Papers/Machine learning

High Performance Front Camera ADAS Applications on TI's TDA3X Platform

tomato13 2019. 1. 25. 17:52

https://ieeexplore.ieee.org/iel7/7397295/7397593/07397661.pdf


Please, quotation as below. 


Government safety regulations and independent rating systems such as Euro NCAP are driving wider development and adoption of the ADAS. 

Following key safety aspects are defined by Euro NCAP.

- Automatic emergency braking (AEB)

- Lane Departure Warning (LDW) systems, and

- Speed & impaired driving (SID) etc. 

Depending on the safety aspects, ADAS can be based upon sensor systems such as radar, camera, light detection and range (LIDAR) and ultrasound based systems. Additionally they can integrate and use external information sources such as global positioning systems, car data networks, and vehicle-to vehicle or vehicle to infrastructure communication systems to efficiently and accurately achieve desired goals. 


To sum up, 

TDA3X platform provides mutil-SOC architecture. In other words, we can use many cores like CPU, GPU, DSP and etc. TDA3X provides an ISP libraries by which we can make communication among different SOCs. It explained examples where many algorithms can be used on different cores simultaneously. 


'Papers > Machine learning' 카테고리의 다른 글

AP(Average Precision), mAP(mean Average Precision)  (0) 2019.02.14
How gpu and tpu works  (0) 2019.01.30
Generative adversarial network  (0) 2019.01.14
Threats in deep learning.   (0) 2018.12.29
Transistor principal  (0) 2018.12.26