DENSO TEN Uses MATLAB to Develop Mobile Cyber-Physical System
System Collects Real-World Data to Solve Mobility Problems
Key Outcomes
- DENSO TEN used MATLAB to develop and deploy a production-ready system that spanned hardware and cloud applications without manual recoding
- MATLAB enabled DENSO TEN to easily implement complex algorithms that utilized AI, image processing, probability calculations, and statistics
- MATLAB Production Server allowed DENSO TEN to run sophisticated analytics in a centralized location on the cloud
DENSO TEN is a car electronics manufacturer that builds a wide variety of products from car navigation and sound systems to engine control units for gas, electric, and hybrid vehicles.
As part of the company’s VISION 2030 plan, DENSO TEN is taking steps to develop mobility solutions that harness data using cyber-physical systems. These systems will collect real-world data and analyze it in cyberspace to derive solutions for mobility problems such as traffic congestion.
To achieve this, DENSO TEN developed local and cloud-centralized systems to collect and analyze vehicle data through edge devices. To make this possible, DENSO TEN used Computer Vision Toolbox™ and Deep Learning Toolbox™ to develop algorithms to process image data on the edge devices. Additionally, MATLAB Coder™ and GPU Coder™ were used to generate codes and deploy them on the edge devices. Finally, MATLAB Production Server™ helped to optimize the cloud system performance by aggregating data and generating results that can be visualized and acted upon. All this was performed from a single, consistent MATLAB® codebase without manually recoding into other languages for deployment.
A portion of the MATLAB code that performs image preprocessing can also be generated into C/C++ so that it can be integrated into a low-power edge device running in the vehicle. The more compute-intensive MATLAB algorithms run on MATLAB Production Server.