Murilo Fernandes Martins

About

Murilo Fernandes Martins is a robotics and artificial intelligence researcher whose work spans reinforcement learning, robot manipulation, multi-objective optimization, and autonomous mobile systems. His most influential contribution, "Learning Gentle Object Manipulation with Curiosity-Driven Deep Reinforcement Learning" (2019, 45 citations), introduced a framework enabling robots to handle fragile objects carefully during both exploration and execution — a critical advance for deploying robots in delicate real-world settings. Complementing this, his benchmarking framework for soft robotic end effectors (2018, 17 citations) provided a rigorous evaluation methodology for compliant grippers handling deformable objects like fruits and vegetables in industrial contexts. Martins has also made notable strides in multi-objective policy optimization, proposing distributional approaches to handle competing objectives at different scales (2020, 23 citations), and in vision-based robot learning, demonstrating fast real-world training through multi-task reinforcement learning (2019, 15 citations). His earlier work addressed multi-robot learning from demonstrations (2010, 22 citations) and qualitative probabilistic approaches to mobile robot localization (2013, 9 citations), reflecting a career-long commitment to making autonomous systems more adaptive and practically deployable. Across more than a decade of research, Martins has consistently bridged foundational machine learning with applied robotics challenges.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning Gentle Object Manipulation with Curiosity-Driven Deep Reinforcement Learning

   45 citations · 2019
2. 2
   A Distributional View on Multi-Objective Policy Optimization

   23 citations · 2020
3. 3
   Learning multirobot joint action plans from simultaneous task execution demonstrations

   22 citations · 2010
4. 4
   A Benchmarking Framework for Systematic Evaluation of Compliant Under-Actuated Soft End Effectors in an Industrial Context

   17 citations · 2018
5. 5
   Simultaneously Learning Vision and Feature-Based Control Policies for Real-World Ball-In-A-Cup

   15 citations · 2019
6. 6
   Heuristically-Accelerated Reinforcement Learning: A Comparative Analysis of Performance

   11 citations · 2014
7. 7
   A Qualitative-Probabilistic Approach to Autonomous Mobile Robot Self Localisation and Self Vision Calibration

   9 citations · 2013
8. 8
   On the construction of a RoboCup small size league team

   8 citations · 2011
9. 9
   Probabilistic self-localisation on a qualitative map based on occlusions

   8 citations · 2016
10. 10
    Simultaneously Learning Vision and Feature-based Control Policies for Real-world Ball-in-a-Cup

    6 citations · 2019

Key Collaborators

Contact & Links