How Driverless Cars Work—And Why They Sometimes Fail

DAYTONA BEACH, Florida. In the search for safer driving experiences, many experts believe that driverless cars will deliver. Researchers believe that driverless cars will ultimately make fewer mistakes than their human counterparts, and that driverless technology might be able to ultimately reduce city traffic by making driving more efficient. Yet, it is important to remember that driverless technology is still being tested and explored. While the technology shows immense promise, there are still many areas where driverless technology must be improved before it can be made available for the general public’s use. In terms of public opinion, driverless cars are either feared or embraced, or there is some mixture of both. But how does driverless technology really work, and it is really safe? When it fails, why does it fail? These questions are important for the general public to consider and also important for personal injury lawyers to consider as the technology becomes more ubiquitous.

According to the Telegraph, driverless cars detect their surroundings through radar sensors, cameras, and lidar sensors. Each of these sensors plays a different role in helping a driverless vehicle detect its environment. For example, radar sensors can detect other vehicles nearby, while cameras can detect pedestrians, bikers, and road signs. Lidar sensors can detect the edge of the road. A central computer in the vehicle processes all this information and then makes “decisions” for the car. However, these sensors can fail if a camera fails to detect a pedestrian or bicyclist in time. If a sign is covered with graffiti, it’s possible for a camera sensor to misread the information. And lidar sensors can only work so well if a highway lane ends, or if construction causes a lane to cross into a wall or other obstacle.

According to Wired, modern autonomous vehicle radars are not sophisticated enough to tell the car’s computer whether there’s a bicyclist on the road or if the object on the road is another object altogether, but radar can detect motion. This failure of discrepancy can be a problem though, because if a self-driving car weights evading inanimate objects less heavily than objects carrying passengers, the misidentification of a bicyclist can have major consequences for the computer’s accident evasion system.

Wired also reports that camera systems have their own limitations, too. Just as human eyes struggle under glare and in poor weather, so do camera systems.

Finally, machine learning systems are also subject to their own flaws and mistakes. Machine learning is only as good as the data fed into the system, and we have seen situations where a machine can work well in 99% of situations and fail miserably in 1%. It’s that one percent that is concerning, because this is where accidents can be tragic.

Another question many personal injury lawyers are asking about this new technology is where liability will lie when injures and accidents do occur. Does it lie with the programmer? The car manufacturer? The company that manufactures the car’s video cameras or computers? These questions are likely to be raised as more self-driving vehicles hit the road and as inevitably, more accidents occur.

Until then, humans need to do what humans have always done behind the wheel—which is take steps to prevent accidents. This means putting down your cell phone, slowing down, and paying attention to the road. The Graham Law Group, L.L.C. are personal injury lawyers in Daytona Beach, Florida who are closely watching as more self-driving car prototypes hit the road. Until then, if you’ve been in a car accident involving another driver, you may have the right to seek damages.

Do you have questions about a recent car accident? Contact the Graham Law Group, L.L.C. today to learn more and to protect your rights. can connect you with a qualified lawyer in your area who may be able to help you seek the compensation and justice you may deserve.


By | 11:48 am | Categories: Legal News | 0 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *