Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

What are Social Norms for Low-speed Autonomous Vehicle Navigation in Crowded Environments? An Online Survey

Mahsa Golchoubian, Moojan Ghafurian, Nasser Lashgarian Azad, Kerstin Dautenhahn

TL;DR

This study addresses how low-speed autonomous carts should navigate crowded, unstructured indoor spaces while maintaining safety and social acceptance. It uses an online survey (n=116) presenting nine interaction scenarios to elicit pedestrians' and passengers' perceptions of various cart behaviors, analyzed with linear mixed-effects models and thematic coding. The key contributions show a tendency for participants to favor road-like norms (e.g., passing on the right, overtaking on the left), path separation, and explicit signaling to communicate intent, along with a clear preference to stop for children and to use velocity-based avoidance. These findings yield practical design guidelines for socially compliant navigation algorithms in airports, shopping malls, and similar crowded environments, and highlight limitations related to sketch-based, remote testing and cultural scope.

Abstract

It has been suggested that autonomous vehicles can improve efficiency and safety of the transportation systems. While research in this area often focuses on autonomous vehicles which operate on roads, the deployment of low-speed, autonomous vehicles in unstructured, crowded environments has been studied less well and requires specific considerations regarding their interaction with pedestrians. For making the operation of these vehicles acceptable, their behaviour needs to be perceived as safe by both pedestrians and the passengers riding the vehicle. In this paper we conducted an online survey with 116 participants, to understand people's preferences with respect to an autonomous golf cart's behaviour in different interaction scenarios. We measured people's self-reported perceived safety towards different behaviour of the cart in a variety of scenarios. Results suggested that despite the unstructured nature of the environment, the cart was expected to follow common traffic rules when interacting with a group of pedestrians.

What are Social Norms for Low-speed Autonomous Vehicle Navigation in Crowded Environments? An Online Survey

TL;DR

This study addresses how low-speed autonomous carts should navigate crowded, unstructured indoor spaces while maintaining safety and social acceptance. It uses an online survey (n=116) presenting nine interaction scenarios to elicit pedestrians' and passengers' perceptions of various cart behaviors, analyzed with linear mixed-effects models and thematic coding. The key contributions show a tendency for participants to favor road-like norms (e.g., passing on the right, overtaking on the left), path separation, and explicit signaling to communicate intent, along with a clear preference to stop for children and to use velocity-based avoidance. These findings yield practical design guidelines for socially compliant navigation algorithms in airports, shopping malls, and similar crowded environments, and highlight limitations related to sketch-based, remote testing and cultural scope.

Abstract

It has been suggested that autonomous vehicles can improve efficiency and safety of the transportation systems. While research in this area often focuses on autonomous vehicles which operate on roads, the deployment of low-speed, autonomous vehicles in unstructured, crowded environments has been studied less well and requires specific considerations regarding their interaction with pedestrians. For making the operation of these vehicles acceptable, their behaviour needs to be perceived as safe by both pedestrians and the passengers riding the vehicle. In this paper we conducted an online survey with 116 participants, to understand people's preferences with respect to an autonomous golf cart's behaviour in different interaction scenarios. We measured people's self-reported perceived safety towards different behaviour of the cart in a variety of scenarios. Results suggested that despite the unstructured nature of the environment, the cart was expected to follow common traffic rules when interacting with a group of pedestrians.

Paper Structure

This paper contains 11 sections, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Research questions
  4. Method
  5. Procedure and Measures
  6. Participants
  7. Data Analysis Approach
  8. Results
  9. Discussion
  10. Conclusion
  11. Scenarios

Figures (2)

  • Figure 1: The results of the multiple choice question scenarios. (a) Pedestrian group's selected reactions about the cart approaching from behind in Scenario 3 (see Table \ref{['tab:scenarios']}). (b) Expectation of the passenger group in the same situation. "Ped" stands for pedestrian. (c) Participants’ preferred side for the cart’s operation in a corridor with an unidirectional flow of pedestrians as in scenario 4. (d) Participants’ expectations of the cart when it gets behind a bidirectional flow of pedestrians as in scenario 7.
  • Figure 2: The results of the pre-survey and post-survey questions