Jitesh Gosar Jitesh Gosar
Menu

Autonomous Driving: Concepts, Levels, and System Architecture

2021-12-21

Introduction

Autonomous driving refers to a vehicle’s ability to sense its environment, make driving decisions, and operate with minimal or no human intervention. Over the last decade, advances in robotics, artificial intelligence, sensors, and computing hardware have pushed autonomous vehicles from research labs into real-world pilots and limited commercial deployments.

This post introduces:

  • the core concepts behind autonomous driving
  • the SAE J3016 automation levels, which are the industry standard
  • a high-level view of the system components used in autonomous vehicles

The goal is to build intuition rather than provide implementation-level details.

Concepts of Autonomous Driving

At its core, autonomous driving is about replacing or augmenting the human driving task with software and hardware systems. These systems must handle:

  • perception of the environment
  • understanding and prediction of other agents
  • decision-making and planning
  • precise control of the vehicle

To standardize how autonomy is discussed, the Society of Automotive Engineers (SAE) defined a widely accepted classification known as SAE J3016.

SAE J3016 Automation Levels

The SAE J3016 standard defines six levels of driving automation, from no automation to full automation. The key distinction between levels is who is responsible for monitoring the environment and handling failures.

Human driver monitors the driving environment

Level 0 – No Automation

  • The human driver performs all driving tasks.
  • Systems may provide warnings or momentary assistance, but they do not control the vehicle.
  • Example: basic lane departure warnings.

Level 1 – Driver Assistance

  • The system assists with either steering or acceleration/deceleration, but not both.
  • The human driver remains responsible for monitoring the environment.
  • Example: adaptive cruise control or lane-keeping assist (used individually).

Level 2 – Partial Automation

  • The system controls both steering and acceleration/deceleration in specific driving modes.
  • The human driver must continuously monitor the environment and be ready to take over.
  • Example: highway driving assistance systems.

Automated driving system monitors the driving environment

Level 3 – Conditional Automation

  • The system performs all aspects of driving within specific conditions.
  • The human driver must respond when the system requests intervention.
  • This level introduces challenges related to human reaction time and attention.

Level 4 – High Automation

  • The system can perform all driving tasks even if the human does not respond to an intervention request.
  • Operation is limited to certain environments or conditions (for example, geo-fenced urban areas).
  • Example: autonomous shuttle services in controlled areas.

Level 5 – Full Automation

  • The system can drive under all roadway and environmental conditions that a human driver can manage.
  • No human driver is required.
  • This level represents the long-term goal of autonomous driving research.

Possible Technical Approaches

Different organizations explore various approaches to autonomy, but most systems share a similar architectural structure. One example is the system design explored by Tartan Racing, a team known for autonomous vehicle research.

Core System Components (Tartan Racing Example)

An autonomous vehicle system is typically divided into the following major components:

  1. Mission Planning
    Defines the high-level goal, such as reaching a destination or completing a route.

  2. Behavior Generation
    Decides driving behavior, such as lane changes, overtaking, stopping, or yielding.

  3. Motion Planning
    Generates safe and feasible trajectories based on the selected behavior.

  4. Perception and World Modeling
    Uses sensors (cameras, LiDAR, radar, GPS) to detect and track objects, lanes, and road boundaries.

  5. Mechatronics
    Handles low-level control, including steering, throttle, braking, and vehicle actuation.

Autonomous driving system components

Current Challenges in Autonomous Driving

Despite rapid progress, several challenges remain:

  • handling rare and unpredictable scenarios
  • robust perception in poor weather or lighting
  • safe human–machine interaction
  • validation and safety assurance at scale

These challenges explain why fully autonomous (Level 5) vehicles are still an active research topic rather than a commercial reality.

Status

Work in progress.
This post serves as a conceptual overview and will be expanded with:

  • real-world examples
  • references to datasets and simulators
  • deeper dives into perception and planning algorithms