This pre-study project was aimed at application of artificial intelligence (AI), in the analysis of sensors’ output data for the measurement of the diameter distribution, of clouds of airborne nanoparticles in the exhaust gas in the tailpipe of a vehicle in order to provide a technique for validating compliance. The end objective of the AI algorithm in this project is to provide information on the number of nanoparticles in several diameter groups in a sample containing roughly 100-10000 of particles per cm3, based on an optical scattering sensor. Sensor operation is based on the angular distribution of optical scattering spectrum of a cloud of soot particles. The approach proposed here uses the transition from Rayleigh to the Mie regime using scattering spectrum at several angles as essential information to analyze the particle number distribution in a cloud of particles. Processing such a massive amount of data using conventional algorithms limits the dynamic response of the sensor.
The sensor architecture is currently being investigated in our group (EMSL, MC2, Chalmers). The final measurement result is the histogram of the distribution of particle counts in the clusters within a measurement time limited to about one minute. In this pre-study, we have investigated the machine learning methods to obtain efficient and fast data processing algorithm.
Researcher at Chalmers, Microtechnology and Nanoscience (MC2), Electronics Material and Systems Laboratory
Full Professor at Chalmers, Microtechnology and Nanoscience (MC2), Electronics Material and Systems Laboratory
Associate Professor at Chalmers, Microtechnology and Nanoscience (MC2), Electronics Material and Systems Laboratory
Funding Chalmers participation during 2018
Areas of Advance