Physical computing across magnetic and RF domains for passive wireless information retrieval
A passive physical computing system that transforms local magnetic signals directly into RF backscatter, retrieving magnetic information wirelessly without on-device digital signal processing.
Mechanically programmable metasurface for ultralow-power subterahertz beam steering
A mechanically programmable metasurface that steers sub-terahertz beams through its surface geometry, holding beam states without power for continuous steering and tracking.
Event-based Visible Light Backscatter Communication System in Underwater Environment
An underwater communication system that combines liquid-crystal polarization modulation with event-camera decoding for low-power data links in underwater IoT.
02
Wireless without Batteries
Connecting Devices without Batteries
A low-power single-channel EEG sensor for fatigue monitoring and command interaction
A lightweight single-channel EEG sensor that combines dry electrodes, low-power backscatter, and neural reconstruction for fatigue monitoring and hands-free command interaction.
A magneto-mechanical bio-interface with ultra-low-power analog backscatter for inclusive cardiac monitoring
A wearable bio-interface that converts heart-induced skin micro-vibrations into magnetic signals and transmits them via analog backscatter for low-power cardiac monitoring.
Mighty: Towards Long-Range and High-Throughput Backscatter for Drones
Mighty pairs a low-power backscatter radio with a lightweight video pipeline that shifts heavy processing to the ground controller, cutting the energy cost of long-range drone video streaming.
mmExpert: Integrating Large Language Models for Comprehensive mmWave Data Synthesis and Understanding
mmExpert is a scalable framework that generates diverse mmWave-language data from the world knowledge of large language models and enables open-vocabulary mmWave understanding through language-aligned pre-training and a versatile interpreter.
Privacy Preserving Two Hand Reconstruction Using Millimeter Wave Radar for Sign Language Understanding
mmPRISM reconstructs two-hand skeletal poses from raw 4D millimeter-wave radar and feeds them to a large language model for privacy-preserving, open-vocabulary sign language understanding.