Smart Assistive Traffic System with Haptic Wearable for the Visually and Hearing Impaired

ESP-NOW Wireless Haptic Communication, FFT Siren Detection, and Road Traffic Management

Category: IoT • Assistive Technology • Digital Signal Processing • Embedded Systems
Tools & Technologies: 2× ESP32, ESP-NOW Wireless Protocol, MAX4466 High-Sensitivity Microphone, FFT (Fast Fourier Transform), Haptic Vibration Motor, 20×4 I2C LCD Display (PCF8574 backpack), LM2596 Buck Converter, TP4056 USB-C Charger, HWE 18650 E6 Battery, Arduino IDE

Status: Completed

Project Overview

This system bridges the accessibility gap at road crossings for pedestrians with visual or hearing impairments. It consists of two nodes: a Road Unit (infrastructure) and a Hangable Wearable Unit (assistive device worn at the belt or pocket).

The Road Unit manages a traffic light cycle (90 s vehicle phase / 30 s pedestrian phase) and monitors for emergency vehicle sirens using FFT spectral analysis on the MAX4466 microphone, detecting the characteristic frequency sweep of ambulances and fire trucks (400–2500 Hz). When a crossing state changes or a siren is detected, the road unit broadcasts the event wirelessly via ESP-NOW to the wearable.

The Wearable Unit translates road data into a distinct haptic language of vibration patterns: a 10-second continuous vibration when vehicles are moving (danger), a rhythmic triple-burst pattern during safe pedestrian crossing, and rapid 100 ms pulses during emergency override. A 2-second heartbeat ping from the wearable back to the road unit creates a connection-loss safety alert if the pedestrian moves out of range.

Live demonstration: The completed system was tested and demonstrated with both units operating in real time. The road model physical prop, built from PVC board and vinyl materials, was used to present the system at demonstration. All three haptic patterns and the FFT-based siren detection were verified during testing.

Key Features

  • ESP-NOW peer-to-peer wireless: no Wi-Fi router, no internet dependency, ultra-low latency communication between the two ESP32 units
  • FFT-based intelligent siren detection: MAX4466 microphone feeds audio into FFT spectral analysis; ambulance and fire truck frequency sweep (400–2500 Hz) detected in real time
  • 3-state haptic language: Danger (continuous 10 s vibration during vehicle phase), Walk (triple-burst pattern during pedestrian phase), Emergency (rapid 100 ms pulses on siren detection)
  • Bidirectional heartbeat: wearable pings road unit every 2 s; connection loss triggers a safety alert on both devices simultaneously
  • Three intelligent emergency scenarios: siren during vehicle phase, siren during pedestrian phase, and a full recovery protocol that resets to a 90 s vehicle phase after the siren clears
  • 20×4 I2C LCD on road unit: displays live phase name, countdown timer, and wearable connection status; driven via PCF8574 I2C backpack
  • Blue LED connectivity indicator: fitted on both units to show whether the ESP-NOW link is active
  • Self-contained power systems: road unit powered from a 12 V AC adapter regulated to 5 V via LM2596 buck converter; wearable runs on a 3200 mAh 18650 LiPo cell with USB-C charging via TP4056

System Architecture

The system operates as two cooperating ESP32 nodes communicating exclusively over ESP-NOW with no infrastructure requirement. The road unit owns the traffic state machine and all sensing; the wearable unit owns the haptic output and the user-facing experience.

Road Unit (ESP32 #1)

12 V AC Adapter → LM2596 Buck Converter → 5.16 V regulated rail MAX4466 Microphone → ADC sampling FFT Analysis: 400–2500 Hz siren sweep detection Traffic Cycle State Machine
90 s VEHICLE → 30 s PEDESTRIAN
Emergency Override: siren detected → all-red + alert broadcast 20×4 I2C LCD (PCF8574): phase + countdown + wearable status ESP-NOW TX → phase event + siren flag broadcast

Wearable Unit (ESP32 #2)

TP4056 USB-C → 18650 3200 mAh LiPo → ESP32 Vin ESP-NOW RX: phase event received from road unit PEDESTRIAN phase → triple-burst haptic pattern
(100 ms on ×3 + 1.5 s pause, repeating)
VEHICLE phase → 10 s continuous vibration (danger alert) EMERGENCY → rapid 100 ms pulse bursts (siren active) Blue LED: connection status indicator Heartbeat ping → road unit every 2 s (link-loss detection)

