Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Testing predictive automated driving systems: lessons learned and future recommendations

Rubén Izquierdo Gonzalo, Carlota Salinas Maldonado, Javier Alonso Ruiz, Ignacio Parra Alonso, David Fernández Llorca, Miguel Á. Sotelo

TL;DR

This article presents and analyze the results of physical tests on the proving grounds of several predictive systems in automated driving functions developed within the framework of the BRidging Gaps for the Adoption of Automated VEhicles (BRAVE) project.

Abstract

Conventional vehicles are certified through classical approaches, where different physical certification tests are set up on test tracks to assess required safety levels. These approaches are well suited for vehicles with limited complexity and limited interactions with other entities as last-second resources. However, these approaches do not allow to evaluate safety with real behaviors for critical and edge cases, nor to evaluate the ability to anticipate them in the mid or long term. This is particularly relevant for automated and autonomous driving functions that make use of advanced predictive systems to anticipate future actions and motions to be considered in the path planning layer. In this paper, we present and analyze the results of physical tests on proving grounds of several predictive systems in automated driving functions developed within the framework of the BRAVE project. Based on our experience in testing predictive automated driving functions, we identify the main limitations of current physical testing approaches when dealing with predictive systems, analyze the main challenges ahead, and provide a set of practical actions and recommendations to consider in future physical testing procedures for automated and autonomous driving functions.

Testing predictive automated driving systems: lessons learned and future recommendations

TL;DR

This article presents and analyze the results of physical tests on the proving grounds of several predictive systems in automated driving functions developed within the framework of the BRidging Gaps for the Adoption of Automated VEhicles (BRAVE) project.

Abstract

Conventional vehicles are certified through classical approaches, where different physical certification tests are set up on test tracks to assess required safety levels. These approaches are well suited for vehicles with limited complexity and limited interactions with other entities as last-second resources. However, these approaches do not allow to evaluate safety with real behaviors for critical and edge cases, nor to evaluate the ability to anticipate them in the mid or long term. This is particularly relevant for automated and autonomous driving functions that make use of advanced predictive systems to anticipate future actions and motions to be considered in the path planning layer. In this paper, we present and analyze the results of physical tests on proving grounds of several predictive systems in automated driving functions developed within the framework of the BRAVE project. Based on our experience in testing predictive automated driving functions, we identify the main limitations of current physical testing approaches when dealing with predictive systems, analyze the main challenges ahead, and provide a set of practical actions and recommendations to consider in future physical testing procedures for automated and autonomous driving functions.
Paper Structure (23 sections, 17 figures, 5 tables)

This paper contains 23 sections, 17 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Overview of predictive perception systems
  3. Experimental validation of predictive automated systems
  4. Description of Experiments
  5. Assessment of Vehicle-VRU Interaction
  6. VRU-1 Configuration 3: pedestrian walks in parallel, turns head towards the car, and crosses the street
  7. VRU-2 Configuration 2: occluded pedestrian crossing at 50% impact (stop)
  8. VRU-3 Configuration 3 – Reduce, follow, and overtake
  9. Assessment of Vehicle-Vehicle Interaction
  10. VEH-1 Configuration 2 – Change lane
  11. VEH-2 Config. 2 – Turning at intersections
  12. Discussion
  13. Main findings
  14. Main limitations
  15. Limited fidelity
  16. ...and 8 more sections

Figures (17)

  • Figure 1: Types of scenarios, probability of occurrence in real-world traffic (log-scale), complexity, and risk. Long-tail distribution. We refer to GRVA2019 for more details on the scenarios.
  • Figure 2: Proving grounds at UTAC's premises.
  • Figure 3: Use case VRU-1 Conf. 3 - Walk, stop, and cross.
  • Figure 4: Detection of dummy face using computer vision.
  • Figure 5: Data logged in VRU-1 Conf. 3 at 40 km/h.
  • ...and 12 more figures