Back
Virtual Lab
RF Spatial Research
Camera-independent spatial inference from RF measurements, motion sensing, and repeated physical observations.
Experimental ArchitectureProbabilistic RF Spatial Inference
Purpose
To what degree can coarse 3D physical occupancy and environmental geometry be probabilistically reconstructed from mobile phone trajectory tracking and ambient Wi-Fi RSSI attenuation?
How it works
Synthesizes phone inertial trajectory with ambient Wi-Fi signal strength drops. Casts 3D attenuation rays through a discretized spatial voxel grid to iteratively update probabilistic occupancy maps without camera imagery.
- Probabilistic Rather Than Geometric: RF attenuation represents Bayesian belief rather than directly observed physical boundaries.
- Multipath Acknowledgment: Explicitly models multipath reflection, antenna radiation patterns, and wall absorption uncertainty.
Key findings
Multipath Material Ambiguity
OBSERVEDRF attenuation alone cannot disambiguate thin metal sheets from thick concrete structures.
Body Shielding Distortion
OBSERVEDThe operator holding the sensing device introduces significant orientation-dependent attenuation.