AImotive: Accelerating the Scalable Future of Autonomy

April 20, 2018 0 comments

Automakers are facing a self-driving-car disruption largely driven by the tech industry, and consumers are expecting their next cars to be fully autonomous. The technology which powers these vehicles features software and hardware solutions utilizing computer vision and artificial intelligence. A company which leverages these technologies to offer OEMs with automated driving functionality is AImotive.

AImotive is a self-driving car technology company headquartered in Budapest, Hungary. AImotive’s mission is to create globally accessible and scalable self-driving solutions. The company was founded in 2015 and has grown to a team of over 170 members in 4 offices covering three continents. AImotive has also secured self-driving test licenses at four locations worldwide ­which include Hungary, Finland, California, and Nevada and is currently working towards the fifth one. Their systems have also been trialled in France, under a partnership with Groupe PSA.

The company follows a vision-first approach to autonomous mobility. The technology relies on a vision-first sensor setup and artificial intelligence for advanced recognition and trajectory planning tasks. AImotive’s full-stack self-driving solution, aiDrive is developed and verified through tests in their photorealistic simulator, aiSim, and road tests around the world. The final point of their offering is aiWare, a generic artificial intelligence accelerator design. aiWare is engineered to accelerate neural networks in use cases where low power consumption and low latency are crucial.

AImotive’s approach to self-driving is deeply rooted in the use of artificial intelligence. This makes their technology scalable and safe. It is impossible to pre-program every possible traffic scenario a vehicle may encounter, and a properly trained neural network can adapt to unknown situations to ensure the safety of passengers and those around the vehicle. AImotive also initiated the Neural Network Exchange Format working group at the Khronos Group with a member of the AImotive team being the spec-editor of the standard. This open standard supports the development of new artificial intelligence-based chips and systems.

Promising Solutions Which Power Autonomous Vehicles

AImotive’s definitive innovations include the company’s industry-leading development pipeline and their complex approach to self-driving. AImotive’s technology has been developed to solve the challenges of self-driving. Creating a full-stack self-driving solution required development tools and a new breed of artificial intelligence accelerator. Utilizing the scalability of computer systems by incorporating virtual testing into their development decreases the time-to-market of the solution while ensuring their safe development. New feature requests are first reviewed by the company’s safety team and implemented by their experienced engineers following safety guidelines. New versions of the self-driving software are first tested in aiSim. This allows quick iteration and bug-fixing. Once updates are safe and stable enough, they are finetuned on a closed test course.

Finally, the new update hits the roads at their testing locations around the world. Running artificial intelligence-based systems is a demanding task. Currently, common GPU-based methods are not viable for use in production vehicles owing to their drastic power consumption, size and heat output. To solve this challenge AImotive’s team has designed aiWare, a new breed of neural network accelerator designed for real-world use. aiWare provides the same computational power as GPU-based acceleration with a fraction of the power consumption and a much smaller size factor.

A Leader with Unmatched Industry Knowledge

AImotive is headed by László Kishonti, Founder & CEO of the company. AImotive evolved from his first venture, Kishonti Ltd., which quickly became a leading high-performance graphics and computing solutions company. Under Laszlo’s leadership, the firm then turned its focus to automotive as artificial intelligence and autonomous driving started to gain global momentum. Laszlo continues to guide the company’s global efforts and innovation. Speaking at key events around the world László is a thought-leader and influencer in the autonomous mobility industry.

Partnerships and Recognitions

AImotive was featured in the CB Insights 100 Promising Artificial Intelligence Startups in January 2017. The company works with international partners to develop its technologies. Cooperation with some of the world’s leading brands in the automotive space, such as Groupe PSA, Volvo, SAIC, Samsung and Kyocera, validate their scalable approach to self-driving. AImotive has proven its technology on highways in the US and Hungary. They have also showcased their automated valet parking functionality in Hungary, the US, Japan and China.

When the Going Gets Tough

One of the greatest challenges AImotive faced was its fast growth. Integrating up to 60-70 new employees per year while also making sure that development was fast-paced and effective is a great challenge to the company.

Artificial Intelligence, Safety and Regulations

Full autonomy is unimaginable without the advanced use of artificial intelligence-powered solutions. Traditional decision-tree based solutions that use computer vision instead of artificial intelligence are not robust enough to guarantee the safe operation of self-driving cars. Recognition and trajectory planning, primarily based on visual information and carried out by artificial intelligence offers the autonomous vehicle the deepest understanding of its environment. Artificial intelligence cannot only recognize a partially occluded object but adapt to its environment. For example, the vehicle will be able to discern between safe and dangerous movements from pedestrians on the side of the road and will also be robust enough to understand that it can hit a plastic bag but a rock on the road must be avoided.

General consumers will first experience limited autonomy in the form of ADAS functionalities. These will appear in high-end models from top OEMs, with lower tier models following hot on their heels. To increase road-safety lawmakers may even make elements of self-driving technology compulsory for all cars produced after a certain date. These could range from already common collision avoidance systems to full highway autonomy. Early on, there will be several players in the autonomous space. This early period will be followed by industry consolidation. However, the first few players to achieve full autonomy will be those that gain the largest stake in the long run.

Times Ahead

Artificial intelligence is a quickly advancing field, however, its current uses are rather limited and the specializations of artificial intelligence are very narrow. Narrow artificial intelligence can, of course, be trained and developed for several tasks and will revolutionize many industries, and people’s lives in the coming years. Nevertheless, as the technology develops, general artificial intelligence will be reached at some point. The effects and ramifications of that technology, however, are part of a different conversation.

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.

Your data will be safe!Your e-mail address will not be published. Also other data will not be shared with third person.