Akademik Veri Yönetim Sistemi
2025 10th International Conference on Computer Science and Engineering (UBMK), İstanbul, Türkiye, 17 Eylül 2025, (Tam Metin Bildiri)
Pedestrian safety remains a critical concern in modern traffic systems. This paper presents an Internet of Things (IoT)–based surveillance system that integrates an Espressif Systems ESP32 microcontroller, an HC-SR04 ultra-sonic distance module, artificial-intelligence (AI) voice interac-tion, and real-time vehicle-to-vehicle communication to enhance pedestrian protection. The system detects pedestrians within 200 cm with 96.4 % accuracy and 134 ms average latency, reducing false positives by 24 % through signal conditioning. Upon detection, it triggers (i) spoken alerts generated locally by a large language model (LLM, Ollama) and a Google text-to-speech (TTS) engine (3.1 s response latency) and (ii) wireless warnings to nearby vehicles (180 ms average transmission delay). Experiments confirmed 94.7 % hot-word activation in quiet environments and zero packet loss for inter-vehicle messages up to 20 m. The proposed framework demonstrates a cost-effective approach to real-time pedestrian safety using embedded hardware and AI.