Вештачка интелигенција

The New Age of AI-Driven Driving Technology

Summary

In a groundbreaking initiative, the Toyota Research Institute (TRI) unveiled its latest AI-supported driving technology last month. At the Thunderhill Raceway in Willows, CA, TRI invited journalists from the US and Europe to experience their new program, Accelerated Concepts, in […]

The New Age of AI-Driven Driving Technology

In a groundbreaking initiative, the Toyota Research Institute (TRI) unveiled its latest AI-supported driving technology last month. At the Thunderhill Raceway in Willows, CA, TRI invited journalists from the US and Europe to experience their new program, Accelerated Concepts, in action. This event marked a significant milestone for TRI, providing reporters with the opportunity to ride in research vehicles and simulators and firsthand experience TRI’s approach to autonomy.

TRI’s Human Interactive Driving (HID) team focuses its research not on autonomous robotaxis, but on active safety, including a high level of autonomy and driver involvement. This approach aims to make driving safer and more enjoyable by keeping drivers engaged and working alongside their vehicles as truly intelligent partners.

One of the highlights of the event was the introduction of the concept called the Driving Sensei. Using AI, the Driving Sensei assists drivers in honing their driving skills through a combination of AI-powered instruction and driver support. This innovative technology helps individuals become better and safer drivers while ensuring their engagement behind the wheel. Gill Pratt, the CEO of TRI, emphasized Toyota’s commitment to safety, stating, “Our human-centric approach uncovers better and safer ways for people and AI to collaborate. We enhance humans by developing models that predict driver actions and advancing AI to improve driving performance.”

Throughout the event, various research vehicles and simulators showcased the different technologies developed within HID’s research pillars:

  1. Human-Centered Learning: This technique leverages data-driven learning to create models of human behavior. From driver awareness to actions and intentions, these models ensure that technology understands drivers and supports them in the most natural way possible.
  2. Driver/Vehicle Performance and Safety: The HID team aims to develop expert-level driving skills with the help of AI. Reporters had the chance to experience this firsthand in an autonomous Toyota Supra model capable of drifting, pushing the boundaries of control. They also tested the abilities of a fully autonomous Lexus LC500 model, which demonstrated its capability to avoid obstacles at high speeds. Teaching these driving skills to AI lays the foundation for autonomous technology that helps drivers avoid accidents by navigating unexpected obstacles or dangerous road conditions.
  3. Shared Autonomy: In this aspect, AI and drivers work together towards a safer and more immersive driving experience. Participants had the opportunity to test TRI’s customized Global Research Innovation Platform (GRIP), a research vehicle with four-wheel steering and wheel-mounted electric motors that accelerate driving research advancements. The vehicle’s dynamic emulation allowed for new driver training scenarios in a controlled environment. Journalists learned how to control drifting through a combination of AI-supported instruction and guidance before putting their skills to the test on a simulated ice surface. They also experienced TRI’s AI-supported driving simulator on the track. This AI trainer used real-time natural language instructions to guide reporters based on their current actions. Journalists could also communicate with the agent through a chatbot-like interface to understand their performance and receive tips for improvement.

Through reimagining the interaction between humans and technology, TRI believes it can save more lives by bringing autonomous driving technology to more people and places sooner. Avinash Balachandran, director of HID at TRI, posed the question, “Can we save more lives by bringing automated driving technology to more people in more places, sooner?” The answer, according to TRI, is a resounding yes.

FAQ

What is the purpose of TRI’s HID team?

The HID team focuses on active safety, aiming to develop a high level of autonomy while involving the driver for a safer and more engaging driving experience.

What is the Driving Sensei concept?

The Driving Sensei concept uses AI to assist drivers in improving their driving skills through AI-guided instruction and support.

What are the research pillars of HID?

The research pillars of HID include human-centered learning, driver/vehicle performance and safety, and shared autonomy.

How does TRI enhance human drivers?

TRI enhances human drivers by developing models that predict driver actions and advancing AI to improve driving performance.

What technologies were showcased during the event?

The event featured various technologies, including AI-driven instruction, autonomous drifting capabilities, obstacle avoidance at high speeds, and immersive driving simulators.

For more technical information about the showcased technologies, visit the TRI Medium blog here.

Source: Toyota Research Institute