Companies have poured billions into funding and resources to develop autonomous vehicles and yet, after years of research and development, have achieved at best Level 2 autonomy — which is far from the promises of fully self-driving Level 5. Currently, the most commercially accessible ADAS system comes from Tesla and is still rife with safety challenges. Where did we go wrong?
Autonomous mobility has always been viewed through an individualist lens rather than collectivist, but as more companies try — and fail — to achieve Level 5 autonomy, we are now understanding that vehicles need to share vision in order to safely operate in a shared environment. This is the outlook for Autonomy Through Infrastructure™ (ATI™), an approach my company — Seoul Robotics — is pioneering whereby sensors equipped with 3D perception software are placed on infrastructure — such as street lights, buildings, and highway overhangs — to fully capture the environment and communicate with other sensors and the 4/5G systems that come standard on vehicles today.
By placing 3D sensors on infrastructure rather than individual vehicles, the vision scope transforms from the vantage point of one vehicle to the entire surrounding environment, eliminating blindspots. ATI solves all of the pervasive barriers to Level 5. Beyond the benefits of improved vision, ATI is extensible at a far lower cost than traditional autonomous mobility solutions. Where AVs require thousands of dollars of hardware for each vehicle, infrastructure-based autonomy can manage thousands of vehicles with only a few sensors.
For industry applications, the case for investing is easier to make; however, the technology is still not even close to achieving Level 5 autonomy because it lacks the ability to both perceive and anticipate obstacles effectively. When vehicles are guided by their internal software with information collected from their own hardware sensors and cameras, there will always be a level of miscommunication, creating risk and limiting the ability to reach Level 5 autonomy.
Taking an infrastructure-based approach has massive potential to fuel autonomous mobility as it centralizes the data collected by 3D sensors, delivering more effective V2X communications and enabling true Level 5 autonomy. ATI enables vehicles to move without any human involvement within a specific mapped area, making it an extremely powerful solution particularly for Original Equipment Manufacturers (OEMs).
Using this concept, our team at Seoul Robotics designed Level 5 Control Tower (LV5 CTRL TWR), a mesh network of sensors and computers on infrastructure that guides vehicles autonomously. LV5 CTRL TWR acts as the brains of the mechanism, collecting all the 3D data and then automating vehicles accordingly using V2X communications. When vehicles have shared vision through a united network of sensors and decisions are made through a shared system, vehicles can move simultaneously with precision.
LV5 CTRL TWR is currently in the early stage of commercial deployment with BMW to automate last-mile fleet logistics at their manufacturing facilities in Germany. With full vision of the environment, the system can autonomously guide hundreds of vehicles simultaneously from the factory floor to a parking facility where they can be housed before moving to dealerships, creating a safer and more efficient process. This successful implementation of the technology shows that, when applied in specifically mapped and controlled settings, ATI achieves complete Level 5 autonomy today.
Beyond OEMs, infrastructure-based automation has the potential to transform operations for a wide range of business applications ranging from vehicle distribution centers to car rental companies and trucking logistics. When applied in these settings, businesses are able to improve efficiency significantly as well as reduce labor costs and safety risks.
To truly impact the way society moves, we need to look beyond autonomous vehicles and towards incorporating infrastructure-based automation throughout the public and commercial spaces we travel through in our everyday lives. With ATI, the valet can become autonomous, car rental companies can move their fleets and public transportation can safely navigate congested streets in high density areas. When used in conjunction with other solutions, ATI will realize everyday Level 5 autonomy within smart cities, freeways, and beyond.
As AV companies work to develop their vehicles, the missing puzzle piece remains a universal connection. Being able to communicate with one another is an important element that has been addressed; however, without infrastructure-based sensors, blindspots will always exist and vehicles will continue to operate on an individual basis. Looking more broadly into the future, Level 5 autonomy will only be possible with the inclusion of ATI.
Infrastructure-based automation is a collectivist approach to autonomous mobility that has proven to deliver more effective results than the individualist approach. As the industry works to bring Level 5 to the masses, ATI will become an increasingly useful approach and key factor in delivering our autonomous future.
HanBin Lee is the CEO of Seoul Robotics.
© Geospatial Media and Communications. All Rights Reserved.
