<![CDATA[The modern EFI system has become the backbone of engine performance control, yet diagnosing faults within it remains a challenge when no Diagnostic Trouble Code (DTC) has been triggered. This study investigates the diagnostic potential of oxygen (O₂) sensor waveforms and Short Term Fuel Trim (STFT) data streams acquired via the OBD-II protocol on a Toyota Innova 1TR-FE. Six experimental conditions were tested—normal idle, normal at approximately 2,500 rpm, injector fault at idle, injector fault at 2,500 rpm, vacuum leak at idle, and vacuum leak at 2,500 rpm—using Toyota Global TechStream (GTS+) as the data acquisition interface. The analysis employed a residual generation approach combined with rule-based decision-level data fusion to distinguish between the three operating states. Results reveal that vacuum leaks produce a distinctive STFT elevation of up to +19.70%, while injector faults cause O₂ sensor voltage to fall to extremely low levels (0.011–0.042 V) with a complete cessation of switching activity (0 Hz). Integrating both parameters through decision-level data fusion enabled unambiguous differentiation among normal, vacuum leak, and injector fault conditions across all tested scenarios. The proposed method achieved a classification accuracy of 100%, surpassing the minimum target of 85%. These findings suggest that combined O₂ sensor and STFT data streams offer a practical and effective foundation for early-stage OBD-II-based fault diagnosis in EFI vehicles.]]>
