How will Artificial Intelligence create a real-world simulation for AID?

By CIOReview | Tuesday, August 7, 2018
69
82
20

Autonomous Intelligent Driving (AID) is the subsidiary of the major automobile giant- Audi AG which has a number of test vehicles that are running an autonomous vehicle stimulation platform from Cognata, a company which incorporates Israeli Artificial Intelligence and deep learning.

Danny Atsmon, CEO, Cognata said, “Previously autonomous vehicles were a hyped AI experiment and companies were at the stage of trying out prototypes on the road. Autonomous vehicles have now evolved into a more mature product that demands human-like performance, which means zero tolerance for safety issues and efficiency in a  high traffic environment. To achieve these goals, simulation is a key technology for developing and validating autonomous vehicles on a large scale in parallel to expansive road tests."

AI-based traffic models are recreated by Cognata in order to simulate real-world traffic conditions. It also includes recording certain specific sensors' interactions from real-world objects like buildings, streets and even the cracks in the road. They also conduct trials like testing multiple weather conditions and also allowing someone to run in front of the vehicle.

AID's multi-year partnership with Cognata has aided to implement a large-scale, cloud-based simulation solution. The efficient and rapid testing they say will help in increasing the safety while speeding up time for marketing AID-enabled autonomous vehicles. Cognata is currently providing AID with their end-to-end simulation offering and will support the autonomous vehicle's entire product lifestyle.

Alex Haag, AID's CTO said that from his outlook, apart from real-world driving, simulation is also crucial to develop and test AID's self-driving system technology.

"Simulation is needed to validate the autonomous vehicles in the virtual world, often before hitting the road. Safety plays a significant role for us and thus, simulating different scenarios and situations helps to continuously improve our software and reach the highest standards. To work with a reliable partner for simulation helps us accelerate the process and deployment of software," said Haag.

Haag also stated that developing the autonomous driving function is a challenging and a complex phenomenon in the context of current technological realisation.

"Next to technical maturity, some key factors will need to be adopted until self-driving vehicles will be on the market and commercially used. Looking at for instance at legislation concerning type approval and road traffic laws there still isn't an exhaustive, worldwide solution," said Haag.

He also mentioned that customers are readily accepting what the market offers, although they are adjusting to the new technology their acceptance varies by market.

"The development of the self-driving software comprises activities such as perception, prediction, trajectory planning as well as validating software by simulation. At AID we intend to build up a deep understanding of these tasks and integrate them in our software platform," said Haag. "There are currently many players that focus on either of these components and offer tools for fulfilling them and one promising partner for simulation is Cognata. As they can quickly integrate virtual 3D models from cities from all over the world a partnership with them helps us verify our software in different scenarios."

"With the help of the data of our autonomous vehicles stemming from different sensors, the simulated environment almost reflects a real-world scenario," added Haag.

Hagg also said that the aim behind AID is to offer the self-driving system for the urban environments, level 4 and up to mobility service providers from 2021 onwards.