Driving Scenario Perception-Aware Computing System Design in Autonomous Vehicles

Abstract

Recently, autonomous driving development ignited competition among car makers and technical corporations. Low-level automation cars are already commercially available. But high automated vehicles where the vehicle drives by itself without human monitoring is still at infancy. Such autonomous vehicles (AVs) rely on the computing system in the car to to interpret the environment and make driving decisions. Therefore, computing system design is essential particularly in enhancing the attainment of driving safety. However, to our knowledge, no clear guideline exists so far regarding safetyaware AV computing system and architecture design. To understand the safety requirement of AV computing system, we performed a field study by running industrial Level-4 autonomous driving fleets in various locations, road conditions, and traffic patterns. The field study indicates that traditional computing system performance metrics, such as tail latency, average latency, maximum latency, and timeout, cannot fully satisfy the safety requirement for AV computing system design. To address this issue, we propose a "safety score" as a primary metric for measuring the level of safety in AV computing system design. Furthermore, we propose a perception latency model, which helps architects estimate the safety score of given architecture and system design without physically testing them in an AV. We demonstrate the use of our safety score and latency model, by developing and evaluating a safety-aware AV computing system computation hardware resource management scheme.

Zhuyi: Perception Processing Rate Estimation for Safety in Autonomous Vehicles

“…Moreover, the work may increase based on the scenarios. For example, a recent study showed that higher obstacle density around an AV can intermittently increase the computational demand [15]. As the compute demand created by real-time AV systems can be very high, it is important to quantify the perception requirement for safe operation, provision the fixed resources in the in-vehicle computer to prioritize important tasks, and use the leftover resources for tasks that will further improve safety and comfort.…”

Towards Computational Awareness in Autonomous Robots: An Empirical Study of Computational Kernels

“…However, this work lacks sufficient understanding on how the computation correlates with the driving environment. Zhao et al [29] perform a field study for autonomous vehicles (AV) to gain insights on how the computing systems should be designed. This work looks into the sequential operation of the AV computational kernels from a safety perspective and present a innovative safety criteria for AV.…”

“…This work looks into the sequential operation of the AV computational kernels from a safety perspective and present a innovative safety criteria for AV. Both of these works [27,29] are solely concentrated on AV which have different constraints than mobile robots such as an Unmanned Aerial Vehicle (UAV). Finally, Carlone and Karaman [30] consider the visual-inertial navigation problem, and propose anticipation with environmental cues to build a simplified model on robot dynamics and produce a computationally efficient system.…”

RoboRun: A Robot Runtime to Exploit Spatial Heterogeneity

“…Since navigation is a primitive task, this frees up resources for higher-level cognitive tasks. Related Work: System Design for Robotics: Prior works have explored individual kernel accelerators with statically set knobs [11], [16], made algorithmic improvements ignoring system implications [9], [14], developed runtime control targeting algorithmic specific knobs [15], [17] or individual kernels [7], [12], and demonstrated the importance of system-environment synergy without exploiting them [5], [8]. In this work, we instead take a holistic view of the end-to-end pipeline, study spatial heterogeneity systematically and design a middleware that adapts our compute subsystem online to this heterogeneity by targeting fundamental and inherent physical space characteristics.…”