Miguel Angelo de Abreu de Sousa

Federal Institute of São Paulo, TU Wien, Universidade Politecnica, Universidade de São Paulo

Papers

Total Citations

H-Index

About

Miguel Angelo de Abreu de Sousa is a researcher at the forefront of neuroengineering and autonomous robotics, specializing in the hardware implementation of neural computation models and intelligent robotic control systems. His most impactful work, "An FPGA distributed implementation model for embedded SOM with on-line learning" (22 citations), pioneered efficient hardware architectures for executing both learning and recall phases of artificial neural networks directly on embedded systems. This contribution is critical for advancing real-time, on-chip machine learning in resource-constrained environments. In the domain of robotics, Sousa developed the DeWaLoP remote control system for in-pipe robots (13 citations), addressing complex networked control challenges for infrastructure inspection. His foundational research on adaptive automata for robotic mapping and navigation in unknown environments (11 and 5 citations) established novel methods for autonomous exploration without prior maps. Additionally, he created CoopDynSim (4 citations), a multi-robot 3D simulator designed to seamlessly transfer controller code from simulation to real hardware. Sousa’s work bridges the gap between theoretical neural computation and practical, deployable robotic systems, making significant contributions to embedded AI and autonomous navigation.