Seeing with sensors

All About Space/Shutterstock.com

Today, cars and vehicles have sophisticated connected systems, which allow them to see, hear and interact with copious amounts of data enabling enhanced monitoring and better diagnostic reporting.

In autonomous or self-driving vehicles, gathering information from sensors, processing it efficiently and responding appropriately in time is of the utmost importance. For autonomous cars to be safe on the roads, they must do one thing extremely well: object detection. How does it work?

Sensors for object detection

Connected cars (or vehicles) can communicate with systems outside the car. Sensors at various points in the car collect different data like speed or position and communicate it to the telematic control unit, which is responsible for navigation, control of vehicle speed, location tracking and other vehicle diagnostics or maintenance alerts.

For object detection, the primary sensors in play are:

• Camera sensors, capture images of surroundings and provide visual input.
• RADAR sensors use radio waves to detect the distance of objects and can supplement the camera’s input during poor visibility conditions like rain or fog.
• LiDAR sensors which use laser pulses to build an accurate map of the surroundings.

Often, all these sensors are deployed and information from them works together in fusion.

All these combined can help the autonomous vehicle see – detect the distance to an object, or information about speed, estimate the 3D shape of the object, and assist with adaptive cruise control and blind-spot monitoring.

Computer vision

In autonomous vehicles, computer vision is key to object detection on the road and enabling perception of the car’s environment with high levels of accuracy. Computer vision involves analyzing images or visual information with an AI model to automate actions based on that understanding. They can have a substantial impact on telematics with applications ranging from smart parking in general cars to avoiding collisions in autonomous driving.

How are standards helping?

Technical committees look at crucial elements of connected car systems, developing standards contributing towards a securely connected road infrastructure.

IEC TC 47 prepares international standards for semiconductor devices. One of its standards, IEC 62969-1 specifies requirements for power interfaces in automotive vehicle sensors, ensuring reliable communication and compatibility. It is also in the process of looking into testing methodologies for the performance of different detection modules in autonomous land vehicles such as the ultrasonic and visual imaging modules.

The information from the sensors then needs to be relayed to and from across various connected internal and external systems. This is how a car service provider can keep track of the car’s route and performance or alerts can be sent out during an emergency.