Posted in:

How Are Self-Driving Cars Developed and Programmed? 

Automakers have toyed around with creating autonomous vehicles since the first car hit the road. But it is only in the last two decades that autonomous vehicle technology has made tremendous strides, with all major automakers competing to have the most advanced autonomous driving system. 

While it is illegal to have a fully autonomous vehicle on American roads, several automakers have developed cars with features that allow vehicles to move from point A to point B with little to no human intervention. If you are fascinated with self-driving cars or even intend to buy one, you could be curious to learn how they are developed or programmed.

What if a Self-Driving Car Gets Into an Accident?

While some cars like Tesla have advanced autonomous capabilities, they are still considered advanced driver assist technologies, meaning that the driver is responsible for controlling the vehicle. Also, they will be liable for resulting damages.

There are situations where the automaker could be liable. For example, if a system malfunction is the cause of the accident. 

If you are unsure of what to do in the wake of a car accident involving an autonomous vehicle, talk to an attorney with experience in autonomous car accidents. 

The Society of Automotive Engineers (SAE) Autonomous Car Classification

SAE classifies autonomous car capabilities into five categories, including;  

  • Level one/driver assistance

Level one autonomy includes features that offer driver assistance, such as automated accelerating, braking, and minor steering, such as lane assist features. Most new vehicles have this level of autonomy.

  • Level two/partial automation

Level two autonomy allows the vehicle to have two or more autonomous functions work simultaneously, such as cruise control and autonomous braking. 

  • Level three automation/conditional automation

Conditional autonomous vehicles can operate with limited to no human intervention but only in some conditions, such as a highway. But the driver must remain on guard and take over in case of malfunction.

  • Level four/high autonomous

Vehicles under this category can self-drive under almost all conditions but will still require little human intervention.

  • Level five/ fully autonomous

This autonomy allows the vehicle to operate without human intervention in all conditions and scenarios. Fully autonomous vehicles may not even feature a steering wheel.

How Are Self-Driving Cars Programmed?

Many car automakers are racing to build the most sophisticated autonomous driver technology. These automakers include the traditional car makers such as Ford, Nissan, Audi, and Mercedes and new entrants such as Tesla, Uber, and Google Waymo. While each company has a different approach to developing this technology, the concept is the same. 

Self-driving technologies adopted by automakers must perform three critical actions in controlling the vehicle to replace humans; perceive its environment, compute the data collected, and control the vehicle by steering, accelerating, and braking. 

Autonomous vehicles utilize cameras, Light Imaging Detection and Ranging (LIDAR), and Radio Detection and Ranging (RADAR) to gather data and perceive their environment. The data is computed through computer code lines, which send signals to the vehicles’ control electronics to initiate braking, steering, stopping, or accelerating, depending on how the vehicle AI interprets the gathered data. 

Semi Autonomous vehicles also do a greater deal of integrating GPS into their systems. However, they go further than GPS to compensate for GPS inaccuracies allowing the vehicles to achieve better accuracy through machine learning. 

Shortcomings in Self-Driving Cars

Different automakers take different approaches to self-driving technology. For example, one automaker could use RADAR instead of LIDAR or cameras. The problem is that every option has its good side and a not-so-good side. 

For example, LIDAR technology works best in ideal weather but not so much in dusty or foggy weather. The same may be true with cameras, although they offer the best resolutions and gather precise data in ideal conditions. Radar, on the other hand, works in almost any environment but is limited in identifying specific objects and obstacles.