Approximation-Aware Design of an Image-Based Control System

Abstract

Image-based control (IBC) systems are common in many modern applications. In such systems, image-based sensing imposes massive compute workload, making them challenging to implement on embedded platforms. Approximate image processing is a way to handle this challenge. In essence, approximation reduces the workload at the cost of additional sensor noise. In this work, we propose an approximation-aware design approach for optimizing the energy, memory and performance of an IBC system, making it suitable for embedded implementation. First, we perform compute-and data-centric approximations and evaluate its impact on the energy efficiency, memory utilization and closed-loop qualityof-control (QoC) of the IBC system. We observe that the workload reductions due to approximations allow mapping these lighter approximated IBC tasks to embedded platforms with lower power consumption while still ensuring proper system functionality. Therefore, we explore the interplay between approximations and platform mappings to improve the energy-efficiency of IBC systems. Further, an IBC system operates under several environmental scenarios e.g., weather conditions. We evaluate the sensitivity of the IBC system to our approximation-aware design approach when operated under different scenarios and perform a failure probability (FP) analysis using Monte-Carlo simulations to analyze the robustness of the approximate system. Finally, we design an optimal approximation-aware controller that models the approximation error as sensor noise and show QoC improvements. We demonstrate the effectiveness of our approach using a concrete case-study of a lane keeping assist system (LKAS) using a heterogeneous NVIDIA AGX Xavier embedded platform in a hardware-in-the-loop (HiL) framework. We show energy and memory reduction of up to 92% and 88% respectively, for 44% QoC improvements with respect to the accurate implementation. We show that our approximation-aware design approach has an FP (per km) ≤ 9.6 × 10 −6 %.

An Evaluation Framework for Vision-in-the-Loop Motion Control Systems

“…Recent works are reported on evaluation framework of vision-based control systems. Authors in [8], [9] proposed the IMACS framework for evaluating the performance of image approximation in a closed-loop image-based control system for automotive control applications. The framework helps to evaluate the impact of image approximation in the imagebased control system.…”

Optimising Multiprocessor Image-Based Control Through Pipelining and Parallelism

“…20. This graph structure is for the approximate setting 'S3' of the IBC system implemented in [56], [57]. The execution time numbers for the actors in the graph and the rates in the channels are obtained by mapping and profiling the IBC application on the NVIDIA AGX Xavier platform.…”

Hardware- and Situation-Aware Sensing for Robust Closed-Loop Control Systems

“…Fig. 2 shows an overview of our setup adapted from [9], [11]. Below we provide an overview of LKAS from an algorithmic as well as hardware perspective.…”

“…On the other hand, traditional lane detection algorithms proposed in [4]- [7] detect lane segments based on diverse handcrafted cues like color-based features, structure tensor, bar filter and so on. Similarly, recent research [8], [9] has shown that approximate sensing can reduce high compute workloads at the cost of additional sensor noise which is tackled by robust control design. This class of algorithms achieve high FPS on edge devices, but they suffer from robustness issues due to situation variations raising safety concerns when used in closedloop control systems.…”

“…This class of algorithms achieve high FPS on edge devices, but they suffer from robustness issues due to situation variations raising safety concerns when used in closedloop control systems. Contribution: We make the hardware-efficient traditional approaches [8], [9] robust to situations by introducing light-weight CNN classifiers and use this situational knowledge for runtime dynamic tuning of the required feature values. This translates to hardware-efficiency and situation-robustness, leading to failsafe operation with a higher quality of control (QoC) in closedloop systems.…”