Intelligent transportation systems (ITS) rely on the real-time detection of traffic signs to support both autonomous vehicles and driver-assistance systems. Unfortunately, due to limited computing power, memory, and environmental conditions (such as low light, glare, rain, and fog), developing an accurate detection system that runs on edge devices remains highly challenging. To address these issues, this research presents an environment-aware framework that leverages a YOLOv8 model (a version of YOLO optimized for edge devices) combined with adaptive preprocessing; therefore, it can adjust, in real time, both preprocessing functions and detection parameters using information from environmental conditions to ensure accurate detection of traffic signs on CPU-only devices. Additionally, the pre-trained weights of the YOLOv8 model are adjusted based on a carefully curated dataset, and the resulting model can be quantized for use with ONNX. The proposed system achieves a mean Average Precision (mAP) of 91.0%, a precision of 96.1%, and a recall of 92.3% at up to 22 frames per second on low-resource platforms. This work contributes an environment-aware framework that fuses adaptive environmental condition analysis with YOLOv8 to achieve accurate traffic sign detection in real time.