Test Driving Autonomous Vehicles: Smart Becoming Smarter


Traditional RF tuner testing required multiple instruments to generate different broadcast standards. Multiple instruments were required to generate multiple carrier signals to test the critical functionalities in the infotainment systems, including channel search, alternate frequency selection, emergency and traffic announcements, among others.

Multiple high-definition acquisitions and synchronisation is required to detect issues arising from such systems. Surround sound, encryption schemes and full-HD video in latest infotainment systems have made multimedia test applications increasingly more complex.

One example of automotive test that requires an integrated, high-performance solution is ensuring the synchronisation of multiple screens for an infotainment system. Hebbur says, “Validations are also done to confirm the robustness of the UI where you try multiple button presses in a crazy sequence to check if the system handles all the presses gracefully.”

Testing electronics

All electronics in a vehicle must do their tasks properly. Verma says, “Different modules are supposed to work as an environment and, hence, call for testing at various stages.”
“For proper autonomous vehicles, we work on individual modules, testing these against the required parameters. After module testing, the environment as a whole is tested,” adds Gupta. “Lab car is used to simulate the vehicle’s running environments, and parameters are measured to check its proper functioning,” explains Nirmalkumar.

“We simulate the running conditions in a lab and then capture the readings,” explains Verma.

Testing these modules, however, raises the question of quantity. Lexus LIT IS, for example, is one-of-a-kind vehicle with about 42,000 LEDs. These LEDs in 2017 model of Lexus LIT IS are programmed to change colour in response to human gestures and music. People’s fascination aside, testing the electronics is already a humongous task with such devices.

Automated testing is the way to go.

Testing of automotive systems has been growing in size and complexity over the past few years. A modern luxury car may very well have up to 100 engine control units. At such high numbers, regular testing just does not cut it, and calls for appropriate test platforms to address the need for automation, automotive specific I/O and flexibility.

Automation has become increasingly valuable to automotive applications due to the complexity involved in testing whole infotainment systems. Testing an infotainment system without automation may take weeks and many man-hours. An automated infotainment test system, however, does it much faster. Test time can now be reduced to a day or two at most. Test procedures can be run in parallel for multiple software stacks for highly-aggressive development.

Safety and security are important

Current systems employ RF based protocols for detection and authentication of owners to unlock doors and start the engine. These protocols allow remote eavesdropping and have done so in the past. Due to the large code size and complexity, these extend the range of remote attacks to many tens of metres, or even kilometres.

The megamos crypto transponder, commonly found in keys and key fobs for wireless entry systems, is supposed to stop an engine from starting without the transponder being near the vehicle. However, fake frequency modulation transmitters broadcasting radio data system traffic message channel (RDS-TMC) information, adversely influencing the navigation system in cars, have been found to attack.

Tyre-pressure-monitoring systems also employ RF protocols to send pressure sensor information from within the tyre to an engine control unit in the vehicle. These protocols have been sniffed and spoofed to fool the engine control unit into reporting a false tyre-pressure warning to the driver. An attack on the system at the wrong time could have safety consequences in case the driver is distracted or alarmed.

The need for smart(er) cars

Smartcars still lack reliable means of communication. With vehicles, even a minuscule chance of error could lead to major disasters. Human involvement in driving provides for swift action being taken in real time. Autonomous vehicles leave a lot of open areas for question. A smart vehicle runs on algorithms, which is to create a parallel to the human mind. However, unlike the human mind, any number of algorithm or security-breach problems can lead to disaster. Tesla or Google cars’ accidents are prime examples. Proper testing can help root out problem areas, so that the next Uber you take is much safer.

Saurabh Durgapal is working as technology journalist at EFY



Please enter your comment!
Please enter your name here