Optimization of Antenna Coupling through Machine Learning for “Smart” TPMS Readers
- Delivery
- Provide download link
- Format
- Price
- Non-members (tax incl.)：￥6,600 Members (tax incl.)：￥5,280
- Paper/Info type
- SAE Paper
No.2021-01-0154
- Pages
- 1-6(Total 6 p)
- Date of publication
- Apr 2021
- Publisher
- SAE International
- Language
- English
- Event
- SAE WCX Digital Summit 2021
Detailed Information
| Author(E) | 1) Saranraj Karuppuswami, 2) C. J. Reddy |
|---|---|
| Affiliation(E) | 1) Altair Engineering, 2) Altair Engineering |
| Abstract(E) | Tire pressure monitoring system (TPMS) is becoming ubiquitous in modern day vehicles with advanced safety and driver assist systems and plays a key role in predictive maintenance. One of the key challenges to realize an efficient TPMS system is to ensure good antenna coupling between the reader antenna in the cabin or on the roof of the vehicle and the antennas in the tires. Understanding the different external factors that affect the antenna coupling is vital to realize an efficient design. Computer aided simulations on antenna coupling is a cost-effective method to reduce the chances of failure before a TPMS is deployed in an actual vehicle. In this work, a computational approach is presented to optimize the antenna coupling and hence the link budget between the reader antennas and the TPMS antennas at 915 MHz. This is achieved by employing machine learning based optimization using commercially available tools, Altair’s HyperStudy and Altair’s Feko. A powerful combination of machine learning technique (regression-based mathematical modelling) to develop a surrogate mathematical model coupled with Global Response Search Method (GRSM) optimization is demonstrated for achieving the goals with very few design iterations A case study is presented that demonstrates the workflow process of optimizing the TPMS antenna coupling using body in white of an automobile. A comparison is also shown between traditional GRSM reader antenna position optimization and optimization coupled with machine learning showcasing significant reduction in computational time and memory. |
