1 of 6

Real-Time System Design for Autonomous Systems

Hyunjong Choi, Ph.D.

Department of Computer Science

​

Real-Time Embedded & Control Systems Laboratory

https://rtec.sdsu.edu

2 of 6

Real-Time Requirements

2

Laser

scanners

Radars

Cameras

Pose

estimator

Sensing

Brake

Throttle

Steering

Gear

selector

Actuation

Physical environment

Perception & Planning

≤ Deadline

Real-Time Performance for Intelligent CPS & IoT Applications

  • Just fast is not enough
  • Should be analyzable and predictable
  • Timing constraints imposed by the physical world
    • e.g., self-driving car

3 of 6

Real-Time foundation for intelligent applications

4 of 6

Systems support for intelligent applications�

< CBS-based Linux scheduling >

< Autoware, ROS 2, AUTOSAR >

Theoretical study

Applications

< Analysis framework >

< Effectiveness of weakly-hard >

5 of 6

Systems support for intelligent applications�

  • F1Tenth autonomous racing

​

​

​

​

​

  • Smart IoT monitoring

​

​

​

< Conventional full autonomous software stack >

< E2E stack >

6 of 6

Thank you

Real-Time Embedded & Control Systems Laboratory

https://rtec.sdsu.edu

​

Hyunjong Choi

hyunjong.choi@sdsu.edu