Researchers have allegedly developed a new algorithm that could help self-driving cars navigate narrow and crowded streets. Drivers are able to complete this task, however, not always without close calls and frustration. This means that programming an autonomous vehicle to do the same without a human driver or knowledge of what another driver on the narrow road might do was uniquely challenging for researchers. However, at Carnegie Mellon University’s Argo AI Center for Autonomous Vehicle Research, students, scientists, and professors doubled down to find a solution to the problem.
The team stated that their research is the first dive into this specific challenge facing autonomous vehicles as it requires drivers to collaborate to make it past each other safely. Human drivers must balance aggression with cooperation with other vehicles or obstacles on the road. Autonomous vehicles have been identified as a potential solution to the challenges of transportation and delivery services. However, this skill will be key to allowing autonomous vehicles to operate in the real world without human drivers. The results of Carnegie Mellon’s project were promising and its algorithm performed better than current models that disregard this unique challenge.
Read More: Algorithm Could Improve How Self-Driving Cars Take on Narrow Streets
