WiMoCSIWiFi Mobility & Health Monitoring

Research Platform

WiMoCSI: Advancing Reliable Wi-Fi Mobility Sensing

Using physics-based simulation, controlled commodity Wi-Fi CSI experiments, and evidence-aware signal processing to investigate walking-speed measurement reliability and future healthy-aging and ADRD research.

Conceptual illustration of Wi-Fi mobility sensing: a wireframe living room with an older adult walking between a wall-mounted transmitter and receiver, a router, a laptop showing signal waveforms, and icons for walking, heart activity, and brain health

WiMoCSI Platform at a Glance

Physics-Based Wi-Fi Simulation

Model human movement and wireless signal propagation in realistic indoor environments.

Demonstrated

Real-World Wi-Fi Measurements

Collect Channel State Information (CSI) using commodity Intel AX201 devices during controlled walking experiments.

Demonstrated

Reliability-Aware Mobility Estimation

Evaluate walking-speed estimates, uncertainty, and when sensing results should be withheld.

Under Evaluation

Healthy Aging & ADRD Research

Pursue future in-home mobility monitoring applications for older adults and Alzheimer's disease and related dementias (ADRD) research.

Long-Term Research Goal

Two Complementary Ways to Investigate Human Movement

Simulate Human Movement

Older adult walking through a simulated senior-living room with Wi-Fi devices and multipath propagation paths

Create controlled virtual walking experiments using physics-based ray tracing. Configure the room, antenna locations, and walking motion to investigate how human movement changes Wi-Fi Channel State Information (CSI) under known conditions.

  • Known walking-speed reference
  • Reproducible experimental scenarios
  • Configurable propagation conditions

Measure Real-World Movement

Older adult walking a green-taped path in a senior living community hallway between two tripod-mounted Wi-Fi transmitter and receiver devices

Use commodity Wi-Fi devices to record how wireless signals change during controlled walking trials. Compare CSI-based motion and walking-speed estimates with independently measured walking references to assess accuracy and reliability.

  • Intel AX201 CSI measurements
  • Controlled walking experiments
  • Independent video-based references

From Controlled Modeling to Real-World Evidence

WiMoCSI combines physics-based simulation with real-world Wi-Fi measurements to investigate how human movement affects wireless signals and how reliably those changes can be measured. Controlled simulations allow researchers to study signal behavior under known conditions, while physical walking experiments reveal the opportunities and limitations of commodity Wi-Fi sensing. Together, these complementary approaches strengthen scientific validation, guide the development of more reliable mobility sensing, and support future applications in healthy aging and in-home monitoring.

From Wi-Fi Signals to Reliable Mobility Measurements

A transparent research workflow connecting simulated and measured Wi-Fi signals with motion analysis, reliability checks, and independent validation.

CSI Sources

Simulated CSI

Physics-based ray tracing with known human motion.

Measured CSI

Intel AX201 recordings with independent walking references.

Signal Analysis
1

Signal Preparation

Align and preprocess Wi-Fi CSI for consistent motion analysis.

2

Data Quality Assessment

Check packet timing, missing data, and signal suitability.

3

Motion Feature Extraction

Identify temporal and frequency-domain patterns linked to human movement.

4

Evidence-Aware Estimation

Estimate walking speed when supported by the data; otherwise withhold the result.

Validation
5

Independent Validation

Compare estimated mobility with reference measurements and quantify errors and failures.

Why Reliability Matters

For future mobility monitoring in older adults and people living with cognitive impairment, knowing when a Wi-Fi measurement cannot be trusted is just as important as obtaining an accurate estimate.

WiMoCSI virtual laboratory rendering: a 3D room with a walking person, Tx and Rx devices, direct and reflected propagation paths, the experiment configuration panel, and simulated CSI plots

Inside the virtual laboratory

A ray-traced virtual laboratory for indoor CSI generation, where room, radio, and motion conditions can be configured to study how the environment and human motion affect wireless sensing.

  • Room geometry and materials
  • Transmitter and receiver placement
  • Walking trajectories and speed
  • Multipath and human-scattering assumptions
  • Reproducible simulation configurations
View Simulation Details

Experimental evidence & results

Ray-Traced Speed Recovery

Scatter plot comparing recovered speed with commanded speed for four speeds and two stochastic seeds, with a 1:1 reference line and a shaded error band

How well walking speed can be recovered from simulated CSI under different propagation assumptions.

Source: WiMoCSI simulation code

AX201 Walking Experiments

Isometric floor plan of the 11 by 11 foot experiment room showing the transmitter on the bookshelf, the receiver laptop, the 7 foot line of sight, and the 7 foot walking path between points A and B

Speed estimates from commodity Wi-Fi CSI recorded during controlled walking trials, compared with independent references.

Source: controlled experimental records

Retrospective Timing Analysis

Retrospective lap-timing recovery chart comparing estimated lap periods with reference timing, including cases where the estimate is withheld

Retrospective lap-timing recovery, including cases where the estimate is withheld because evidence is insufficient.

Source: retrospective evaluation

Figures are generated from WiMoCSI code and experimental records and should be read within their experimental conditions and scope.

Engineering Trustworthy Wi-Fi Mobility Sensing

Understanding measurement uncertainty, recognizing unreliable estimates, and validating Wi-Fi sensing under real-world conditions.

Sources of Measurement Uncertainty

A walking person between a Wi-Fi router and a CSI laptop, with indicators for packet loss, irregular sampling, and timing perturbations

Measurement Challenges

Real-world Wi-Fi measurements are affected by packet loss, irregular sampling, timing variations, and changing signal conditions. WiMoCSI investigates how these factors influence motion detection and walking-speed estimation.

Reliability-Aware Estimation

We investigate signal-quality checks, estimation safeguards, and uncertainty assessment to distinguish potentially useful mobility measurements from cases where the available evidence is insufficient.

Scientific Validation & Next Steps

Current evidence comes from physics-based simulations and controlled Wi-Fi walking experiments. Broader testing across participants, environments, and everyday activities is needed before considering real-world health monitoring applications.

Why Reliability Matters for Healthy Aging

For future monitoring of older adults, including people living with Alzheimer's disease and related dementias (ADRD), inaccurate or missed mobility measurements could lead to misleading interpretations. Understanding when Wi-Fi sensing works, when it fails, and how uncertainty should be reported is essential before these technologies can support healthcare research and everyday care.

Open Research, Reproducibility & Collaboration

Explore the scientific methods, software, publications, and experimental resources behind WiMoCSI.

Source Code & Development

Explore the WiMoCSI software architecture, simulation tools, CSI analysis methods, and ongoing development.

Scientific Publications

Read research on Wi-Fi motion sensing, walking-speed estimation, numerical validation, and measurement reliability.

Experimental Configurations

Review documented simulation scenarios, antenna geometry, walking references, and acquisition settings.

Methods & Reproducibility

Understand the research workflow, validation procedures, quality checks, and known limitations.

Built for Reproducible Research

WiMoCSI aims to make its methods, experimental assumptions, and validation evidence transparent so researchers can evaluate findings and build on the platform.