Last month, one of Google's autonomous vehicles was involved in a minor car crash with a bus, and for the first time since the company's AVs have been driving on public roads, this was the first car accident that was caused by Google's autonomous system rather than human error. Clearly the system is not perfect, and as the US transport secretary told the press during SXSW recently, Google's autonomous vehicles should not be considered perfect to begin with. They have a much lower rate of error compared to human drivers, but the technology isn't fully mature yet. Interestingly enough, however, following these events Google has obtained a patent called "Bus Detection for an Autonomous Vehicle" which, by the looks of it, should lower the rate of error even further.
Although the patent was awarded on March 8, it appears that Google originally filed for the patent a couple of years ago, back in March of 2014. Whether the recent events involving the aforementioned car crash have influenced the patent's release earlier this month is unknown, but whatever the case may be, the patented technology seems to aim at making Google's autonomous vehicles more aware of school buses on public roads. The system uses image recognition technology to gather data on vehicles located in the AV's vicinity; it determines the size and color of the vehicles and performs several other steps in order to figure out whether one or more of these vehicles are school busses. The system appears to be designed so that Google's autonomous vehicles would drive more cautiously in the vicinity of school buses, effectively minimizing the rate of error and risk of crashing.
It should be noted that the bus involved in last month's crash was not a school bus, so whether or not the accident could have been avoided if this particular system was already in place is unknown. Whatever the case may be, self-driving cars are not infallible, but the system will continue to improve throughout the next few years. Until then, more patents are likely to be filed and awarded, and US legislation is also bound to adopt a number of changes in order to better accommodate autonomous vehicles on public roads.