New Research from Pedro Cesar Lopes Gerum

Probabilistic Deep Learning for Traffic Density Prediction

CSU Researcher and Assistant Professor Pedro Cesar Lopes Gerum has created an AI model that accurately predicts traffic congestion even during unexpected disruptions. Unlike traditional tools that offer a single forecast, his MQRNN-monotonic provides a range of possible outcomes and their likelihood, helping cities assess risks and make better decisions. The model excels in normal conditions but also anticipates congestion levels when accidents, severe weather, or special events disrupt usual patterns. This precision could transform how cities manage traffic signals, dynamic tolling, and emergency response. "This research enables smarter, more responsive transportation systems, making traffic smoother, more sustainable, and more predictable for everyone," says Gerum.

Check out these videos about his research: 

Interested in learning more? Review the paper 

The authors involved in the paper contributed as follows: 

• Study conception and design: M. Baykal-Gürsoy, P.C. Lopes Gerum, A. Reed Benton;
• Data collection: M. Baykal-Gürsoy, P.C. Lopes Gerum;
• Analysis and interpretation of results: A. Reed Benton, P.C. Lopes Gerum;
• Draft manuscript preparation: M. Baykal-Gürsoy, P.C. Lopes Gerum, A. Reed Benton.
• All authors reviewed the results and approved the final version of the manuscript.