Inspiration
3D mapping with LiDAR
What it does
How we built it
Standard Wi-Fi signal strength (RSSI) is far too noisy to detect humans through walls. Instead, we deployed two ESP32 microcontrollers running Espressif’s esp-csi framework over ESP-NOW on 2.4 GHz Channel 11. Node A acts as a continuous 100 Hz packet transmitter, while Node B acts as the receiver, extracting raw Channel State Information (CSI)—the exact Imaginary ($I$) and Real ($Q$) phase and amplitude components across 50+ Orthogonal Frequency-Division Multiplexing (OFDM) subcarriers—and streaming them over USB Serial.
Challenges we ran into
Running custom Firmware on the ESP32 was challenging, especially getting the ESP32 to convey reliable data over serial.
Accomplishments that we're proud of
We are very proud of making something that we initially thought of as beyond our capabilities.
What we learned
We learned how to run custom firmware on microcontrollers.
What's next for Marco-Polo
Training a lightweight neural network to better classify and use CSI data.
Log in or sign up for Devpost to join the conversation.