Downloads: 17 | Views: 234 | Weekly Hits: ⮙3 | Monthly Hits: ⮙3

Prototypes and Models | Engineering Applications of Artificial Intelligence | India | Volume 12 Issue 7, July 2023

Smart Cabin Application on High Performance Computing Machine - AC Personalisation

Gowrishankar S ^[3] | Kuldeep Kumar ^[2] | Sreevishakh KP | Muralidhara K V | Shruthe R

Abstract: Heating Ventilation and Air Conditioning (HVAC) systems have been used for many years to create comfortable temperature environments in enclosed spaces such as residential, industrial, space, and automotive industry. Vehicle HVAC systems work to keep passengers at a comfortable temperature. However, a variety of environmental elements have an impact on thermal comfort, and a person's thermal preferences can vary substantially due to physiological, behavioral, and cultural factors. In this paper we came up with AC personalization application prototype on high performance computing along with machine learning support to predict preferred ac temperature values based on previous learning data sets and three real time input values. These three inputs include ambient temperature (cabin temperature), heart rate and body temperature of the driver. A combination of classic and adaptive Autosar hardware platforms are used for the prototype. A Service-Oriented Real-Time Communication (SOA) scheme over SOME/ip is used for communication between classic and adaptive nodes also we used ANDi tool for showcasing the outputs rather than HVAC module. We used a combination of three machine learning algorithms-Decision Tree Regressor, Polynomial Regressor, Adaboost Regressor for ac temperature prediction model.

Keywords: Adaptive & Classic Autosar, SOME/IP, Service Oriented Architecture (SOA), Machine learning, RCar-H3e, AURIX? TC37x

Edition: Volume 12 Issue 7, July 2023,

Pages: 916 - 921