With the deadlines approaching in various countries for bans on the sale of gas-powered automobiles, the global auto market is now looking to electric vehicles and their skyrocketing sales. Vendors in Taiwan have also stepped up to form an electric vehicle alliance to integrate Taiwan's industrial capabilities, in order to seize the huge business opportunities that come with this rapdily expanding market. Compared with pricey electric vehicles, gas-powered cars with partial self-driving capabilities have been much more appealing to consumers with their promise to offer a next-generation smart driving experience paired with an affordable price tag. Ford's fourth-generation Focus, which is strategically equipped with level 2 self-driving capabilities, hit a sweet spot in global market positioning, and immediately stimulated other carmakers to follow suit with similar products. Priced at around NT$800,000, nearly half the price of an electric car, the Focus has pushed global smart car development to new frontiers.
Major car manufacturers have been able to launch cars packed with a variety of smart features after just a few years of R&D. This is not only due to the rapid advancement of semiconductor architectures driving down the prices of chips and sensors, but also thanks to assistance from startups making clever use of AI technology. In the sea of startups from around the world, Mindtronic AI prevailed against intense competition at the Consumer Electronics Show (CES) 2020, and was recognized for its self-developed Driver Monitoring System (DMS), Advanced Driver Assistance System (ADAS), and Human Machine Interface (HMI). In fact, the company has already received five CES awards in the four years since its founding, which is a testament to the company's outstanding capabilities for technological innovation.
Mindtronic AI technical director Mike Huang said that car makers typically specialize in areas like transmission systems, chassis, vehicle designs, production, and sales. AI-based intelligent vehicle monitoring systems, however, are completely alien to them. Because auto makers are unlikely to get immediate results from their in-house R&D teams, and because it would be difficult to compete with outside organizations that are already familiar with AI technology, seeking external help naturally becomes their best option. Mike said that after winning the CES award, Mindtronic AI started to work with Japanese, European, and Chinese car manufacturers. They have already made their way into the auto supply chain and are expecting to see their products used in new cars by 2022.
Mindtronic AI has been working in the field of artificial intelligence for many years with a focus on applying AI applications to automotive safety. In early 2018, the company launched its first smart AI dashboard solution. Later in the same year, it was nominated as one of the 10 most promising startups in the world by the CES Consumer Technology Association (CTA). Because of its strong R&D capabilities, the company was selected as one of the Hello Tomorrow APAC Challenge Finalists in France. In 2021, it was selected by the German Trade Office Taipei as the best startup in Taiwan and one of the top three in Asia.
Overcoming vehicle AI challenges with proprietary edge AI solution
Many startups have been drawn to enter the smart car market because it offers extremely high potential for rewards, but few have been successful in working out deals with car brands. The reason is that AI technology alone is not enough to overcome challenges of real-time processing, uneven lighting, and unstable in-vehicle environments. Director Huang pointed out that when a vehicle is in motion, the system must be able to continuously recognize its surroundings, which means being able to make at least 30 calculations per second to ensure driving safety. Secondly, image clarity suffers greatly if the camera is capturing footage in uneven lighting. The AI must be able to correct the footage and make corresponding judgments in real-time to prevent the environment from impacting driving safety.
Mindtronic AI uses a lightweight deep learning framework that is fast, accurate, and power-efficient, which will automatically initiate self-driving system takeover if it detects that the driver is engaged in dangerous behaviors while the vehicle is in motion. Furthermore, the company has also introduced a self-driving edge computation AI technology that monitors the driver's facial expression and eyes to detect driver fatigue, making the self-driving takeover mechanism even more accurate.
In 2020, Mindtronic AI received mentorship support and startup resource assistance from TTA, the Ministry of Science and Technology, and the Ministry of Economic Affairs to help the company grow. TTA, in particular, connected the company with National Yang Ming Chiao Tung University's IAPS accelerator program. This led to the company's participation in the Taiwan Excellence Awards and other activities for increased global brand visibility and exposure to international fundings. Mindtronic AI plans to continue to invest in product optimization and look for collaboration opportunities with U.S. automakers to seize the huge business opportunities of the smart vehicle market.
Driver Monitoring System
Photo: Mindtronic AI
(Editor's note: The original article was published in TTA Magazine Issue 9. Read more startup stories in TTA Magazines.)