Felipe Tommaselli I am a first-year PhD student at the University of São Paulo (USP), advised by Prof. Marcelo Becker and co-advised by Dr. Ricardo V. Godoy. My research focuses on learning-based control and decision-making for robots, with a particular interest in understanding and improving learned representations. I also earned a BSc in Electrical Engineering at USP, with a minor in Computer Science, under the supervision of Prof. Robson Cordeiro. Previously, I was a visiting researcher at UIUC with Prof. Girish Chowdhary and worked on AI infrastructure as a software engineer at CloudWalk. Scholar / LinkedIn / Github / X

f.tommaselli@usp.br

Research

My experience spans legged, aerial, and manipulation systems. I began with applied field robotics and am now pursuing more fundamental questions in robot learning throughout my PhD. I am currently working on:

• motion retargeting and whole-body policy learning for Boston Dynamics’ Spot;
• contact-rich manipulation benchmarks with the Franka Research 3; and
• representation analysis during training for legged-locomotion policies.

Publications

* denotes equal contribution.

	Mind the Phase: Effective Rank and Representation Health in Legged Locomotion NEW F. Tommaselli, T. Segreto, J. Negri, R. V. Godoy, M. Becker Conference on Robot Learning (CoRL), 2026 project page / paper (code coming soon)
	Towards Capability-Aware Traversability Navigation for Unstructured Environments G. Capezzuto, F. Tommaselli, M. P. Angarola, R. V. Godoy, M. Becker IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026 project page / paper / video
	LeCropFollow: Latent Space Planning for Navigation in Unstructured Crop Fields F. Tommaselli, F. Affonso, A. Rocha, G. Capezzuto, A. N. Sivakumar, G. Chowdhary, M. Becker IEEE Robotics and Automation Letters, vol. 11, no. 9, 2026 project page / paper / video / code
	Learning to Walk with Less: A Dyna-Style Approach to Quadrupedal Locomotion F. Affonso, F. Tommaselli, J. H. Aléssio, V. S. Medeiros, M. V. Gasparino, G. Chowdhary, M. Becker IEEE Access, vol. 14, 2026 paper
	VLN on the Fly: An Onboard Vision-Language Navigation Stack for Aerial Robots M. S. Tayar*, F. Tommaselli*, G. Capezutto*, P. A. R. Saraiva*, P. H. V. de Freitas, L. Kido, G. Sonego, R. V. Godoy, M. Becker International Micro Air Vehicle Conference (IMAV), 2026 project page / paper / code
	CROW: A Self-Supervised Crop Row Navigation Algorithm for Agricultural Fields F. Affonso*, F. Tommaselli*, G. Capezzuto, M. V. Gasparino, G. Chowdhary, M. Becker Journal of Intelligent & Robotic Systems, vol. 111, no. 1, article 28, 2025 paper / code
	Autonomous UAV Flight Navigation in Confined Spaces: A Reinforcement Learning Approach M. S. Tayar, L. K. de Oliveira, F. Tommaselli, J. D. Negri, T. H. Segreto, R. V. Godoy, M. Becker Latin American Robotics Symposium (LARS), 2025 paper

Selected Projects

I enjoy building open software and hardware. As an undergraduate, I developed projects for hackathons and competitions; now, I seek to keep building research that others can reproduce, understand, and extend.

	HiveBoard An open, modular, fully 3D-printed benchmark of 13 industrial mechanisms, validated across Spot, ANYmal, a tabletop arm, and a prosthetic hand. Under review project / code / video
	Swarm in Blocks A block-based platform for programming and monitoring multi-UAV swarms, developed with SEMEAR-USP. CopterHack, 1st place, 2022 code / video