I am a Brazilian master’s student in Electronic and Information Engineering at Tsinghua University, supported by a full Chinese Government Scholarship.
I want to use technology to make people’s lives easier, and I see sensing as a great opportunity to do that. I am interested in how different kinds of sensors capture the world, what each one can and cannot see, and how each of them perceives people and the environment in its own way. My current research is on multimodal wireless sensing, using everyday signals such as Wi-Fi channel state information (CSI) and visible light to recognize human activity and detect falls indoors, without cameras or wearables. I built a synchronized visible-light + Wi-Fi testbed and study when the two modalities complement each other, together with a Bayesian fusion that stays robust when one of them is degraded.
Although my background is in computer science, I have always been drawn to the physical side of systems: building setups, running experiments and understanding how hardware actually behaves. Before graduate school, I spent five years at Stone (NASDAQ: STNE), where I led the integration of Android payment hardware, bringing nine point-of-sale devices into production and working closely with manufacturers such as Ingenico, Sunmi and Positivo on embedded and IoT integration. I hold a B.Sc. in Computer Science from Universidade Federal Fluminense, Brazil.
IEEE ICASSP 2027 Workshop on Wireless AI Sensing (WASAI)
A synchronized visible-light + Wi-Fi sensing testbed, a 12-participant dataset including falls, an information-theoretic model of light/Wi-Fi complementarity, and a Bayesian decision-level fusion that stays robust when one modality is degraded.
Extended Proceedings of the XXIII Brazilian Symposium on Human Factors in Computational Systems (IHC 2024), Brasília, pages 86 to 90
Original title in Portuguese: "Como podemos ajudar pessoas com deficiência visual a se sentirem seguras durante pagamentos no mundo físico"
★ Top scores for presentation and communication, Posters & Demos track
@inproceedings{mousinho2024ihc,
author = {Mousinho, Dario and Beltrão, J. and Trevisan, D. and Romero, V.},
title = {Como podemos ajudar pessoas com deficiência visual a se sentirem seguras durante pagamentos no mundo físico},
booktitle = {Anais Estendidos do XXIII Simpósio Brasileiro sobre Fatores Humanos em Sistemas Computacionais (IHC 2024)},
pages = {86--90},
year = {2024},
publisher = {SBC},
address = {Brasília, Brazil},
doi = {10.5753/ihc_estendido.2024.243884}
}Synchronized testbed with a modulated LED and four photodiodes sampled at 200 kHz on an FPGA, plus an ESP32-C5 802.11ax link reporting CSI on 245 subcarriers. Dataset of 12 participants, 7 activities including falls, 3 room conditions (252 recordings, 5,688 windows).
Raspberry Pi system that captures Wi-Fi frames and estimates heart rate from CSI in real time using band-pass filtering, outlier removal, moving averages and Fourier analysis. Live demo at the VIII MidiaCom Laboratory Workshop.
Time-Varying Data Science Group, Tsinghua University
Institute of Future Human Habitat, Tsinghua University
MidiaCom Lab, Universidade Federal Fluminense
Tsinghua University
Stone (NASDAQ: STNE), Brazil
Tsinghua University
Universidade Federal Fluminense (UFF), Brazil