Real-Time Adaptive Traffic Signal System

Abstract

Signalised intersections that rely on timetable-bound phase rotation are inherently mismatched to the stochastic nature of vehicular demand, yielding persistent queue inequity and avoidable fuel expenditure. To rectify this mismatch without incurring the infrastructure overhead of vision-centric or cloud-dependent solutions, the present study engineers a wholly self-contained, microcontroller-resident intersection governor. A trio of HC-SR04 acoustic ranging modules furnish per-lane gap telemetry at 500 ms cadence; firmware resident on an Arduino Uno (ATmega328P) translates these measurements into occupancy tiers and computes phase windows whose magnitudes scale with measured lane saturation, thereby channelling intersection capacity toward genuinely congested approaches while suppressing idle green allocation on vacant lanes. An EM-18 reader interrogating a 125 kHz radio-frequency band monitors the intersection perimeter concurrently; detection of an authorised credential mounted on an emergency vehicle triggers an unconditional phase override within 115 ms, with deterministic restoration of the interrupted adaptive cycle upon credential withdrawal. A character-mapped LCD broadcasts operational state continuously, eliminating the requirement for external diagnostic tooling during commissioning. Six structured bench scenarios — spanning solo, dual, and trilateral lane loading, tag-entry pre-emption, tag-exit resumption, and zero-occupancy cycling — returned unanimous pass verdicts, validating both functional subsystems without exception. The resulting platform is cost-accessible, field-configurable, and architecturally open to incremental capability extension.