MIT Scientists Develop Autonomous System With "Better Reasoning"

MIT Scientists Develop Autonomous System With

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Researchers from MIT and Toyota have collaborated to bring a new autonomous system to help vehicles navigate tricky intersections.

The system helps to mitigate the problem of blind spots for humans and driverless vehicles alike. By improving safety in intersections, the research signals a future of much safer roads.


Navigating intersections

In 2016, approximately 23 percent of fatal and 32 percent of nonfatal traffic accidents in the U.S. occurred at intersections, a 2018 Department of Transportation study says.

Even driverless car systems can fail when visibility is obstructed at busy intersections.

Impressively, the new navigation system uses uncertainty as data. As per MIT News, the group of researchers developed a method that estimates the risk of going into situations without adequate data.

The system weighs several critical factors. For example, it assesses visual obstructions, sensor noise and errors, the speed of other vehicles, and even the attentiveness of other drivers.

The navigation system measures the risk and then advises the car to stop, merge into traffic, or inch forward in order to gather more data.

“When you approach an intersection there is potential danger for collision. Cameras and other sensors require line of sight," Daniela Rus, director of the Computer Science and Artificial Intelligence Laboratory (CSAIL) said in the MIT press release.

"If there are occlusions, they don’t have enough visibility to assess whether it’s likely that something is coming.”

Trials in a mock city

In their study, the researchers tested their system in more than 100 trials with remote-controlled cars turning into a busy intersection in a mock city.

The system helped the cars to avoid a collision 70 to 100 percent of the time. They also tested other variations of the system that were less successful.

“In this work, we use a predictive-control model that’s more robust to uncertainty, to help vehicles safely navigate these challenging road situations,” Rus said.

While the research needs more work, such as factoring in the presence of pedestrians, it is much needed another step towards safer roads.

Watch the video: Lecture A brief overview of Hessian-free optimization Neural Networks for Machine Learning (July 2022).


  1. Jerold

    She should tell you you are wrong.

  2. Black

    I am sorry, that has interfered... This situation is familiar To me. Write here or in PM.

Write a message