Remote human operators: Enhancing the safety of autonomous vehicles

MIT engineer Cathy Wu recognized that safety and reliability are the main roadblocks to the widespread adoption of autonomous vehicles. To overcome these challenges, Wu proposes a hybrid system where autonomous vehicles handle simpler tasks independently, such as cruising on highways, while remote human operators guide more complex maneuvers. Drawing inspiration from air traffic controllers, this approach envisions human supervisors directing autonomous vehicles in critical situations.

Framework for remote human supervision

In a recent IEEE Transactions on Robotics publication, Wu and her team introduced a framework for efficient remote human supervision of autonomous vehicles without compromising passenger safety. They emphasized that if autonomous vehicles coordinate with each other on the road, the instances requiring human intervention can be significantly reduced.

To evaluate the feasibility of this system, the team focused on merging, a maneuver where autonomous vehicles often struggle. By having remote human supervisors momentarily take control while merging, the team aimed to assess the efficiency and safety of the system. The researchers determined the maximum time each human supervisor should spend on a single merge to ensure effectiveness. They also investigated whether a few supervisors could manage a larger group of autonomous vehicles and improve the human-to-car ratio while maintaining safety.

Surprisingly, the researchers discovered that when autonomous vehicles coordinate with each other, the need for human intervention decreases significantly. By adjusting speeds and creating space for merging vehicles, cooperating autonomous vehicles eliminate risky merging situations. The team devised mathematical frameworks to assess the potential of scaling remote supervision and quantified the influence of cooperative autonomous vehicles on traffic, thereby enhancing reliability.

Although these concepts are hypothetical and not currently in use, the researchers believe that exploring these possibilities can drive advancements in autonomous vehicles. They emphasize the importance of finding a balance between automation and human guidance to ensure passenger safety and maximize the benefits of autonomous driving.

Looking ahead, the researchers plan to focus on other challenging scenarios on highways, such as lane changes and overtaking slower vehicles. While the full automation of highway driving could save commuters a significant amount of time, safely navigating complex or unexpected situations remains a challenge. Remote human supervisors could fill this gap, ensuring passenger safety and allowing for a seamless driving experience.

The study opens up new possibilities for the future of autonomous vehicles, suggesting that a hybrid system combining autonomous capabilities with human supervision can pave the way for safe and reliable deployment. By reimagining the concept of autonomous vehicles and embracing innovative solutions, we can create a future where passengers can enjoy the benefits of self-driving technology while maintaining the highest standards of safety.