The document summarizes an optimization technique used to adjust air pollution emissions rates in an air quality model using data from low-cost air quality sensors. The technique develops an inversion method to automatically adjust emissions inputs to improve model predictions against monitored concentrations. Preliminary tests of the technique in Cambridge, UK optimized NOx emissions rates from 305 road sources against data from 20 low-cost sensors and 5 reference monitors. The optimization reduced errors between modeled and monitored concentrations and adjusted emissions profiles and rates in a physically reasonable manner.