Safely on the Road With Your Ears Open

Tags: Artificial Intelligence
Who would have thought just a few years ago that cars would listen to the outside world, draw conclusions thanks to artificial intelligence and then be able to perform an evasive maneuver automatically?
Michael Scheibe, May 02, 2018
Michael Scheibe has been writing for the online magazine and our other publications in his capacity as editor in the ZF publishing team since 2016. He also functions as the link between external authors and ZF experts.
ZF allows vehicles to see, think and act, but why not also "hear"? This was the question that Florian Ade and Julian Fieres asked themselves. After all, hearing is an important factor when it comes to road traffic safety. ZF’s Sound.AI enhances vehicle sensors, giving cars a sense of hearing. Last year, the two economists began thinking about how they could enhance the radar and lidar technology for the car’s seeing capability to also include sound recognition. In the process, the vehicle’s hearing capability was to be developed based on artificial intelligence. Ade and Fieres therefore also included this technology in the name of their product: "Sound.AI".

They then took their idea of the hearing vehicle and submitted it to the Innovation Challenge 2017, a company-wide idea competition that was held for the first time last year. With this competition, ZF is breaking new ground to find interesting and marketable products of the future. The idea for “Sound.AI” came to Ade and Fieres during last year’s ZF Pitch Night held at the ZF Forum. There, an external startup presented an acoustic-based system for identifying problems in power grids and high-pressure gas systems. “Hey, why not develop something similar for cars?” recalls Fieres.

It all began with the Innovation Challenge

It all began with the Innovation Challenge

So they came up with various scenarios and then entered their idea in the “Innovation Challenge” competition. Yet joy and pride in making the cut dissipated quickly when faced with the question of how to transform their idea into a minimum viable product (MVP) within just a few weeks. An MVP is the “smallest possible feasible product” and shows that the product fundamentally works. Florian Ade works in ZF's Strategic Corporate Development, and Julian Fieres is Head of Strategy and Corporate Development for the E Division at the ZF location in Schweinfurt – neither of them are engineers. The fact that they were able to get the MVP developed in just five months is due, in part, to assistance from the Aachen-based Forschungsgesellschaft Kraftfahrwesen (fka) – a research institution for the automotive industry. Its experts supported the project by assisting with the sound hardware, building a test vehicle and optimizing the acoustics. An internal ZF support team consisting of trainees as well as artificial intelligence and sound experts was also involved.
The team around Florian Ade (front, 2nd from the left) and Julian Fieres (front, 3rd from the left) at the final functional test of "Sound.AI".

Wide scope for use

Wide scope for use

Just before Christmas 2017, they had reason to celebrate: The team presented the “Sound.AI” prototype at the final round of the competition. The system detects signals issued by sirens from approaching police, fire or emergency medical service vehicles (“siren detection”). Using a display screen, “Sound.AI" informs the driver which direction the emergency vehicle is coming from and suggests what to do. If it comes from behind, it recommends the driver to pull over to the right, or – on freeways – to form an emergency lane. However, “Sound.AI” can do more than just detect sirens, it is also suitable for other purposes. ZF has decided to develop the MVP in a corporate start-up up to a production-ready stage – with ZF contributions from the Electronic Interfaces Business Unit, Advanced Engineering and Digitalization. There have been first talks with automotive manufacturers.
For ZF, "Sound.AI" is also a successful example of a different approach to working in the company. This start-up culture allows technical modifications to result in specific market opportunities more quickly.
An emergency vehicle is approaching from behind: "Sound.AI" hears the signal and informs the driver via a display in the vehicle.

The car is listening…

Watch the video