Multi-Way Smart Traffic Control

Problem

Adama's population is projected to more than triple by 2040, and its fixed-timer traffic signals already can't cope. The conventional system allocates signal time on presets rather than real conditions, gives no priority to ambulances or fire trucks, doesn't enforce traffic rules, and largely ignores pedestrian safety — pushing congestion from one intersection to the next.

Approach

A multi-way smart traffic control system that senses real conditions and adapts:

• Detection: YOLO-based vehicle detection and classification from CCTV feeds, combined with RFID tags for fast, reliable vehicle identification at intersections.
• Smart Time Allocation (STAS): fuzzy-logic signal timing driven by live vehicle counts, congestion, and peak-hour context.
• Special Vehicle Identification: automatic recognition and prioritization of emergency/public-service vehicles for unobstructed passage.
• AI decision-making: reinforcement learning trained on historical data and simulation to coordinate signals across intersections.
• Supporting subsystems: pedestrian signaling, a dual-mode (mobile + roadside) driver notification system, redundant power, and real-time data collection feeding the model.

Intersection phasing was modeled and evaluated in AnyLogic before hardware decisions.

Results

• An integrated, simulation-validated design coordinating detection, prioritization, and adaptive timing across multiple intersections.
• A staged rollout plan — baseline signals → pedestrian safety → CCTV → STAS → emergency prioritization → RL decisioning — with a citywide budget study for Adama.
• Emergency-vehicle prioritization aimed at cutting response times at busy junctions.

Learnings

• Combining RFID (identification) with vision (classification) is more robust than either alone.
• Fuzzy logic is a practical bridge to adaptive control before committing to full RL.
• Simulation-first design de-risks expensive physical infrastructure decisions.