Looking for MSCA PF 2026 candidates: Universidad de Sevilla (Spain) - Department of Engineering

Brief description of the Centre/Research Group 

The University of Seville is one of the most important higher education institutions in Spain, with 70,000 students and 4,600 professors. Moreover, Seville is the largest city in the south of Spain. Situated on the plain of the river Guadalquivir, and with a population of about 1.5 million, Seville enjoys favorable wheather conditions and contains three UNESCO Word Heritage Sites.  The researcher will work with the Multi-robot & Control Systems (MACS) group, with extensive experience in cooperative teams of heterogeneous robots, including autonomous ground, aerial, and marine vehicles. MACS members have participated in 15 European and 11 national projects, developing multi-drone systems for cinematography (MULTIDRONE, H2020-731667), power line inspection (AERIAL-CORE, H2020-871479), solar plant inspection (Interreg DURABLE), or human-robot teaming for construction (COBUILD, Spanish government).   MACS Principal Investigators led the USE team that participated in in the Mohammed Bin Zayed International Robotics Competition (MBZIRC 2017, 2020), winning one of the Challenges in 2020. This is a remarkable achievement, as MBZIRC’20 proposed several heterogeneous multi-robot challenges and many of the world’s leading robotics labs participated. In terms of technology transfer activities, MACS has a longstanding tradition, having collaborated with multiple companies such as Airbus, Iturri, EMASESA, Navantia S.A., or EDP Portugal.

Project description

This project will address multi-agent planning and learning methods to enhance the efficiency and effectiveness of collaborative human-robot tasks in dynamic scenarios such as construction or logistics, by addressing uncertainty, battery constraints, and adapting to changing conditions.  The focus will be on high-level task planning under uncertainty (due to action duration, probabilities of task failure, human performance or intentions, etc.). On the one hand, new online Monte-Carlo Tree Search algorithms, both centralized and decentralized, will be developed to solve task allocation problems for human-robot teams modeled as Partially Observable Markov Decision Processes, seamlessly integrating robot recharges and multi-agent tasks requiring synchronization between agents. Recharges are crucial for extending operational periods, while synchronized multi-agent tasks are needed for load transportation or human-robot interaction.  On the other hand, novel multi-agent reinforcement learning techniques will be developed for mixed human-robot teams, enabling robots to adapt to Dynamic environments by learning probabilistic models online. Using realistic simulation tools, the system will learn worker behaviors and routines (e.g., worker effectiveness or usual working hours) to optimize joint plans, anticipating factors like fatigue and human presence. Direct worker feedback, e.g.,  expressed as preferences, will then be used to guide the reinforcement learning algorithms and shape team plans online during operation. These interactions will also help learn the effectiveness of each worker and optimizing task allocation, as the level of satisfaction may decrease when a person receives many unwanted tasks, potentially impacting the worker’s performance when completing tasks.

Evaluation panel:

Information Science and Engineering (ENG)

 IMPORTANT: Applicants should check their elegibility to apply to the MSCA PF 2026 Call: 

• Any nationality (European Fellowship) / European national or long-term resident of a Member State or Horizon Europe Associated Country (Global Fellowship)

• Be in a possession of a doctoral degree (PhD)

• Have a maximum of 8 years full-time equivalent experience in research, measured from the date of award of the doctoral degree

• Comply with the mobility rule: not have resided or carried out their main activity (work, studies, etc.) in Spain (for European Postdoctoral Fellowships), or in the country of the host organisation for the outgoing phase (for Global Postdoctoral Fellowships) for more than 12 months in the 36 months immediately before the call deadline.

 For a correct evaluation of your candidature, please send the documents below to Professor  Jesús Capitán (jcapitan@us.es)

• Detailed CV

• Brief description about research interests

• Name and contact of 1/2 references (if available).

If you are ultimately selected to submit your application alongside Professor Jesús Capitán, you will be able to benefit from the Coaching and Advising Programme offered by the International Projects Office at USE, providing additional resources and a proposal review service to the MSCA PF 2026 candidates.