Amid growing concerns about the safety of self-driving vehicles, Henry Liu, a leading expert in autonomous vehicle testing, has called for the federal government to create a national driver’s test that self-driving vehicles would need to pass before being allowed on public roads. Liu, who leads the University of Michigan’s autonomous vehicle testing center, believes such a regulation would set minimum standards to ensure that these vehicles can safely navigate various traffic situations. He stresses the importance of federal oversight in establishing these safety standards to boost public confidence.
Safety Concerns and Public Trust
The increasing number of high-profile accidents involving autonomous vehicles has raised significant public concerns about their safety. Liu argues that successful testing of these vehicles in diverse traffic conditions would help improve public trust in their ability to drive safely. While he acknowledges that there is still much research to be done, Liu believes that autonomous vehicles, once fully developed, could potentially save lives and enhance the efficiency of the transportation system.
At present, there are no specific federal regulations governing self-driving vehicles, though some states have their own requirements. The National Highway Traffic Safety Administration (NHTSA) has been gathering data on crashes involving autonomous vehicles but has only issued voluntary guidelines, which do not include driving tests. As a result, self-driving cars must meet the same safety standards as traditional vehicles, with investigations into incidents occurring only after significant accidents.
Proposed Competency Test for Autonomous Vehicles
Liu proposes that a national driver’s test, or even a voluntary assessment, be developed to ensure that autonomous vehicles can perform basic driving tasks, such as making left turns at intersections without traffic lights, stopping at stop signs, and yielding to pedestrians. He believes such a test would prevent poorly performing vehicles from being deployed on the roads and help developers gain smoother approval for their systems in various cities. Although he acknowledges that no test can guarantee the elimination of all crashes, he sees it as a crucial step in ensuring safety.
As autonomous vehicles advance with machine learning technologies to make real-time decisions, Liu predicts widespread deployment of these vehicles on U.S. roads within the next five to ten years. He points out that Europe and China already require third-party testing for autonomous vehicles, while the U.S. continues to rely on self-certification by the companies themselves. Liu argues that basic competency tests would foster innovation in the industry, improve safety, and ensure that autonomous vehicles are ready for large-scale deployment, without hindering progress.
