Open-Source Platform RuView Beta Released: Detecting People Through Walls Using WiFi Signals
RuView is a spatial sensing platform that uses ESP32 sensors to detect human presence and vital signs via WiFi signals, all without cameras, ensuring privacy.
RuView Beta: Transforming WiFi Signals into “Eyes”
A new technology has emerged that detects human movement on the other side of a wall without using cameras. The open-source project “RuView” leverages existing WiFi signals to create a spatial sensing platform. Far from being a mere research prototype, RuView is implemented as a low-cost edge computing system using ESP32 microcontrollers.
Capabilities of RuView
RuView detects subtle disruptions in WiFi signals caused by human movements. The key to this functionality lies in the data known as “Channel State Information” (CSI). ESP32 sensors capture this data and analyze it using AI models.
Its primary features include:
- Presence and Occupancy Detection: Detecting human presence through walls, tracking movement, and counting people entering or exiting spaces.
- Monitoring Vital Signs: Measuring respiration and heart rate non-invasively, even while sleeping or sitting still.
- Activity Recognition: Recognizing activities such as walking, sitting, gestures, or falls using time-series CSI patterns.
- Environmental Mapping: Identifying rooms, tracking furniture movement, and detecting new objects using RF fingerprinting.
One of the standout aspects of RuView is its camera-free design, making it suitable for environments requiring privacy, as well as dark or obstructed settings.
Technical Mechanism and Current Limitations
RuView operates primarily on edge computing. Each ESP32 node costs approximately $9 and does not require cloud connectivity or internet access. The system adapts locally to its environment and uses spiking neural networks for quick learning within 30 seconds.
Currently in beta, the project is under active development. Documentation highlights several limitations:
Firstly, ESP32-C3 and original ESP32 models are not supported due to their single-core design, which is insufficient for CSI digital signal processing (DSP). Additionally, using a single ESP32 node limits spatial resolution; using two or more nodes, or an additional module called “Cognitum Seed,” is recommended.
Accuracy is also an area for improvement. The current camera-less pose estimation precision (PCK@20) is approximately 2.5%. However, when trained on ground truth data captured via cameras, the pipeline targets over 35% accuracy, with data collection and evaluation phases currently underway.
Potential of Open-Source Development
RuView is available on GitHub, welcoming contributions and bug reports. Its foundational technology stems from Carnegie Mellon University’s “DensePose From WiFi” research. Pose estimation outputs 17 COCO key points and trains solely on 10 sensor signals without cameras.
The project highlights the possibility of repurposing existing wireless infrastructure for innovative sensing applications. Clever designs, such as multi-frequency mesh scanning that utilizes nearby routers as “free radar illuminators,” showcase its ingenuity. All measurements are cryptographically secured using Ed25519 proof chains.
RuView could have future applications in smart homes, healthcare, security, and other fields. However, ethical discussions surrounding privacy are inevitable. Deploying technology capable of detecting human movements and vital signs through walls requires careful societal dialogue.
Frequently Asked Questions
- Can RuView be used with standard home WiFi routers alone?
- No, RuView requires dedicated ESP32 sensor nodes. While it utilizes existing WiFi signals, ESP32 sensors must be installed as receivers to obtain CSI data. Modifications to the router itself are unnecessary.
- Is the camera-free accuracy practical?
- The current beta version has limited camera-free pose estimation accuracy. However, the project has implemented a pipeline to improve accuracy by training on camera-collected data. Initial applications are likely to focus on presence detection and vital sign monitoring.
- Are there privacy concerns with RuView?
- While RuView does not record images, making it less intrusive than cameras, it can still detect individual behaviors and vital signs. Data handling requires careful consideration. The system's edge-based design, which avoids cloud data transmission, helps mitigate privacy risks.
Comments