LiveSense: A Real-Time Wi-Fi Sensing Platform for Range-Doppler on COTS Laptop
| Authors | Jessica Sanson et al. |
| Year | 2026 |
| Field | AI / Agents |
| arXiv | 2603.06545 |
| Download | |
| Categories | cs.AI |
Abstract
We present LiveSense - a cross-platform that transforms a commercial off-the-shelf (COTS) Wi-Fi Network Interface Card (NIC) on a laptop into a centimeter-level Range-Doppler sensor while preserving simultaneous communication capability. The laptops are equipped with COTS Intel AX211 (Wi-Fi 6E) or Intel BE201 (Wi-Fi 7) NICs. LiveSense can (i) Extract fully-synchronized channel state information (CSI) at >= 40 Hz, (ii) Perform time-phase alignment and self-interference cancellation on-device, and (iii) Provide a real-time stream of range, Doppler, subcarrier magnitude/phase and annotated video frames to a Python/Qt Graphical User Interface (GUI). The demo will showcase the ability to detect (i) Distance and radial velocity of attendees within a few meters of the device, (ii) Micro-motion (respiration), and (iii) Hand-gesture ranging. To the best of our knowledge, this is the first-ever demo to obtain accurate range information of targets from commercial Wi-Fi, despite the limited 160 MHz bandwidth.
Engineering Breakdown
Plain English
LiveSense transforms a standard Wi-Fi network card on a laptop into a real-time sensing system that can detect people's distance and movement within a few meters, all while maintaining normal internet connectivity. The system extracts channel state information (CSI) at 40+ Hz from Intel Wi-Fi 6E/7 NICs and performs signal processing on-device to align timing, cancel self-interference, and output range-Doppler data with centimeter-level accuracy. This enables practical applications like contactless gesture recognition, micro-motion detection, and occupancy sensing without specialized hardware or disrupting network use. The key innovation is making this work with unmodified commercial Wi-Fi chips that weren't designed for sensing, and doing it fast enough (40+ Hz) for real-time applications.
Core Technical Contribution
The core novelty is demonstrating that COTS Wi-Fi 6E/7 NICs can be repurposed for precise Range-Doppler sensing through software-only processing pipelines, without hardware modification or firmware changes. The authors developed on-device signal processing algorithms for time-phase alignment and self-interference cancellation that extract usable sensing data from standard CSI measurements while preserving normal Wi-Fi communication. Previous work either required specialized hardware modifications, proprietary access to radio firmware, or operated at lower frame rates; LiveSense achieves 40+ Hz synchronized CSI extraction with real-time GUI streaming on unmodified Intel NICs. This is a systems contribution showing how to bridge the gap between research sensing platforms and commercially deployable solutions.
How It Works
The pipeline starts with a COTS Intel AX211 or BE201 Wi-Fi NIC that captures channel state information (CSI) — the phase and amplitude response of the wireless channel across multiple subcarriers. The system extracts this CSI at 40+ Hz or higher, ensuring time synchronization across antenna elements so range and velocity estimates are coherent. On-device processing performs two critical operations: (1) time-phase alignment to correct hardware timing skew and phase offsets, and (2) self-interference cancellation to remove the direct transmit leakage that would otherwise dominate the sensing signal. The aligned CSI is then transformed via FFT and Doppler processing to generate range bins and velocity estimates for each target, outputting magnitude/phase per subcarrier along with annotated video frames to a Python/Qt GUI for real-time visualization and logging. This entire pipeline runs on the laptop itself, keeping latency low enough for live interactive applications.
Production Impact
For engineers building sensing systems, LiveSense eliminates the need for specialized hardware — you can deploy on any laptop with Wi-Fi 6E/7, dramatically reducing hardware costs and complexity. Real-world applications include smart home presence detection (no camera privacy concerns), occupancy-based HVAC control, gesture-based interfaces in industrial settings, and healthcare monitoring (fall detection, respiration sensing) without wearables. Integration is straightforward since the system outputs standard Python-consumable data streams (range, Doppler, CSI) and works alongside normal network traffic, making it feasible to retrofit into existing deployments. The trade-off is that you're limited to line-of-sight ranges (typically 5-10 meters depending on environment) and depend on Wi-Fi transmission opportunities; you also need enough CSI resolution in your NIC, which newer Intel cards provide but older ones may not. Latency is manageable (40+ Hz) for most use cases, but dense multi-user environments or heavily congested Wi-Fi bands may degrade performance.
Limitations and When Not to Use This
The approach is fundamentally limited by Wi-Fi's transmission bandwidth and CSI granularity — you cannot detect targets beyond line-of-sight and range resolution degrades with fewer subcarriers (40 MHz bandwidth = ~3.75 m range resolution). The system requires COTS Wi-Fi 6E/7 NICs with accessible CSI output; older Wi-Fi 5 or non-Intel hardware may not support the necessary CSI extraction or frame rate, limiting deployability to recent devices. Environmental factors like multipath propagation, metal/water interference, and interference from nearby Wi-Fi networks can degrade sensing accuracy, and the paper doesn't thoroughly characterize robustness across diverse real-world indoor/outdoor settings. The self-interference cancellation algorithm assumes relatively stable transmit-receive isolation and may fail in certain antenna configurations; multi-user scenarios with many simultaneous transmitters could overwhelm the signal processing pipeline. Follow-up work is needed on: cross-platform support (non-Intel NICs), environmental robustness benchmarking, privacy-preserving filtering (to avoid eavesdropping on others' data), and scaling to larger spaces or outdoor range.
Research Context
This work builds on a decade of Wi-Fi sensing research that has shown CSI can reveal fine-grained motion and presence, but prior efforts (WiGig, FMCW radars adapted for Wi-Fi) required either custom hardware or proprietary firmware access, limiting real-world adoption. LiveSense draws from recent advances in commodity Wi-Fi NIC capabilities (Intel's CSI extraction APIs on newer generations) and signal processing techniques for MIMO radar that the sensing community adapted from automotive radar literature. The research extends benchmark efforts like Widar and RT-Fall that quantified Wi-Fi sensing accuracy but relied on specialized setups; LiveSense aims to bring those capabilities to unmodified COTS hardware. This opens a new research direction: treating commercial Wi-Fi as an ambient, ubiquitous sensing modality, similar to how recent work exploits FMCW radar in smartphones — if successful at scale, it could shift Wi-Fi sensing from a niche research topic to a practical systems primitive in deployed networks.
:::tip Subscribe Get weekly breakdowns of papers like this in AI Letters - the newsletter for engineers building production AI systems. :::