Cognitive Computing

Cognitive computing will help autonomous vehicles share the road

Share this post:

It’s no secret that automakers and technology companies are racing to develop self-driving vehicles that will transform how we travel from place to place. What you may not know is that IBM has been inventing, patenting and innovating new technologies for the automotive industry for decades. For example, IBM recently partnered with Local Motors to introduce Olli, the first self driving mini-bus to integrate the power of IBM’s cognitive computing platform Watson to enable natural interaction between the vehicle and passengers.

IBM was also recently granted patents focused on improving autonomous vehicle safety through machine learning. The cognitive computing technologies could enable automakers to provide consumers with a new sense of confidence as self-driving vehicles and human drivers increasingly share the road.

Modeling human behavior

As autonomous vehicles are incorporated into our transportation networks, they must communicate with each other to share states and intentions. For example, caravanning — in which vehicles follow each other at impossibly close distances and at high speeds — depends, in part, on this communication.

We are working in IBM’s Thomas J. Watson Research Center in Yorktown Heights, N.Y. to understand and model human behavior, such as reaction times and next likely action based on past observations. As computational neuroscientists, we draw on our understanding of biological cognition and behavior generation in the brain. With that background, these cognitive computing models can help human drivers and autonomous vehicles better share the road.

U.S. Patent 9566986 (Controlling driving modes of self-driving vehicles for the invention) was inspired by our recognition that neither a human driver nor a self-driving vehicle is immune from failure while driving. Interestingly, each seems to fail at different things. Therefore, it seemed natural to train a cognitive computing system that can estimate and quantify risk of errors based on the current state of either the driver or the self-driving vehicle. Based on this estimate, the invention can begin to transition to the less risky mode of driving gradually, based on alerting and a hand off.

Estimating confidence around a current problem is something IBM Research excelled at in the Watson Jeopardy! grand challenge. Here, the same cognitive computing strategies are used to help driver and vehicle collaborate for a safer ride.

Another patent my co-inventors and I worked on similarly recognizes that human drivers have different styles of driving that can impact safety, especially when self-driving vehicles and humans share the road. Furthermore, certain super-human functions of self-driving vehicles, such as caravanning, are impossible when human drivers are involved. By better modeling of individual drivers, and sharing of these models among self-driving vehicles, this invention helps mitigate the danger of certain actions on the road.

U.S. Patent 9361409 (Automatic Driver Modeling For Integration Of Human-controlled Vehicles Into An Autonomous Vehicle Network) models human driving and allows self-driving vehicles to better anticipate the actions of human drivers. This invention employs the same interfaces self-driving vehicles typically use to communicate with one another. The vehicles can also share these models with each other, allowing them to better identify and react to human drivers as the models become more informed and accurate.

The National Highway Traffic Safety Administration (NHTSA) has proposed a system to classify vehicles with respect to autonomous capabilities. These range from Level 0, in which the driver is always in control, to Level 4, where the driver is not expected to control the vehicle at any time. Many years will be spent slowly replacing existing vehicles with versions along this scale. We can expect a diverse mix of vehicles with various features and abilities driving alongside manually-driven vehicles and the implications for their interaction.

IBM can leverage its cognitive computing capabilities around modeling human behavior and estimating confidence to help make entire transportation systems, such as shared roads, safer, regardless of which self-driving vehicle technologies and modes of operation become predominant in the next decade. It’s just another way that IBM Research is helping to make smarter systems for a better planet.







More Cognitive Computing stories

Meet Ari, the Smart Bike that Helps you Catch Green Lights

As more bikes take to the city streets than ever, researchers at RMIT have designed Ari, a smart e-bike that could help riders cruise the ‘green wave’.

Continue reading

IBM Researchers Design Winning AI Algorithms for Drones

For the second year in a row, researchers at the IBM-Illinois Center for Cognitive Computing Systems Research(C3SR) won a competition challenging experts worldwide to design low-power embedded systems for Internet-of-Things (IoT) applications. The 2019 Design Automation Conference (DAC) System Design Contest’s objective: create algorithms that can accurately detect and locate objects from images taken by […]

Continue reading

Data-Driven Aquaculture Management

Aquaculture requires innovative solutions to address challenges including population growth, climate change and limited natural resources.

Continue reading