Overview on FPGA Design Implementations for Autonomous Cars

Session

Computer Science and Communication Engineering

Description

Autonomous cars are top representative product of cyber-physical systems, Industry 4.0 and 5G communications. Besides vehicle producer companies (Audi, BMW, etc.), also semiconductor device producers (Intel, NVIDIA, Xilinx, Qualcomm), ICT companies (Google, Nokia, etc.) and Computer-aided design (CAD) tool provider companies (Mentor Graphics) invest in development of devices for autonomous cars. They are on a search for the most appropriate hardware platform and device for autonomous vehicle control system development. Several contests were organized recently for design of control systems for autonomous cars with different hardware platforms. Complex and time-consuming algorithms for detection, tracking and localization, based on deep neural networks, evolutionary algorithms for artificial intelligence (AI) and machine learning are embedded in autonomous cars. There are very strong constraints for frame-rate and processing latency for reacting in real time, based on data collected from sensors and this, with high safety. In the same time factors as power and fuel consumption, thermal effects and system cooling, storage, electromagnetic compatibility and vibration effects have to be considered when selecting the hardware solution. Most of the autonomous car processing and control systems are based on multicore CPUs (in some cases on controllers and/or System-on-Chips - SoCs), working in parallel with accelerators. Field-Programmable Gate Array (FPGA) is one of the candidates for accelerator device, together with Graphics Processing Unit (GPU) and Application Specific Integrated Circuits (ASICs). The paper makes an overview of the existing applications of FPGAs for design of control systems in autonomous cars and analyses the pros and cons for their role in this sector.

Keywords:

Autonomous cars, FPGA, GPU, AI accelerators

Session Chair

Bertan Karahoda

Session Co-Chair

Krenare Pireva

Proceedings Editor

Edmond Hajrizi

ISBN

978-9951-437-69-1

Location

Pristina, Kosovo

Start Date

27-10-2018 10:45 AM

End Date

27-10-2018 12:15 PM

DOI

10.33107/ubt-ic.2018.100

This document is currently not available here.

Share

COinS
 
Oct 27th, 10:45 AM Oct 27th, 12:15 PM

Overview on FPGA Design Implementations for Autonomous Cars

Pristina, Kosovo

Autonomous cars are top representative product of cyber-physical systems, Industry 4.0 and 5G communications. Besides vehicle producer companies (Audi, BMW, etc.), also semiconductor device producers (Intel, NVIDIA, Xilinx, Qualcomm), ICT companies (Google, Nokia, etc.) and Computer-aided design (CAD) tool provider companies (Mentor Graphics) invest in development of devices for autonomous cars. They are on a search for the most appropriate hardware platform and device for autonomous vehicle control system development. Several contests were organized recently for design of control systems for autonomous cars with different hardware platforms. Complex and time-consuming algorithms for detection, tracking and localization, based on deep neural networks, evolutionary algorithms for artificial intelligence (AI) and machine learning are embedded in autonomous cars. There are very strong constraints for frame-rate and processing latency for reacting in real time, based on data collected from sensors and this, with high safety. In the same time factors as power and fuel consumption, thermal effects and system cooling, storage, electromagnetic compatibility and vibration effects have to be considered when selecting the hardware solution. Most of the autonomous car processing and control systems are based on multicore CPUs (in some cases on controllers and/or System-on-Chips - SoCs), working in parallel with accelerators. Field-Programmable Gate Array (FPGA) is one of the candidates for accelerator device, together with Graphics Processing Unit (GPU) and Application Specific Integrated Circuits (ASICs). The paper makes an overview of the existing applications of FPGAs for design of control systems in autonomous cars and analyses the pros and cons for their role in this sector.