A Guide to SAE Level 2 and Above Features

Date: September 30, 2020

Author: Akshay Srinivasa

Number of cars with advance technology on the streets is increasing day by day, with more comfort features and faster machine. Features that were once a distant dream is now affordable to masses. Increasing number of cars on the street puts forth the challenge of safety of passengers as well as pedestrians. To address these issues, manufacturers provides safety kits bundled together in form of ADAS (Advanced driver assistance systems). SAE Classifies these safety features based on their intrusiveness with driver action. Fully autonomous being no driver involvement.

The processing powerhouse that drives modern day ADAS in passenger cars

Cars are equipped with various features listed above to make driving safer. Having variety if sensors across the length of the car performing different application requires dedicated processing hardware setup or a powerful GPU performing the desired task of all the sensors. Various semiconductor manufacturers like Texas Instruments, Renesas, NXP, Qualcomm have specific boards meant only for ADAS applications which are developed to handle complex sensor data and machine learning on the edge for certain application. These special purpose hardware or hardware accelerators need to manage flow of data from the sensors to them, process and further to the mechanical actuators that control various aspects of the car like steering, throttle and brakes.

Challenges in developing an SAE Level 2 features

With numerous features arises challenges of integrating multiple of those in a vehicle. Some of the common difficulties or challenges are discussed below;
  • Developing processing capabilities for data handling. Semiconductor manufacturers have also made significant advances in developing systems to incorporate vast amount of data that needs to be processed in fraction of a second.
  • Fusion of multiple sensors, this can mean using multiple sensors of same nature or fusing image sensor with other type like radar.
  • Developing a dependant mapping software when GPS fails (like in tunnels).
  • High cost of sensors like LiDAR which are priced steeply compared to radar or camera
One might argue for or against a single type of sensor for all the ADAS features. But researchers are still a long way from developing a single sensor system for all the features with necessary processing support either on edge or a powerful ECU.

Further reading

You might also like these blogs

An overview of Firmware

Look around and count the number of gadgets you own.…

Read More
Understanding Traffic Sign Recognition system used in ADAS (Advance driver assistance system)

Autonomous machines/ vehicles are all set to drive along the…

Read More
Bayesian Grouping of Multi Sensor Radar Fusion for effective Pedestrian Classification in Automotive Surround View

From long time camera is used as the sensor for…

Read More
Back to Top