Distributed and Decentralized mixed-integer programming for collaborative autonomous driving

• Grant number: 25/01170-4
• Support Opportunities: Scholarships abroad - Research Internship - Post-doctor
• Start date: October 01, 2025
• End date: September 30, 2026
• Field of knowledge: Engineering - Electrical Engineering - Industrial Electronics, Electronic Systems and Controls
• Principal Investigator: Janito Vaqueiro Ferreira
• Grantee: Angelo Caregnato Neto
• Supervisor: Tamas Keviczky
• Host Institution: Faculdade de Engenharia Mecânica (FEM). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Institution abroad: Delft University of Technology (TU Delft), Netherlands
• Associated to the scholarship: 24/04703-0 - Distributed and decentralized mixed-integer programming for collaborative autonomous driving, BP.PD

Abstract

This research project addresses the motion planning problem for collaborative autonomous vehicles using distributed or decentralized mixed-integer programming (MIP). Unlike typical mathematical programming formulations, MIP offers the possibility of optimizing integer variables, which allows a multitude of relevant problems to be solved within the same framework, from obstacle avoidance to connectivity maintenance and decision-making. Although MIP-based algorithms have been extremely successful in jointly addressing the issues of motion planning and decision-making for multi-agent systems, the ensuing optimizations are typically centralized, imposing a limit on the scale and complexity of the solved problems. Several techniques for decentralized or distributed MIP have been proposed but their evaluation considering concrete engineering problems in motion planning and decision-making is still limited. This project presents a potential international collaboration in the form of a research internship with a group specialized in distributed optimization at Delft University of Technology, The Netherlands. The objective is to form an international group to address the gap between theory and practice within this field by investigating the viability of distributed and decentralized mixed-integer methods in the design of motion planning and decision-making modules considering the concrete engineering challenge of collaborative autonomous driving.