The latest in autonomous technology
Testing of self-driving cars, shuttles, and delivery vehicles has been happening on the streets for years. With industry leaders like Waymo, Tesla, Cruise, and Uber jumping into the autonomous pool, there has been ongoing competition for the first and most road-worthy vehicles. While you may think of futuristic spaceships when we say “autonomous vehicle,” the process is much longer and detailed than that. Research and development are focused on incremental improvements to technology that assists the driver.
Have lane departure warnings or parking assist in your car? You’ve got a low level of assistance on a universal classification system developed by the Society of Automotive Engineers. As the levels increase up to level five, so does the vehicle’s amount of automation. It's these improvements that spell out interesting changes for the future of autonomous vehicles.
Autonomy in many forms
The technology has been applied to delivery services, dispatching small robot-like services to drop off packages or food. In the era COVID-19, this contactless option is in high demand and an interesting pivot for the autonomous vehicle community. Testing autonomous technology for relatively innocuous deliveries is decidedly less risky than transporting humans, allowing for more lenient regulations and restrictions.
Google’s autonomous vehicle shop, Waymo, launched its Via service to do exactly that, while Walmart is working with six different companies to provide autonomous delivery services. And don’t forget about the sky — unpiloted drones are also being used for aerial drop-offs, parachuting everything from COVID-19 tests to online orders. Most delivery options remain decidedly in the testing phase, with regulators rushing to catch up to the innovative applications.
On-road testing
When it comes to transporting people, the same sort of race exists. The City of Toronto has recently announced its plans to test an autonomous shuttle in 2021, offering a driverless experience to eight passengers at a time. Toronto’s test is a first for the city, and follows many of the guidelines proposed by other municipalities — a limited route, on-board safety steward, and smaller operating capacity. These somewhat-closed testing circuits allow operators to control variables in a safer ecosystem.
At the other end of the spectrum is the Tesla model, wherein the electric vehicle manufacturer has just released its beta software update — Full Self-Driving mode, also called Autopilot. Available to select Tesla owners, Autopilot can control the car’s travel path down a lane, around curves, and adjust speeds accordingly — features that were all developed through a constant feed of data from the Tesla neural network of vehicles. The limited release allows for some of its nearly 1M owners to pilot the programming on busy city streets, a plan that has critics disputing the dangerous move.
While the car still requires a driver to be ready to take control if there is an issue, many are concerned that this new technology will create inattentive drivers. The Autopilot function requires the driver’s hands to be on the wheel to function, theoretically allowing for a quick reaction, should something go awry. Technology is still in development, and with that comes limitations. If a roadway has faded or missing street marking, the functionality may be limited, inaccurate, or fail altogether.
Other critics point out that the name — Full Self-Driving — is misleading. The car can’t fully operate independently, technically classifying it as a driver assist feature rather than an autonomous vehicle. Studies have shown that driver behaviour changes when they think a car is more autonomous than assistive, leading to over-confidence in the onboard system. This trend is troubling, as it may lead to drivers being more distracted and less reactive to potential collisions.