Carlos Quintero-Peña

PhD Student in Computer Science

Biography

I am a PhD student in Computer Science at Rice University, working under the supervision of Lydia Kavraki and Anastasios Kyrillidis. I am honored to have received a Fulbright Scholarship in 2019 to start my doctoral studies.

I am interested in understanding how to plan safe motions for robots in unstructured environments. My goal is to provide robots with the capabilities of autonomously deciding and acting in environments that may have noisy or incomplete information. To this end, I have explored the use of optimization and data-driven models that enable the robots to perform complicated tasks safely and reliably.

Before joining Rice, I was an instructor and researcher at Universidad de los Andes and Universidad Santo Tomás where I thaught courses mainly in Electronic Engineering, Robotics and Machine Learning. During this time I had the chance to join the RoboCup community by becoming part of the STOx’s team of the Small Size League (2014-2017), serving as TC member in 2018 and by co-founding the SinfonIA team of the @Home Social Standard Platform League in 2019.

Download my resumé.

Interests

Robotics
Optimization
Machine Learning

Education

PhD in Computer Science, 2024 (expected)

Rice University
MSc in Computer and Electronic Engineering, 2011

Universidad de los Andes
BSc in Electronic Engineering, 2009

Universidad de los Andes

News

[03/01/2023] Our paper Optimal Grasps and Placements for Task and Motion Planning in Clutter has been accepted at ICRA 2023!
[12/05/2022] CS Rice wrote the note Robotics researchers develop open-source motion planning benchmarking tool about our work with my collegue Constantinos Chamzas on Benchmarking Motion Planning Algorithms
[10/31/2022] I gave my second COMP600 talk named “Human-Guided Motion Planning in Partially Observable Environments”
[10/23/2022] Check out our workshop on Evaluating Motion Planning Performance: Metrics, Tools, Datasets, and Experimental Design happening at IEEE/RSJ IROS 2022, Kyoto, Japan
[09/01/2022] I was selected to TA COMP450 - Algorithmic Robotics - taught by Dr. Kavraki for the Fall 2022
[06/15/2022] Check out the media article made by Rice University Humans in the loop robots find their way about our work on Human-Guided Motion Planning. IEEE Spectrum also picked it up!
[01/31/2022] Our paper Human-Guided Motion Planning in Partially Observable Environments has been accepted at ICRA 2022!
[11/18/2021] Our paper MotionBenchMaker: A Tool to Generate and Benchmark Motion Planning Datasets has been accepted for publication in the IEEE Robotics and Automation Letters (RA-L)
[05/14/2021] Our paper Motion Planning via Bayesian Learning in the Dark has been accepted for spotlight presentation at the Machine Learning for Motion Planning (MLMP) workshop as part of ICRA 2021
[02/28/2021] Our paper Robust Optimization Motion Planning for high-DOF Robots under Sensing Uncertainty has been accepted at ICRA 2021!
[02/28/2021] Our paper Learning Sampling Distributions Using Local 3D Workspace Decompositions for Motion Planning in High Dimensions has been accepted at ICRA 2021!
[12/10/2020] I was invited to give a talk at Seminario Internacional Sistemas Robóticos Autónomos y Aplicaciones
[12/09/2020] I was invited to give a talk at Uniandes
[10/19/2020] I gave my COMP600 presentation named “Optimization, Robots and Uncertainty”
[09/01/2020] I was selected to TA COMP450 - Algorithmic Robotics - for the Fall 2020
[01/01/2020] I joined the Kavraki Lab
[08/25/2019] I started as a PhD student in the Computer Science Department at Rice University

Projects

Human-Guided Motion Planning in Partially Observable Environments

Human-in-the-loop algorithm to compute joint-space trajectories for high-DoF robots under partial observability

Can Theoretical Algorithms Efficiently Escape Saddle Points in Deep Learning?

Review of optimization algorithms that can escape saddle points in Deep Learning and some experimental results

Inverse Kinematics Robo Picasso

Solving the inverse kinematics problem for a FANUC S-500 robot and using it to draw a Mickey Mouse

Motion Planning with incomplete scene information

A Fetch robot doing Motion Planning and executing a trajectory when the scene representation may be incomplete

Neural Network Pruning - A review

Review of methods to prune neural networks

Robust Motion Planning under Sensing Uncertainty

Optimization-based motion planning algorithm capable of incorporating sensing uncertainty for a variety of noise models

Contact

carlosq [at] rice [dot] edu
Duncan Hall, Rice University, Houston, TX 77005