"Robotaxis (Robotaxi·autonomous taxis) may temporarily cause more traffic congestion while strictly adhering to road rules, including speed limits. However, in the long term, they will reduce the risk of traffic accidents and decrease the demand for private cars and personal parking lots. The popularization of robotaxis is inevitable."
This is the prediction of Ragunathan Raj-kumar, George Westinghouse Professor of Electrical and Computer Engineering at Carnegie Mellon University (CMU). CMU is recognized as the birthplace of autonomous vehicle research in the United States, having started its research in 1984. Professor Raj-kumar led the CMU autonomous driving institute from 2004 to 2020, making him an indispensable figure in global autonomous driving research.
CMU has also shown remarkable achievements in the 'Grand Challenge,' a defense Department of Defense (Ministry of National Defense) autonomous vehicle racing competition that demonstrated the necessity of autonomous driving research. No vehicles finished the first competition held in the Mojave Desert in 2004, but in the second competition the following year, CMU secured 2nd place. CMU further excelled in 2007 by winning the 'Urban Challenge,' which changed the competition route from the desert to the city, outperforming Stanford University, Virginia Tech, the Massachusetts Institute of Technology (MIT), Cornell University, and the University of Pennsylvania. This showcased the world-class capabilities in autonomous driving technology. The students from CMU who participated in the competition later moved on to leading companies like Google and Tesla, driving advancements in global autonomous driving technology.
Professor Raj-kumar noted, "1.3 million people die in traffic accidents every year worldwide." He stated, "Robotaxis based on advanced autonomous driving technology are likely to be embraced by our society as a viable alternative to reduce traffic accidents." The following is a Q&A.
What initiated the autonomous driving technology underlying robotaxis?
"The 2004 DARPA Grand Challenge raised public awareness of the need for developing autonomous driving technology. However, CMU had been developing this technology prior to that. By winning the third Grand Challenge held in the city in 2007, CMU demonstrated the potential for autonomous vehicles to drive autonomously in urban areas. This shifted the perception of owning autonomous vehicles from a matter of 'if' to 'when.' Larry Page and Sergey Brin, the founders of Google, attended the third competition to witness the potential of autonomous driving. This served as the inspiration for Google's autonomous driving research. Notable members of the CMU team at that time include Brian Salaske, co-founder of Nuro (an American autonomous startup established by Waymo engineers, the first to operate on public roads), and Chris Urmson, founder of Aurora World, which gained attention for unveiling autonomous trucks.
Do you foresee robotaxis becoming popular in the near future?
"The popularization of robotaxis cannot be avoided in the future. Google Waymo has already successfully deployed initial models in major cities such as San Francisco, Phoenix, and Los Angeles. Although there are constraints like adverse weather conditions, any technical issues will likely be addressed within a few years. However, robotaxis will not be deployed on a large scale to the extent that they replace traditional cars in the next few years. It may take a significant amount of time for widespread deployment. I believe that robotaxis are a service worth embracing because too many deaths occur from traffic accidents on the roads. In fact, over 1.3 million people die in traffic accidents worldwide every year."
What are the current issues with robotaxis?
"Currently, robotaxis are too expensive due to the high cost of multiple sensors. Despite this, their performance declines under severe weather conditions like heavy rain, snow, and thick fog. They also struggle with poor driving abilities in heavy traffic. In construction sites where new roads are being created or in situations where accidents lead to road closures, they may fail to accurately process information. Roads often present unpredictable new situations periodically, and their responsiveness is lacking. In this sense, robotaxis may react differently from what human drivers would expect, appearing to behave maliciously. However, various studies are underway to address these issues, and they are expected to be resolved from a long-term perspective."
What impact will robotaxis have on road environments, traffic congestion, and car insurance?
"The initial deployment of robotaxis may cause more traffic congestion. This is because robotaxis strictly adhere to road rules, including speed limits. In reality, robotaxis move within speed limits even when there are no other cars on the road and do not stop unless at designated locations. Currently, robotaxis are programmed to operate in a very conservative (safe) manner. Of course, this behavior will improve over time. Robotaxis will replace private vehicles, possibly reducing the demand for personal parking lots and car insurance. Most importantly, as robotaxi technology advances, it is expected that collisions among vehicles, along with the resulting injuries and fatalities, will decrease significantly. This may lead to a substantial reduction in the demand for car insurance and insurance premiums. While this phenomenon won't occur in the short term, it is clearly a consequence in the long run."
In what direction do you expect robotaxis to develop in the future?
"The popularization of robotaxis will necessitate automotive manufacturing, operation, and support (maintenance and repair). Currently, these three aspects are being addressed simultaneously; however, if leading companies like Google Waymo only handle production while transferring operations and support to other companies, the spread of the robotaxi industry may rapidly accelerate. Robotaxis are also a means to respond to climate change. To operate effectively, robotaxis require various sensors and communication networks within vehicles. Such devices are suitable for use in electric vehicles (EVs) with a reliable power supply. Ultimately, this implies that robotaxis are optimized for EVs. The increase in robotaxis corresponds with a reduction in carbon emissions."
What advice would you give to corporations developing robotaxis or municipalities preparing to introduce them?
"If you are developing robotaxis, you need to have a clear understanding of the latest technology trends along with the opportunities, economics, and limitations that robotaxis can create. For municipalities, it is essential to consider how robotaxis will assist public safety. While there are various benefits, such as increased mobility for vulnerable populations and a reduction in accidents for elderly drivers, it is crucial to wisely address the job issues for existing transport laborers. The popularization of robotaxis will inevitably happen, and it's necessary to anticipate and respond to how they can change many aspects of our lives."
Plus Point
DARPA 'Grand Challenge' led to the practical application of autonomous vehicles
The DARPA Grand Challenge, an autonomous vehicle racing competition organized by DARPA in 2004, was held in the Mojave Desert of California. Designed to develop military autonomous technology, the event planned to award a prize of $1 million (approximately 140 million won) to the team that could autonomously complete 240 kilometers within the desert. Although 15 teams participated, none completed the course. Carnegie Mellon University traveled the farthest among the teams, but only managed to cover 11.78 kilometers. In the second competition, 23 vehicles participated, and five completed the course. Stanford University secured 1st place, while Carnegie Mellon University took 2nd and 3rd places. DARPA changed the competition course from the desert to the city for the third competition held in 2007. The race required teams to complete a closed Air Force base route of 96 kilometers within six hours, and the event was renamed the 'Urban Challenge.' The Urban Challenge featured more stringent and complex rules than the earlier Grand Challenge. It required compliance with all traffic laws while recognizing and avoiding other vehicles. The course included intersections, roundabouts, one-way streets, and freeway entrances and exits. This added the requirement for vehicles to not only drive autonomously but also recognize and interact with other vehicles and make intelligent decisions. Carnegie Mellon University won the competition with nearly perfect performance.