From: cleoabram
The journey towards self-driving cars is not a sudden switch, but a gradual progression involving different levels of autonomy [02:21:28]. This evolution has been ongoing for decades [02:35:15].
Historical Progression of Vehicle Autonomy
Early forms of driver assistance laid the groundwork for today’s autonomous vehicles:
- Mid-1900s: Basic cruise control allowed drivers to set a speed [02:38:26].
- Early 2000s: Cars began offering help in critical moments, such as emergency braking [02:43:08].
- Later advancements: More continuous support systems like lane centering or adaptive cruise control, which matches the speed of the car in front [02:48:08].
As technology advanced, the industry began to categorize the journey towards full self-driving into specific levels [03:01:31].
Standardized Levels of Self-Driving
The commonly referred-to levels of self-driving are [03:08:14]:
Level 0: Momentary Help
At this level, the car provides only momentary assistance to the driver [03:12:35].
Level 1: More Continuous Support
Vehicles offer more ongoing support for the driver [03:15:23].
Level 2: Newer Systems
This level includes highway piloting systems [03:18:03]. Most cars that are easily available for purchase today operate at Level 2, with humans still largely in charge [03:24:25][03:33:04].
Higher Levels: Transitioning Control
As levels increase, the car takes on more responsibility [03:38:15]. The ideal progression involves:
- Human Monitoring: Initially, humans would monitor the car [03:41:06].
- Full Charge in Certain Situations (Implicit Level 4): The car becomes fully in charge, but only within specific situations or geographic areas [03:45:00]. Examples include driverless taxis operating in specific cities [04:03:05].
- Car in Charge Anywhere, Anytime (Level 5): This is the ultimate goal, where the car is fully autonomous in any condition, any time, with no human intervention needed [03:50:38]. This is often referred to as the “sci-fi version” and is still some distance away [04:21:18][04:29:12].
Current Focus and Competing Approaches
The current focus in self-driving technology is not on achieving the full Level 5 immediately, but on advancing within the capabilities of full self-driving with constraints [04:36:12][04:40:48]. This involves two main divisions [05:38:39]:
Geographic and Complexity Constraints
Companies are developing solutions for:
- Less Complex Situations, Anywhere: Training vehicles to drive in areas without excessive complexity, such as highways [04:44:48]. This aligns with long-distance autonomous trucking [08:42:07]. These trucks are already being used for deliveries [09:17:51] and are expected to significantly reshape how goods are moved, similar to how an excavator differs from manual labor [09:07:05][09:34:04].
- More Complex Situations, Geographically Constrained: Training vehicles for complex environments like cities, but only within specific, predefined areas where they have been extensively trained [04:51:30]. This approach is exemplified by robotaxi services, like Cruise in San Francisco, where vehicles operate without a human driver [00:22:20][01:24:06].
Carrying People vs. Goods
Another division is whether the autonomous vehicles carry people or goods [08:31:02]:
- Cities, No People: This includes delivery robots [08:33:43].
- Highways, No People: This describes long-distance autonomous trucking [08:42:07].
- Cities, With People: This is the domain of robotaxis, which provide driverless rides in urban environments [00:06:06]. These vehicles, like the Cruise Origin, are designed without traditional driver controls such as a steering wheel or brake pedals, as drivers are not meant to take over [10:01:21][10:23:03].
Real-World Challenges and Nuances
The deployment of self-driving cars reveals unexpected complexities in the interaction between humans and machines [06:35:48]. Examples include:
- Sleeping Passengers: Driverless taxis have encountered situations where passengers fall asleep, requiring customer support to prompt them to exit the vehicle [06:47:33].
- Interactions with Law Enforcement: Incidents, such as a driverless car attempting to pull over and then moving to a safer spot after police walked away and turned off sirens, highlight the need for sophisticated detection and interaction protocols [07:16:11][07:54:19].
The Future Perspective
The current advancements indicate that self-driving cars are here, though not yet in the single, go-anywhere sci-fi form envisioned [01:29:21][01:41:40]. In the long term, humans might view driving cars as a “barbaric” practice due to the inherent risks and time consumption involved [10:47:31][11:01:03]. The future of autonomous driving is being shaped in the “messy, messy middle” of these evolving levels and competing approaches [12:00:23].