Joan Palacios

mmTracking

My paper "Tracking mm-Wave Channel Dynamics: Fast Beam Training Strategies under Mobility" consisted on a theoretical analysis of the mmWave sparse channel dynamics for fast beam training strategy. The main idea consisted on extracting channel evolution information out of decoded data packages. I published this paper during my PhD studies at Pivotal Commware and it remains as my most cited paper with 200 citations.

Close Window

ACO

My paper Adaptive Codebook Optimization for Beam Training on Off-the-Shelf IEEE 802.11ad Devices consisted on a channel estimation and beam training algorithm for commercial IEEE 802.11ad devices. It leveraged harmonic analysis properties of individual element tweaks to be able to extract a channel estimation out of absolute magnitude measurements. Its ability to work in a real environment with real hardware makes it one of my most useful tools developed during my PhD studies at Pivotal Commware. One of its applications, "Demo: Channel Estimation and Custom Beamforming on the 60 GHz TP-Link Talon AD7200 Router" earned the Best Demo Award at ACM WinTech 2018. Its code is available on Github.

Close Window

Satellites

During my stay as a postdoc at North Carolina State University, my research focused mainly on satellite communications. My first work was on studying the impact of hybrid beamforming on the performance of a satellite communication systems in my paper "A Hybrid Beamforming Design for Massive MIMO LEO Satellite Communications". My second and more important work "A Dynamic Codebook Design for Analog Beamforming in MIMO LEO Satellite Communications" proposed a pre-computed, cycling dynamic codebook that removed the need for users to switch between different satellite sector index.

Close Window

MOMP

The research I'm most proud of: MOMP/SMOMP: (Separable) Multidimendional Orthogonal Matching Pursuit. Developed during my stay as a postdoc at North Carolina State University, (S)MOMP consisted on a generalization of the sparse coding algorithm OMP that leveraged the structure and properties of tensors in order to exploit the nature of sparsity in multi-dimensional problems. Its initial main application consisted in blazingly fast accurate large sparse channel estimation where the number of channel matrix entries exceeded the millions, but since then its applications have extended to other unrelated problems. I turned the work of these paper and other related works into a Python package available on Github. The Github page contains links to the main paper and other derived works.

Close Window

Chamber controls

One of the first tools I ever developed during my work at Pivotal Commware was software control for anechoic chambers. This software abstracted the physical hardware and allowed the user to control all chamber instruments using high level commands. It was able to handle a queue of experiments to be executed and was aware of its current state for setting each experiment. This software had 4 concurrently working interfaces satisfying different user cases:

• Python integration: Experience users that wanted to write their own experimentation scripts.
• Web interface: Operators that wanted to queue/launch pre-defined experiments.
• REST API: Advanced users that wanted to build their own tools that integrated with the chamber.
• Executable GUI: Operators that wanted to use common chamber routines through a simple graphical interface.

This tool ended up becoming so useful that it replaced all previous chamber softwares and increased chamber uptime so much that we reduced the number of anechoic chambers.

Close Window

Cyclops

During my work at Pivotal Commware, I was the researcher behind the Cyclops project. The Cyclops leveraged channel estimation through beamforming sensing and image distortion analysis to provide an intuitive view of radiated signals in the physical world. Cyclops achieved its goal by coordinating its internal antennas, signal sensors and cameras to deliver a picture of an overlapped signal heatmap on a camera feed accompanied by an interactive KPI report. I took care of the entire project during its multiple prototyping stages and did most of the work during its production. Cyclops was awarded the 2025 Fierce Network Innovation Award for the Network Test and Measurement cathegory.

Close Window