.png){kind=icon}
.png){kind=icon}
Return
Enabling Learning-Based Efficiency Optimizer With Shadow Cycles in Resource-Constrained Autonomous Embedded Systems
.png){kind=icon}
PRE
AI
DOI：10.1109/TC.2025.3644184.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The emerging trend of autonomous embedded systems (AES) is promising to minimize human intervention in critical tasks. In the pursuit of maximal per-watt performance, the complex hardware and software of AES require intelligent energy efficiency optimizers (EO), and the stochastic runtime variances require continuous EO. However, deploying the desirable on-device EO causes severe performance slowdown due to contention on limited computing power with the AES pipeline. We find that there are ignored and underutilized heterogeneous resources within AES for costly EO, which results from unbalanced accelerator behaviors and misaligned parallel inference executions. We experimentally and theoretically analyze the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Shadow Cycles</i> within the realistic autonomous Bird’s Eye View pipeline on commercial embedded platforms, categorizing them into vertical and horizontal types with distinct properties. In this paper, we introduce <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SHEEO+</i>, a continuous and intelligent energy efficiency optimizer that utilizes ignored heterogeneous shadow cycles. It achieves continuous and lightweight AES monitoring with the observation module, as well as intelligent and efficient AES power management with the optimization module. On the one hand, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SHEEO+</i> observes both the internal runtime status and external environment variance with portable interfaces to capture shadow cycles and real-time states. On the other hand, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SHEEO+</i> optimizes power configurations per iteration based on deep reinforcement learning (DRL) methods. It tailors DRL for two types of shadow cycles and invocates optimization processes based on resource availability. To extensively evaluate <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SHEEO+</i>, we implement a prototype and deploy it on realistic edge platforms. The evaluation results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SHEEO+</i> utilizes up to 74.2% shadow cycles and achieves up to 18.6% energy efficiency improvements compared to state-of-the-art energy efficiency optimizers with negligible deployment overheads.
Keywords:
Energy efficiency optimization
shadow cycle
reinforcement learning
autonomous embedded system
Journal
IF:
3.8
Papers:
5.3K
Citations:
9.8K