MAINCE PREDICT
Proven AI-based road degradation models
Superior road condition predictability with deep learning
Maince Predict uses deep learning to model how road conditions will change over time. It can forecast degradation based on real-world datasets across various road types — outperforming traditional methods in accuracy and adaptability.
From reactive.
To proactive.
Asset managers and engineers rely on road degradation forecasts for dependable, multi-year maintenance planning. But existing models fall short.
Maince Predict uses AI models that have been trained on data from thousands of km’s of roads of different types and hierarchies. It can predict the degradation of pavement based on survey data from visual inspections, road scans and skid resistance tests.
Maince Predict key features
Automated data preparation and validation
With Maince Predict you can now truly leverage your data to forecast road degradation. It allows for simple, secure and scalable uploading of all your survey data, location-based data, network inventory, and maintenance completed. All input data is validated and stored for further use in Maince Predict or elsewhere.
Predict degradation across survey types
Maince Predict uses state-of-the-art AI models to help you anticipate how and when individual road sections will deteriorate. The models are trained across diverse survey types, road types, geographies and maintenance strategies. Simply choose the number of years and which part of your network you want to forecast for, and which data and models to use.
Understand network & treatment behaviour
All predictions in Maince Predict are made available for further use in Maince Plan or third-party systems. They will improve your understanding of the way your roads behave over time, help you identify future intervention windows and move you from reactive to proactive maintenance. The solution will help you understand how different treatments and materials behave, so you can make smarter, evidence-based maintenance decisions.
“ Maince Predict’s machine learning models set a new benchmark in predicting road degradation — a breakthrough with the potential to transform maintenance planning worldwide. ”
James Stokes, PhD, Jean Lefebvre (UK)
What makes Maince Predict unique?
Proven, deployed ML models for degradation forecasting
Predicts for different road types & geographies
Performs using data from multiple survey types
Outperforms legacy models by more than 20%
Predicts remaining service life
Predictions can be mapped geographically year-by-year
Seamless integration with Maince Monitor & Plan