific parallel solving pattern, FGLQR achieves 10.2x speed up and 32.9x energy reduction compared to the software implementation on an advanced Intel CPU.