Self-Driving Cars, Powered by Humans.....Coming Soon!

As a former automotive engineer at Ford Motor Company, this has been an exciting week. First, Tesla announced that they would launch electric and self-driving pick-up and commercial trucks. It is widely believed that these vehicles will be equipped with self-driving technology as current model Tesla’s have this feature. Toward the end of the week, it was discovered that Apple had received approval to begin testing self-driving vehicle technology by the California DMV. This means that nearly every major automaker along with several leading technology companies are investing time, money & human resources into figuring out how to deliver a self-driving transportation network. So, what is it about the next decade that has some of the best companies in the world trying to figure out how to deliver a self-driving transportation solution? We believe there are a few key changes that help us understand why now. 

One of the keys to solving any solution using technology is a deep understanding of your constraints. Over, the last several years, companies like Google, Apple, Garmin & Uber have spent billions of dollars to map out the streets, roadways and highways. These development efforts help form the boundaries for self-driving vehicles.  In addition, as customers use these mapping technologies, information can be fed back to the algorithm to see when there are deviations in outcome and utilized to improve mapping accuracy.

While it is essential to understand the boundary conditions of the roadways, transportation and technology solutions also require the capability to monitor real-time changes in conditions. The biggest set of variables that need to be accounted for is all the other vehicles on the roadway. Over the years, we have seen the evolution of the capability of these sensors. The first iterations of these sensors capability was with blind-spot monitoring and blind-spot detection.  This evolved to lane departure capabilities.  The sensors that self-driving vehicles currently include are LIDAR (uses laser beams to create 3D images of objects so that the vehicle can calculate the distance from these objects by seeing how long it takes the light beam to return), high-definition camera (detect and record pedestrians, road signs & other vehicles), radar (bounces radio waves off surrounding objects to calculate the distance between by seeing how long it takes the beam to return), GPS (accurately identify the precise position of the vehicle by seeing how long it takes the signal to return from the satellite – remember the mapping technology from above), ultrasonic sensors (these use sound waves to detect objects within very short distances – primarily used for self-parking). One of the big reasons why this dizzying array of sensors is able to be used now is that through continued innovation, the cost of these sensors has been able to dramatically come down over the years.

When you think about the amount of real-time processing power that is needed to ingest all the data from the sensors, it is mind-boggling. By some estimates, self-driving vehicles could generate 1GB of data per second. One of the biggest proposed benefits of self-driving vehicles is that they will improve the safety by making better decisions and a faster rate than humans. For several decades, the technology industry has invested heavily in improving the memory and chips for computers. In many ways, the pace of improvements has outpaced the ability for everyday users to exercise the capability of these powerful machines. However, self-driving vehicles are a perfect application for these powerful computers to be utilized in a way that is beneficial for society.

In order to pull all these capabilities together into a useful product, the most important resources are people. Software developers and hardware engineers who are working on the self-driving transportation network could save lives and help reduce injuries. In 2015, 38,300 people were killed and 4.4 million people were injured by vehicle crashes in the United States. Globally, the number of annual fatalities is 1.3M and 20-50 million people injured in automobile accidents. This is a big opportunity to develop a solution to save lives and reverse a disturbing trend.