CARIN: Wireless CSI-based Driver Activity Recognition under the Interference of Passengers

on YouTube ↗07m

" summary

CARIN uses Wi‑Fi Channel State Information to recognize driver activities despite passenger interference, building an offline database of single-person CSI and combination-based profiling to simulate multi-person signals by summing spectra. It cuts the Cartesian explosion with pressure-sensor reduction of seating combinations and a distance-minimization filtering that compares observed average Doppler shifts to simulated sums. Evaluated in a Toyota with 3,000 traces across 8 driver/10 passenger activities, it achieves 90.9% F1, 72% lower runtime, and +32.2% over baselines.

AI-written summary · about

" chapters (5)

1. 0:00Start
2. 1:28Methodology in CARIN • CARIN leverages a combination-based solution: • Profiles al possible activities by using Cartesian product. . Combination-based solution is not trivial.
3. 2:56Combination Reduction • Key observation: CSI signals from two separate activities can be added to simulate an activity combination.
4. 3:42Combination Reduction (Cont'd)
5. 4:51Experiment Setup