Nvidia May Have A Fix For Self-Driving Cars Detecting Emergency Vehicles
The GPU maker is working to fix a pressing issue related to self-driving cars and how they respond to emergency vehicles with the help of AI.
Nvidia recently filed a patent to enable self-driving cars to respond autonomously to nearby emergency vehicles. The company is primarily famous for making graphics cards that gamers use worldwide. However, Nividia is also renowned for its prowess in developing artificial intelligence-based technologies for computers as well as automobiles.
For starters, Nvidia’s DLSS and Optix are two AI-based applications that increase graphics fidelity in computer games. Furthermore, Nvidia also uses its GPUs with powerful AI capabilities to help design even better graphics cards. Thus, saving resources and improving chip design in the process. Moreover, Nvidia also offers dedicated hardware, including specially designed GPUs for accelerated machine learning development.
With its expertise in artificial intelligence, it only seems natural for Nvidia to develop hardware and software solutions for self-driving cars. Nvidia already has partnerships with some of the biggest automotive names in the industry, including Audi and Mercedes Benz. Moreover, it recently filed a patent with the United States Patent and Trademark Office that explains how it plans to use a specially designed system containing microphones to identify and respond to different emergency vehicles.
Nvidia’s Solution For Self-Driving Cars
The broad explanation of Nvidia’s patent involves identifying and responding to alert sirens of different emergency vehicles with the use of a pair of microphones. The solution proposed by the company requires these microphones to be attached to an autonomous or semi-autonomous vehicle and work in collaboration with a Deep Neural Network (DNN). The microphones will generate audio signals based on the sound alerts present in the surrounding environment. These audio signals would be first used to create a frequency spectrum and then sent to be analyzed by the DNN. After processing the information, the system will return a prediction on various aspects of the audio, including its location, direction and type. Thus, allowing an autonomous or semi-autonomous vehicle to plan its response accordingly.
It looks like a workable solution, but it is still in a very early stage. Furthermore, most patents do not become real products for many different reasons. Since Nvidia has recently filed this patent, developing it into a working product would take a reasonable amount of time. Most autonomous cars use LiDAR sensors, cameras, and AI to implement self-driving capabilities. Tesla, however, is one exception to this rule as it does not rely on the LiDAR sensor in its electric vehicles. Tesla earlier used Nvidia’s chip to power its self-driving cars but later switched to in-house developed silicon to save cost.
