Multi-Robot Coordination in Dynamic Environments Shared with Humans

Alcherio Martinoli, Zeynab Talebpour
2018
Article de conférence

Résumé

This work addresses multi-robot coordination in social human-populated environments based on a Hoplites-based framework for solving the multi-robot task allocation problem. Humans are considered in the proposed coordination mechanism by means of accounting for social costs in bid evaluations and requesting collaboration in socially blocking situations. Initially, the effect of a realistic dynamic noisy environment with varying number of static/moving humans on the behavior and performance of our method is studied through an extensive suite of experiments in a realistic high-fidelity simulator (Webots). Results show that the total traveled distance and time are increased when humans are present in the environments. Localization noise is also increased particularly for the case of static people. In the second part of the experiments, a number of problematic cases resulting in longer modified plans, blocked passages, and long waits have been investigated. A comparative study targeting human-agnostic navigation and planning, human-aware navigation without considering humans in the planning phase, and human-aware navigation and planning has been conducted. Both simulated and real robot experiments confirm the effectiveness of accounting for humans on both team and individual levels for respecting social constraints as well as achieving a better performance based on MRTA metrics.

Official source

https://infoscience.epfl.ch/record/233804?ln=fr

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

In this paper, we propose a risk-based coordination method for the Multi-Robot Task Allocation (MRTA) problem in human-populated environments. We introduce risk-based bids that incorporate human trajectory prediction uncertainties and furthermore, social costs in their formulation. We demonstrate the effectiveness of including a predictive component in the risk formulation despite the lack of accurate position estimation for humans through an extensive suite of experiments. This is done by means of testing different levels of prediction error for known human trajectories and in a separate approach, using a Kalman filter for human trajectory estimation. Furthermore, we propose different risk formulations and evaluate their performance in a high-fidelity simulator. Additionally, a comparative study targeting human-agnostic planning at both navigation and planning levels, human-aware navigation and planning based on deterministic costs, and risk-based human-aware planning with no individual human-aware navigation has been conducted. Results confirm that risk-based bids lead to more socially acceptable team plans that reduce the need for the lower level individual human-aware navigation to be activated. Risk-based plans that account for social costs, prevent difficult social situations that can lead to less effective human-aware navigation, such as traversing narrow passages occupied by humans.

2018

Chargement

Risk-Based Human-Aware Multi-Robot Coordination in Dynamic Environments Shared with Humans

Alcherio Martinoli, Zeynab Talebpour

In this paper, we propose a risk-based coordination method for the Multi-Robot Task Allocation (MRTA) problem in human-populated environments. We introduce risk-based bids that incorporate human trajectory prediction uncertainties and furthermore, social costs in their formulation. We demonstrate the effectiveness of including a predictive component in the risk formulation despite the lack of accurate position estimation for humans through an extensive suite of experiments. This is done by means of testing different levels of prediction error for known human trajectories and in a separate approach, using a Kalman filter for human trajectory estimation. Furthermore, we propose different risk formulations and evaluate their performance in a high-fidelity simulator. Additionally, a comparative study targeting human-agnostic planning at both navigation and planning levels, human-aware navigation and planning based on deterministic costs, and risk-based human-aware planning with no individual human-aware navigation has been conducted. Results confirm that risk-based bids lead to more socially acceptable team plans that reduce the need for the lower level individual human-aware navigation to be activated. Risk-based plans accounting for social costs prevent difficult social situations that can lead to less effective human-aware navigation, such as traversing narrow passages occupied by humans.

IEEE2018

Chargement

Continual developments in robotic technology have enabled the use of robots in everyday applications in domestic, office and public spaces. Although single robot problems have been the main focus of social robotics research, applications of robots in social environments will not be limited to a single robot due to the increasing demand for robotic assistants and multi-robot operations. Multi-robot systems can achieve performances exceeding the sum of the individual robot contributions by exploiting the full potential of the team through information sharing, coordination, and joint decision-making. Robots operating in human-populated environments either directly interact with people or have to share the space with the humans. It is of utmost importance that people co-existing with robots feel safe and comfortable around them. This makes human-awareness essential for long-term sustainable deployment of robots in such environments. Furthermore, for cooperative robots, the presence of humans and their actions can directly affect the robot and team plans, making human-awareness more essential for ensuring high performance as well as social acceptability. Research in the area of socially-aware navigation has received substantial attention in recent years. However, despite their great potential, human-aware teams of robots considering social factors at both individual navigation and collective coordination and planning levels, are currently largely unexplored. In this thesis, we address the problem of human-aware cooperative navigation and coordination for multi-robot systems in realistic social environments. We focus on a class of multi-robot coordination problems known as multi-robot task allocation using a market-based approach. We explicitly consider the challenges of noisy, dynamic and stochastic human-populated environments by means of accounting for perception and prediction limitations and uncertainties in social cost modeling, bid estimation, coordination, and replanning. We construct an end-to-end framework comprising three main components of (i) human-aware navigation, (ii) human-aware coordination and planning for multi-robot systems, and (iii) human-robot interaction in the presence of multiple cooperative robots. We opt for an incremental approach to this problem starting from single robot human-aware navigation with expectation-based social costmaps. Subsequently, we move to multi-robot cooperative navigation in highly stochastic social environments. We propose human-aware coordination strategies based on social costs and social risks. The concept of risk introduced in this thesis incorporates perception and prediction uncertainties as well as social costs for estimating the stochastic costs of tasks that the robots should bid on in the market. Additionally, we introduce an adaptive risk-based replanning method for dealing with the limitations of local perception and unpredicted human behavior in the social environment. Finally, we demonstrate the interactive potential of the team of robots for social multi-robot task allocation by integrating an interaction that actively requests human collaboration and assistance in socially costly and blocking situations, into our adaptive replanning strategy. Extensive experiments with up to four robots and 12 humans in simulation, and up to two robots and two humans in reality have been carried out for evaluating the performance of the proposed methods in this thesis.

EPFL2018

Chargement

EPFL2018