Pedestrian deaths in America hit a 30-year high last year. Official figures show that more than 6,500 pedestrians lost their lives in 2019, and that number is growing substantially each year. Many have argued that the introduction of autonomous vehicles will further add to the number of deaths on America’s roads. But a new motion prediction system designed by a leading technology company promises to make pedestrians on the road safer than ever before.
A unique piece of tech
Southwest Research Institute (SwRI) has developed a computer vision tool that utilizes a deep learning algorithm to predict pedestrian movements. Rather than sticking to traditional motion prediction technology like optical flow, SwRI has used unique temporal convolutional networks instead. The developers say that this has allowed them to produce a vital piece of safety technology that’s highly accurate and provides almost instant results. To do this, the technology focuses on a pedestrian’s pelvic region to predict a sudden change of movement rather than the rest of the body, meaning it can detect when a pedestrian is going to make a potentially dangerous move on the road.
Are autonomous vehicles in need of this technology?
Since March 2018, when 49-year-old pedestrian Elaine Herzberg was hit and killed by an autonomous Uber vehicle, questions have been raised about how safe driverless vehicles are. The investigation into her death determined that various factors were to blame, including the autonomous system which failed to identify her as a human on the road, or her predicted path. When a pedestrian is injured or wrongfully killed on the road, professional assistance is often sought as they can help them and their families to claim compensation and get justice. As autonomous vehicles are a relatively new phenomenon, an attorney will be able to give legal assistance on the best course of action. However, it can also be reassuring for pedestrians to know that they are better protected as a result of SwRI’s new technology.
Will it improve pedestrian safety?
When drivers see a pedestrian, they will typically drive with extra caution, slow down, and second guess the pedestrian’s next steps based on experience and the pedestrian’s most likely cause of action. Despite this, busy intersections and urban roads are still common sites for pedestrian accidents. The team behind SwRI’s technology says that a computer in an autonomous vehicle being able to predict what a pedestrian will do next is pioneering and something that should be appreciated for its ability to process such a complex scenario accurately. One of the biggest benefits that this technology offers over similar motion detection technology is that it doesn’t just predict straight-line movements. This means that it reduces the chance of a pedestrian being injured by an autonomous vehicle, regardless of which direction they choose to travel in.
An increasing number of pedestrians are being injured on America’s roads each year, and the surge in autonomous vehicles over the next decade could increase these numbers further. Thankfully, this pioneering new technology will make autonomous vehicles and roads safer than ever for pedestrians, so it should be welcomed with open arms.