Cooperative planning and negotiation in human-robot teams

(English) As robots become increasingly integrated into everyday environments, rigid role paradigms and unilateral control models fall short of enabling meaningful collaboration. Preserving human autonomy while allowing robots to contribute proactively in shared decision-making tasks introduces the...

Descripción completa

Detalles Bibliográficos
Autor: Dalmasso Blanch, Marc|||0000-0003-1734-7863
Tipo de recurso: tesis doctoral
Fecha de publicación:2025
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/448567
Acceso en línea:https://hdl.handle.net/2117/448567
https://dx.doi.org/10.5821/dissertation-2117-448567
Access Level:acceso embargado
Palabra clave:human-robot interaction
human-robot collaboration
human-robot negotiation
human-robot plan negotiation
human-robot planning
multi-agent planning
004 - Informàtica
Àrees temàtiques de la UPC::Informàtica
Descripción
Sumario:(English) As robots become increasingly integrated into everyday environments, rigid role paradigms and unilateral control models fall short of enabling meaningful collaboration. Preserving human autonomy while allowing robots to contribute proactively in shared decision-making tasks introduces the need for alignment and negotiation between agents. Negotiation arises not merely as a design preference but as a requirement when autonomous entities with partial knowledge, differing capabilities, or misaligned goals must act jointly in real-world settings. This thesis investigates the challenge of integrating robots into human teams in unstructured environments, with a particular focus on Human-Robot Collaborative Navigation (HRCN). It seeks to empower them as active decision-making agents who flexibly and critically adapt to human preferences and needs. This technological development is framed as a social necessity: without it, robots would remain confined to controlled environments, or people would lose agency by having to adapt to rigid robot behaviour. The core contributions of the thesis are threefold. First, it introduces the Social Reward Sources (SRS) model, a shared spatial and task representation for Human-Robot Teams (HRT). Second, it presents a multi-agent planning system leveraging the SRS model to generate collaborative plans for heterogeneous teams. Third, it proposes a negotiation framework for Human-Robot Plan Negotiation (HRPN), incorporating a novel plan characterisation model, the cooperativeness space. These and additional secondary contributions are validated through real-world experiments within the collaborative object search benchmark. Altogether, the thesis offers a pathway for deploying robots as collaborative agents capable of negotiation, thereby supporting agency-preserving human-robot interaction in open-world contexts.