Autonomous vehicles are not simply futuristic options. Though we may not recognize it, AVs have slowly integrated into everyday life. Automakers and tech companies now openly test autonomous prototypes on the road. Perhaps less noticeable, yet far more prolific, are the large number of passenger vehicles on the road equipped with advanced driver-assistance systems. And while not so obvious, automated braking systems, lane-keeping systems and other driver-assist features have already marked the debut of phased automated mobility technology.
The road to autonomous is long and one of incremental progress and deployment dependent upon many variables. While on-road and in-lab testing continues to march forward in the approach to full Level 5 autonomy, challenges to autonomous transportation have remained relatively consistent.
To reach full autonomy, developers must achieve and maintain a robust level of environmental awareness (360-degree perception) in the vehicle, enable accurate localization or ability to understand vehicle position on the road, and have the necessary intelligence to predict and plan human-like driving decisions (driving policy). This means visible AVs that can meet and perform in all traffic instances. Qualcomm has been working on these autonomously driven vehicles and has learned it is simpler to drive in controlled operating domains such as fixed lanes and daylight scenarios. Data is the main limiting factor, and getting data for the foreseeable scenarios is critical.
But as driving scenarios become more complex, it also becomes increasingly difficult to get the necessary data needed to inform the development of more robust systems. In other words, the data acquisition and system development are logarithmic, not linear.
Logistical challenges lie beyond the technological hurdles of simply rolling out the necessary hardware. To achieve fully autonomous transportation, connectivity will need to be constant, reliable and of high fidelity. AVs will need to be connected at all times, because receiving and transmitting data from various cellular vehicle-to-everything sources includes other vehicles and road users, road infrastructure, pedestrians, emergency services and more.
Vehicles will also need input from high-definition maps to provide information about the surrounding geography and what’s ahead on the road. Some of this data will be communicated directly from vehicle to infrastructure/vehicle and some will be 5G network-based. However, it will be key to guarantee fast and reliable connectivity wherever a vehicle may go, from the most congested urban areas to the loneliest of country roads.
Policy questions also remain. For example, right now road users bear relatively straightforward responsibility in the event of an accident or traffic violation. However, once the vehicle takes over the majority of driving functions, the questions about responsibility become more difficult to answer, and new regulatory standards and practices will need to be developed.
While it will take years to develop systems, infrastructure and frameworks needed to achieve full autonomous mobility, by far the biggest benefit of these systems is one we already see on the road: safety.
A relatively basic level of driver-assist functions — such as automatic braking systems — are well-distributed in the market and have a meaningful impact on road safety and collision avoidance. Even at the starting level, these systems have the ability to save lives and form the base layer of future AV systems. Other available semiautonomous systems, such as highway autopilot functions (so long as they provide a high confidence level for drivers), go beyond safety to increase driver comfort and productivity while reducing stress.
As we approach full autonomy, current data points to an incremental transition from built-in and proven safety to increased convenience and productivity for both drivers and passengers. Commercial fleets will also see productivity improvements, as the move away from human drivers will encourage increased fleet utilization across a variety of commercial applications from trucking to robotaxis. Soon, vehicles will spend less time in storage and more time on the road.
These advancements in driver-assist development and adoption will dramatically affect future vehicle design and use. Manufacturers will need to shift their focus away from just driver conveniences and toward the entire user experience. While content consumption within the vehicle is a primary consumer consideration, interior designs must adapt to provide a more comfortable, social and connected experience.
To remain competitive, automakers will also need to prioritize user experience and deliver new features and improvements via over-the-air updates and adopting app-based models. This approach remains in its relative infancy, but the days of the static, nonupgradeable vehicle will soon be a memory.
However, the road to autonomy is a steady march and not a race. It will not be the flip of a switch, but rather a constant iteration of how we think about transportation and the interaction with other vehicles.
Today’s development and deployment of driver-assist systems are just the tip of the iceberg, and consumers stand to benefit each step of the way as roads become safer and transportation becomes more convenient.