NH3 slip identification and NH3-Induced bias correction in remote NOx emissions monitoring of heavy-duty diesel vehicles

Remote in-use emissions monitoring of heavy-duty diesel vehicles (HDDVs) is increasingly adopted to strengthen air-quality governance and ensure real-world compliance with nitrogen oxides (NOx) limits. A persistent challenge is the severe cross-sensitivity of electrochemical NOx sensors to ammonia (NH 3 ) slip from aftertreatment systems. This interference inflates apparent NOx emissions, triggers false exceedances, and undermines the credibility of fleet-scale monitoring data. Here, a telematics-integrated dual-algorithm framework is proposed to detect NH 3 slip events and correct the NOx measurement bias resulting from NH 3 -induced cross-sensitivity using only on-board signals, enabling scalable deployment without hardware modification. NH 3 slip is first identified using a moving-window NH 3 excess index ( EI NH3 ) combined with SCR efficiency thresholds to ensure robust event discrimination under transient driving. Cross-sensitivity artifacts are then corrected by constraining the effective selective catalytic reduction conversion to 99% during slip conditions and applying state compensation derived from Arrhenius-type NH 3 storage kinetics. The framework is validated on multiple HDDVs over real-driving emission (RDE) cycles using portable emissions measurement systems (PEMS) and a laser spectroscopic NH 3 analyzer as independent references. Results show that motorway high-speed operation exacerbates NH 3 slip under elevated space velocity and exhaust temperature. Across RDE tests, the identification module achieves 77-97% slip event recall and >93% classification accuracy, while the correction reduces the mean error of the 90th-percentile specific NOx emission ( SE NOx _P 90) by 94% (0.50 to 0.03 g/kWh), effectively eliminating false exceedances attributable to NH 3 interference. In multi-vehicle compliance screening, several vehicles that would have been falsely flagged as non-compliant based on raw remote NOx data were reclassified as compliant after correction, with their estimated emissions falling below the 0.69 g/kWh regulatory limit, reducing false non-compliance determinations and improving the precision of high-emitter targeting. By enabling scalable and trustworthy NOx quantification, the proposed framework enhances the credibility and cost-effectiveness of telematics-based oversight. It supports cleaner freight operations through more reliable, data-driven emissions governance under real-world driving conditions.