Visit https://brilliant.org/NewMind to learn STEM for free, and the first 200 people will get 20% off their annual premium subscription

Nearly a decade ago, a vast list of technology companies, with a combined investment of more than $100 billion, claimed that within five years the once unimaginable dream of fully self-driving cars would become a normal part of everyday life. These promises obviously did not come true. Despite this abundance of funding, research and development, expectations are beginning to change as the dream of fully autonomous cars proves to be far more complex and difficult to realize than automakers anticipated.

THE LAYERS OF SELF-DRIVEN
Just like humans drive a vehicle, autonomous vehicles use a layered approach to information processing. The first layer uses a combination of multiple satellite-based systems, vehicle speed sensors, inertial navigation sensors and even terrestrial signals such as cellular triangulation and differential GPS, summing the vehicle's motion vector as it travels from its starting point to its destination. The next layer is characterized by the process of detecting and mapping the environment around the vehicle, both for the purpose of following a navigation path and avoiding obstacles. Currently, the main mechanisms of environmental perception are laser navigation, radar navigation and visual navigation.

LIDAR
In laser navigation, a LIDAR system launches a continuous laser beam or pulse at the target, and a reflected signal is received at the transmitter. By measuring the reflection time, signal strength and frequency shift of the reflected signal, spatial cloud data of the target point is generated. Since the 1980s, early computer-based autonomous vehicle experiments relied on LIDAR technology and even today it is used as the primary sensor for many experimental vehicles. These systems can be categorized as single-line, multi-line and omnidirectional.

RADAR
The long-range radars used by autonomous vehicles are usually millimeter wave systems that can provide centimeter accuracy in determining position and movement. These systems, known as Frequency Modulated Continuous Wave RADAR or FMCW, continuously radiate a modulated wave and use changes in phase or frequency of the reflected signal to determine distance.

VISUAL PERCEPTION
Visual perception systems attempt to mimic how humans drive by identifying objects, predicting motion and determining its effect on the direct path a vehicle must follow. Many in the industry, including visual movement leader Tesla, believe that a camera-centric approach, combined with enough data and computing power, can push artificial intelligence systems to do things previously thought impossible.

AI
At the heart of the most successful visual perception systems is the convolutional neural network or CNN. Their ability to classify objects and patterns in the environment makes them an incredibly powerful tool. As this system is exposed to real-world driving footage, either through collected images or from test vehicles, more data is collected and the cycle of humans labeling the new data and training the CNN is repeated. This allows them to measure distance as well as infer the movement of objects and the expected path of other vehicles based on the driving environment.

At the current state of technology, the fatal flaw in autonomous vehicle advancement is the pipeline through which they are trained. A typical autonomous vehicle has multiple cameras, each capturing dozens of images per second. The sheer volume of this data, which now requires human intervention and proper retraining, is now becoming a bottleneck in the overall training process.

HAZARDS
Even in the area of human-controlled driver assistance, more than 400 crashes involving automated technology have been reported to the National Highway Traffic Safety Administration in the past 11 months in 2022. In fact, several notable fatalities have occurred where sensing and decision-making systems have been identified as a contributing factor.

COUNTERPOINT
While the argument could be made that human error statistically causes many more accidents than autonomous vehicles, including the majority of driver-assisted accidents, autonomous systems usually do so in a way that could otherwise be managed by a human driver. Despite autonomous vehicles having the ability to react and make decisions faster than humans, the basis of environmental perception on which these decisions are based is so far removed from the capabilities of the average human that trust in them remains below the majority of the audience lingers.

—
SUPPORT A NEW SPIRIT ON PATREON
https://www.patreon.com/newmind

Please take the opportunity to connect and share this video with your friends and family if you find it helpful.