Hardware Components

Component Model / Specification Unit Role
Microcontroller (Road) ESP32 Dev Board Road Unit Traffic state machine, FFT siren detection, ESP-NOW TX, LCD driver
Microcontroller (Wearable) ESP32 Dev Board Wearable Unit ESP-NOW RX, haptic motor control, heartbeat ping, LED indicator
Microphone Module GY MAX4466 High-Sensitivity Electret Microphone Amplifier Road Unit Audio capture for FFT-based siren frequency analysis
Haptic Motor Coin-type vibration motor Wearable Unit Converts road phase events into tactile vibration patterns
LCD Display 20×4 Character I2C LCD with PCF8574 backpack Road Unit Displays phase name, countdown timer, and wearable connection status
Traffic LEDs Red, amber, green LEDs Road Unit Physical traffic signal indicators for the road model demo
Status LEDs Red, green, blue LEDs Both Units Blue LED shows ESP-NOW link status; red/green indicate phase state
LiPo Battery HWE 18650 E6, 3200 mAh, 3.7 V, 11.84 Wh, with JST connector Wearable Unit Portable power source for the wearable unit
Charging Module TP4056 USB-C LiPo charger module Wearable Unit Manages LiPo charging with overcharge and over-discharge protection
Buck Converter LM2596 adjustable DC-DC step-down converter (set to 5.16 V output) Road Unit Regulates 12 V AC adapter output to stable 5 V logic rail
AC Adapter BT-branded 12 V / 2 A switching AC adapter Road Unit Mains power source for the road unit infrastructure installation
Mains Plug UK BS1363 3-pin plug (3 A, 250 V) Road Unit Connects AC adapter to mains socket
Enclosures Two white PVC electrical junction box cover plates (BS 4007) Both Units Houses all electronics for road unit and wearable unit respectively

Circuit Diagram

Full Proteus circuit diagram showing both units: the wearable ESP32 with vibration motors, TP4056 charger, 18650 battery, and boost converter; and the road unit ESP32 with MAX4466 microphone, 20x4 LCD, traffic and status LEDs, and 220V AC to 5V DC LM7805 power supply.


Haptic Language Design

Rather than a simple on/off vibration, the wearable communicates three distinct states through carefully designed vibration patterns that a visually or hearing-impaired user can learn and interpret without any audio or visual feedback.

VEHICLE PHASE
Pattern: 10-second continuous vibration on crossing phase start.

Meaning: Vehicles are moving. Do not cross. Stay on the pavement.

Duration: 90-second vehicle phase; vibration fires at phase entry.
PEDESTRIAN PHASE
Pattern: Triple-burst: 100 ms vibration ×3 with 100 ms gaps, then 1.5 s silence. Repeats for the full pedestrian phase.

Meaning: Safe to cross now.

Duration: 30-second pedestrian phase; pattern cycles continuously.
EMERGENCY OVERRIDE
Pattern: Rapid 100 ms pulse bursts, continuous, faster than the walk pattern.

Meaning: Emergency vehicle detected. All lights red. Clear the road immediately.

Trigger: FFT detects siren sweep (400–2500 Hz) on road unit microphone.

Process and Methodology

The build spanned six weeks across two parallel tracks: physical road model construction for the demonstration prop, and electronics hardware design, fabrication, and firmware development for both ESP32 units. All work was documented in photographs from the first material cut through to the final live demonstration.

Road Model: Material Preparation, April 2026

Road Model: Surface Construction, April 2026

Electronics Components, April 2026

Enclosure Fabrication, April 2026

System Assembly, April 2026

Completed System and Live Demonstration, May 2026


Challenges & Solutions

  • Challenge: Distinguishing genuine emergency sirens from ambient road noise (vehicle engines, wind, music) on the microphone input.
    Solution: Applied FFT spectral analysis to isolate the 400–2500 Hz frequency range characteristic of ambulance and fire truck sirens. Threshold tuning reduced false positives from broadband noise.
  • Challenge: Designing a haptic vocabulary that is intuitive without training and distinct enough to avoid confusion between states.
    Solution: Three patterns with clearly different rhythms: a long continuous burst for danger (unmistakably "stop"), a rhythmic triple-burst for safe crossing, and a fast urgent pulse for emergency. Rhythm length and density are the differentiators.
  • Challenge: Detecting when the wearable user has moved out of ESP-NOW range without a reliable disconnect event from the protocol.
    Solution: Implemented a 2-second heartbeat ping from wearable to road unit. If the road unit misses heartbeats beyond a timeout threshold, it flags the wearable as disconnected and updates the LCD status field accordingly.
  • Challenge: Supplying stable 5 V to the road unit ESP32 and LCD from the 12 V AC adapter without excessive heat dissipation from a linear regulator.
    Solution: Used an LM2596 adjustable buck converter (switching regulator) trimmed to 5.16 V output. Verified with a multimeter before connecting the ESP32 to avoid overvoltage damage.

Results & Impact

  • Functional haptic communication: All three vibration patterns (danger, walk, emergency) were confirmed working during live testing and demonstration.
  • Reliable wireless link: ESP-NOW peer-to-peer communication operated without a router or internet connection, demonstrating the system's infrastructure-independence.
  • Accurate siren detection: FFT-based analysis on the MAX4466 microphone detected the simulated emergency frequency sweep and triggered the emergency override correctly in tests.
  • Heartbeat safety mechanism: Connection-loss detection via the 2-second heartbeat ping successfully flagged when the wearable moved beyond ESP-NOW range.
  • Accessibility impact: The system demonstrates a low-cost, infrastructure- independent solution for giving visually and hearing-impaired pedestrians real-time road crossing information without relying on audible signals or visual displays.

Project Links


Thank You for Visiting My Portfolio

I sincerely appreciate you taking the time to explore my portfolio and learn about my work and expertise. If you have any questions or wish to discuss potential collaborations, please feel free to reach out via the Contact section.

Best regards,
Damilare Lekan, Adekeye.