Overview of Collision Avoidance System

Collision avoidance is one of the fundamental functions in advanced driver assistance. A collision avoidance system is an active pre-crash system, unlike passive systems like seat belts and airbags that mitigate the post-crash impact. Automated Driving Toolbox™ provides various application examples that show you how to design and test collision avoidance systems. This topic introduces the typical workflow of a collision avoidance system, presents an overview of various application examples, and helps you get started at building a collision avoidance system.

Workflow

The typical workflow of a collision avoidance system consists of four major steps.

Object Detection — Detect objects in the environment of the vehicle using data collected from multiple sensors, such as camera, radar, lidar, and IMU.
Sensor Fusion and Tracking — Fuse information from multiple sensors, and track the positions of detected vehicles across multiple frames.
Collision Risk Evaluation — Evaluate the possibility of a collision based on analysis of the available data.
Action — Alert or assist the driver to prevent a collision.

Typical workflow of collision avoidance systems

Types of Collision Avoidance Systems

You can categorize collision avoidance systems based on the action that the system takes in the event of a potential collision.

Collision Avoidance Alert Systems — Alerts the driver when it evaluates a potential collision. Forward collision warning (FCW) systems are an example of collision avoidance alert systems.
- Forward collision warning system — Warns the driver if the distance between the vehicle and the colliding object, or the time-to-collision (TTC), crosses a safe limit.
Collision Avoidance Assist Systems — Assists the driver in preventing a collision when it evaluates a potential collision. Adaptive cruise control (ACC) and autonomous emergency braking (AEB) are examples of collision avoidance assist systems.
- Adaptive cruise control — Adapts the speed of the vehicle to the traffic environment.
- Autonomous emergency braking — Automatically reduces the speed of or stops the vehicle to prevent collisions.

Application Examples

Automated Driving Toolbox provides these examples for designing and testing collision avoidance systems.

Type	Sensor Input	Task	Example
Forward Collision Warning	Vision detections Radar detections IMU measurements	Design an FCW algorithm.	Forward Collision Warning Using Sensor Fusion
Forward Collision Warning	Vision detections Radar detections IMU measurements	Integrate an FCW application with a vehicle network.	Forward Collision Warning Application with CAN FD and TCP/IP
Adaptive Cruise Control	Vision detections Radar detections	Design an ACC system.	Adaptive Cruise Control with Sensor Fusion
Autonomous Emergency Braking	Vision detections Radar detections	Design an AEB system.	Autonomous Emergency Braking with Sensor Fusion
		Automate testing of an AEB system against multiple driving scenarios.	Automate Testing for Autonomous Emergency Braking
		Cosimulate an AEB system with Simulink^® and RoadRunner Scenario.	Autonomous Emergency Braking with RoadRunner Scenario
		Design an AEB system with 14DOF vehicle dynamics. Cosimulate the AEB system with Simulink, RoadRunner Scenario, and Unreal Engine^®.	Autonomous Emergency Braking with High-Fidelity Vehicle Dynamics
		Analyze the impact of 3DOF and 14DOF vehicle dynamics models on an AEB system.	Autonomous Emergency Braking with Vehicle Variants
		Generate variants of a Euro NCAP seed scenario. Test the AEB system against generated variants using iterative testing method.	Automate Testing for Scenario Variants of AEB System