| Title: | PhD Student |
|---|---|
| Affiliation: | Robotics Research Group, KU Leuven |
| Email: | brendan.pousett@kuleuven.be |
| ORCID: | 0000-0002-8375-8452 |
| Résumé: | |
| Publications: | link to lirias (more coming soon 😀) |
Note: I have not finished translating my previous, industry-focused résumé to the academic CV format, so I have linked it above. My old website has further details on projects I worked on before my PhD.
I received by B.A.Sc. from UBC in Engineering Physics in 2019, during which I completed internships at NZ Technologies, A&K Robotics, Tesla, and Genesis Motion Solutions. Upon graduating with academic distinction (Dean's Honour List, Trek Excellence Scholarship), I continued working with Genesis Motion, focused on designing, analysing, integrating, and testing novel geared-electric actuator technologies, in collaboration with partners such as HYUNDAI and Bastian Solutions. I developed a batch analysis tool which utilized a proprietary efficiency model and an evolutionary optimization algorithm to generate and simulate actuator design variants. This resulted in a 1000x speedup of our simulation process, and improved the operating efficiency of prototypes by a factor of 2.
After a strong focus on industry, I began my academic career as a PhD student at KU Leuven in 2023, supervised by Prof. Herman Bruyninckx and Dr. Wilm Decré, focused on mechanical design and model-based control of robots for paving, structural masonry, and prefabricated building panel installation. My interests lie in co-development of sensing and control methodologies on bespoke hardware. My motivation is to build low-cost prototypes which fulfill the correct accuracy requirements for an application, and are modelled by quasi-linear systems. I seek to combine these with convex optimization-based controllers which are robust to disturbances.
I build my own robot prototypes, which gives me a holistic view of model-based robotics. Model parameters can be optimized during the design phase for better control performance, and the most elegant solutions occur when hardware and control complement each other. Here are some systems I have built during my PhD research: