Self-driving car in downtown Los Angeles
Allen J. Scherben/Los Angeles Times/Getty Images
Artificial intelligence systems that control driverless cars still struggle to predict the sudden appearance of other vehicles, cyclists, and pedestrians, but new algorithms are making it more accurate to detect the presence of such hidden objects. We showed you how to make predictions and predict their movements.
“We definitely captured the complexities of the real world, such as hidden pedestrians and unpredictably moving cyclists,” said Hari Thiruvengada of California-based cognitive computing company VERSES AI. “We added occlusion inference to help predict road user behavior that is hidden from direct visibility.
