Pedro Santos

Solvable does not mean faithful: automated planners routinely validate and execute LLM-generated PDDL instances that completely corrupt the user's intended initial states and goal specifications.

Intermediate task decomposition in LLM-agent systems can improve accuracy but fails to consistently outperform a single agent in VAT determination tasks.

Achieving over 96% accuracy in exercise quality assessment could revolutionize remote rehabilitation by enabling effective, therapist-free patient monitoring.

Reprojection error can mislead 3D accuracy assessments, but a new ray-based calibration framework reveals a more reliable path to precision.

QuadRocket achieves precise trajectory tracking and disturbance compensation, paving the way for innovative thrust-vector control in aerial robotics.