Autonomous Vehicles:
Levels of automation:
Levels of Autonomous Driving:
,What are they? What is the difference between each of these? Advantages and disadvantages for
each level:
• Level 0: No Automation
• Level 1: Driver Assist
o Has driver assistant such as cruise control and adaptive cruise control or automatic
braking
o Requires driver attention
• Level 2: Partial Automation
o Car can do full Steering, acceleration and deceleration
o Advance driving assistance system, the car has full control, but human input may still
be required. Relies on human to monitor the roads and surroundings
• Level 3: Conditional Automation
o Car monitors the road, controls steering, acceleration and deceleration and
monitors surroundings.
o Fallback for self-driving failures
o The car has full control, but human input may still be required.
• Level 4: High Automation
o Geofencing - Full autonomous in a certain area.
o Human can still take control of the car if they want.
• Level 5: Full Automation
o Car might not have a steering wheel; human may not be able to control the car.
Sensors:
How do they work independently? Advantage and disadvantage
Ultrasonic:
• short range used in parking assistance and backup warning system
• Works in all-weather condition.
• Quick reaction time
• Reliable object detection
• Cheap
• Ambient light independent
GPS & INS:
• GPS tracks the position of the car, accurate within several meters.
• INS is used with GPS to improve accuracy. Uses gyroscope and accelerometer to get position
of car.
Radar:
• Radio waves are used to determine distance between the objects and the sensor.
• High precision object detection compared to ultrasonic
• Range – near far
• All weather
• Cheap
• Ambient light independent
, Lidar
• 360-Degree sensor uses light beam to determine the distance between obstacles and the
senor
• Ambient light independent
• Not all weather
• Mid-far range
• Angular resolution high
• Expensive
Cameras
• Cheap to use. Uses complex algorithms to analysis data collected
• Not all weather
• Lane detection
• Ambient light dependent
• Range mid-far
• Angular resolution high
• Cameras have good resolution, is less noisy and small.
• Cameras have bad velocity and don’t work well in all-weather.
• LIDAR okay in resolution
• LIDAR is noisy, bad velocity, bad in all-weather, and is big
• Radar has good velocity, good in all-weather, and is small
• Radar has a bad resolution and is noisy.
Sensor types and how they work together:
Each sensor has different functions. Sensors can work individually for specific features, and they can
also collaborate to work well in other features.
For example, ultrasound is used for parking assistance.
Lidar, radar and camera are used for pedestrian detection, monitor surroundings, collision avoidance
and emergency breaking.
Long range Radar is used for adaptive cruise control.
Kalman filter
Levels of automation:
Levels of Autonomous Driving:
,What are they? What is the difference between each of these? Advantages and disadvantages for
each level:
• Level 0: No Automation
• Level 1: Driver Assist
o Has driver assistant such as cruise control and adaptive cruise control or automatic
braking
o Requires driver attention
• Level 2: Partial Automation
o Car can do full Steering, acceleration and deceleration
o Advance driving assistance system, the car has full control, but human input may still
be required. Relies on human to monitor the roads and surroundings
• Level 3: Conditional Automation
o Car monitors the road, controls steering, acceleration and deceleration and
monitors surroundings.
o Fallback for self-driving failures
o The car has full control, but human input may still be required.
• Level 4: High Automation
o Geofencing - Full autonomous in a certain area.
o Human can still take control of the car if they want.
• Level 5: Full Automation
o Car might not have a steering wheel; human may not be able to control the car.
Sensors:
How do they work independently? Advantage and disadvantage
Ultrasonic:
• short range used in parking assistance and backup warning system
• Works in all-weather condition.
• Quick reaction time
• Reliable object detection
• Cheap
• Ambient light independent
GPS & INS:
• GPS tracks the position of the car, accurate within several meters.
• INS is used with GPS to improve accuracy. Uses gyroscope and accelerometer to get position
of car.
Radar:
• Radio waves are used to determine distance between the objects and the sensor.
• High precision object detection compared to ultrasonic
• Range – near far
• All weather
• Cheap
• Ambient light independent
, Lidar
• 360-Degree sensor uses light beam to determine the distance between obstacles and the
senor
• Ambient light independent
• Not all weather
• Mid-far range
• Angular resolution high
• Expensive
Cameras
• Cheap to use. Uses complex algorithms to analysis data collected
• Not all weather
• Lane detection
• Ambient light dependent
• Range mid-far
• Angular resolution high
• Cameras have good resolution, is less noisy and small.
• Cameras have bad velocity and don’t work well in all-weather.
• LIDAR okay in resolution
• LIDAR is noisy, bad velocity, bad in all-weather, and is big
• Radar has good velocity, good in all-weather, and is small
• Radar has a bad resolution and is noisy.
Sensor types and how they work together:
Each sensor has different functions. Sensors can work individually for specific features, and they can
also collaborate to work well in other features.
For example, ultrasound is used for parking assistance.
Lidar, radar and camera are used for pedestrian detection, monitor surroundings, collision avoidance
and emergency breaking.
Long range Radar is used for adaptive cruise control.
Kalman filter
