[자율주행 스터디] Introduction to Self-Driving cars - Week 2-1

Week 2 - Self-Driving Hardware and Software Architectures

Course info.

• https://www.coursera.org/learn/intro-self-driving-cars?
  • UNIVERSITY OF TORONTO

---

Lesson 1 : Sensors and Computing Hardware

Contents

• Sensor types and characteristics
• Self-driving computing hardware

---

Sensors

• Sensor: device that measures or detects a property of the environment, or changes to a property
• Categorization
  • Exteroceptive (extero = surroundings) (주위 환경)
  • Proprioceptive (proprio = internal) (자기 자신)

---

Sensors for perception

• Exteroceptive (외부 관측을 위한 센서)

1. Camera (카메라)

   • Essential for correctly perceiving environment
   • Comparison metrics
     • Resolution
     • Field of view (horizontal and vertical angular)
     • focal length, depth of field and frame rate
     • Dynamic range
   • Trade-off between resolution and FOV?
     • FOV가 커지면 동일한 품질을 얻기 위해서는 해상도(resolution)가 높아져야한다.
   • Stereo Camera (카메라 두 대를 이용하는 방법)
     • Enables depth estimation from image data

2. LIDAR (라이다)

   • Detailed 3D scene geometry from LIDAR point cloud
   • Comparison metrics
     • Number of beams
     • Points per second
     • Rotation rate (Detection range)
     • Field of view
   • Upcoming : Sold state LIDAR

3. RADAR (레이더)

   • Robust object detection and relative speed estimation
   • Comparison metrics
     • Range
     • Field of view
     • Position and speed accuracy
   • Configurations
     • Wide FOV, short range
     • Narrow FOV, long range

4. Ultrasonic (초음파센서)

   • Short-range all-weather distance measurement
   • Ideal for low-cost parking solutions
     • using in parking assistance
   • Unaffected by lighting, precipitation(비, rain)
     • (역광과 같은 빛의 변화와 관련된 효과에 영향을 받지 않음)
   • Comparison metrics
     • Range
     • Field of view
     • Cost

• Proprioceptive (자차 정보 취득을 위한 센서)

1. Global Navigation Satellite Systems and Inertial Measurement Units (GNSS/IMU)

   • Direct measure of ego vehicle states
     • position, velocity (GNSS)
       • Varying accruacies : RTK, PPP, DGPS
     • Angular rotation rate (IMU)
     • Acceleration (IMU)
     • Heading (IMU, GNSS)

2. Wheel Odometry (휠 센서)

   • Tracks wheel velocities and orientation
   • Uses these to calculate overall speed and orientation of car
     • Speed accuracy
     • Position drift

---

Computing Hardware

• Need a "Self-driving brain"
  • Takes in all sensor data
  • Computes actions
  • Already existing advanced systems that do self driving car processing
    • e.g. DRIVE PX/AGX, Intel & Mobileye EyeQ
  • Image processing, Object detection, Mapping (Parallel computation is needed)
    • GPUs - Graphic processing unit
    • FPGAs - Field Programmable Gate Array
    • ASICs - Application Specific Integrated Chip
  • Synchronization Hardware
    • To synchronize different modules and provide a common clock