top of page


Mr. John Smith

Job title



Brake disc and pad temperatures are an important element in determining the clamping force needed to stop or hold a vehicle from moving or rolling off the hill. Accurate disc and pad temperature estimation is key to parking brake system. Further, real-time accurate estimation of the brake discs temperature is vital to finding the precise clamping force needed to stop an autonomous vehicle, which needs to decelerate or stop by itself. Installing physical sensors is expensive for mass production units and impacts the ecological footprint. To deal with these, Brake Thermal Models (BTMs) which estimate the discs/pads temperature using physical models have been in development and used for several years. Due to very high dimensionality of the brake disc/pad heating and cooling phenomenon, these thermal models fail to achieve a satisfactory accuracy level. BTMs tend to accumulate errors over time, leading to an enlarged error gap during driving. Brakes are unarguably one of the most important safety systems of a vehicle, and hence there is need for a more accurate model. Hitachi Astemo Brakes Systems in collaboration with Hitachi Europe Corporate R&D has developed a very robust and accurate virtual brake disc/pad temperature estimation model based on readily available vehicle data. We expanded the vehicle data using mathematical relations. Our solution is driven by data, and the model is driven by Machine Learning and Artificial Intelligence (AI). The model is trained on real driving data in real conditions, and the trained model is also tested on real data in varying real driving conditions. The robustness of the model comes from training and combining several machine learning / AI models such that the models support one another in making the final temperature estimation (the hybrid approach). The model error does not accumulate over time and easily recovers from erroneous previous prediction(s). The developed model by Hitachi requires minimal effort and results show that it outperforms a given BTM.

Hitachi Astemo: Dr. Anthony Ohazulike, Dr. Abdessamed Ramdane

Robust disc and pad temperature estimation model – a machine learning / artificial intelligence approach

EB2022-FBR-006 • Full • EuroBrake 2022 • Thermal simulation


Sign up or login to the ICC to download this item and access the entire FISITA library.

Upgrade your ICC subscription to access all Library items.