When AI applications permeate our daily lives on every level, fully-autonomous driving technologies remain one of the most crucial aspects; aside from this, AI has other utilizations for future vehicle techs. Thus, on the afternoon of September 16th, our Auto Future editor team came to Taipei to interview Mike Huang, the CTO of Mindtronic AI, a vehicle AI startup, about AI techs’ creative applications.
Mike Huang, the CTO of Mindtronic AI, had an interview with Jeff Lee, the director of Auto Future.
In 2017, multiple core members from Fab giants founded Mindtronic, using their experience to establish cutting-edge vehicle AI tech applications. Since 2018, the company has achieved Taiwanese companies’ best record in CES: five Innovation Awards, demonstrating its outstanding technological prowess and vision.
When talking about Mindtronic AI’s role in the car industry, Mike has a unique take: in the era where developing L4/L5 fully-autonomous driving tech is mainstream, he focuses the company’s resources on the integration of the dashboard and infotainment. Borrowing the concept of domain controllers (the integration of software functions in specific domains) for smart cockpits, Mindtronic provides assistance developed by geometrical algorithms called Geometrical-driven service. For example, many car models on the market have eye-tracking functions that alert drivers when they are sleeping or unconcentrated; they also have POI (Points of Interest: attractions, restaurants, parking lots, etc.) marked on their navigation mapping as default. Mike’s R&D team has evolved these functions to gaze tracking and TOI (Target of Interest), using AI’s deep learning abilities to identify the drivers’ TOI by sampling their interests and referencing the map; and sending the target’s marketing information to the vehicle’s infotainment display for passengers and travelers. This kind of service has great potential to become a monetization model in the future.
Aside from smart cockpits’ application, another R&D focus of Mindtronic is Copilot- driving cooperation between humans and AI. Unlike autonomous driving, Copilot utilizes DMS (Driver Monitoring Systems) to connect ADAS (Advanced Driver-Assistance Systems), such as ACC and AEB, to take over driving when the driver endures cognitive impairments (due to mental or health issues). The “bridging” AI and Copilot function - let’s call them “adaptive blending driving policy”- could be another option for increasing driving safety besides fully-autonomous driving. What’s most exciting is the Self-Learning Framework R&D direction mentioned by Mike: the AI could learn personal driving habits and driving mode customization via facial recognition. A car could be “familiar with” multiple driving habits so it can switch to the Copilot mode most suited for the driver after facial recognition in the driving seat. Most importantly, these AI long-term learning data could be transferred to another car like being stored in a hard disk (so to skip redundant learning processes when replacing cars); anticipations for this technology’s commercialization couldn’t be higher. For Mindtronic, in-cabin understanding and road understanding are its long-term goals for developing and integrating AI techs.
Motivated by multiple CES innovation awards, Mike hopes to maintain this award-winning trend for yearly new product releases. On the business side, he’s been pushing aftersales package called Robot Tour Guide (technologies combining eye tracking and TOI, as mentioned above) to multiple markets like the Middle East and Europe, providing more convenient and safer self-drive trips for travelers with local rentals.
Mindtronic AI’s groundbreaking technologies integration with carmakers or first-tier suppliers will certainly be the next step to success for the company.