Image Credits: VanderWolf Images/shutterstock.com
Self-driving cars were once the subject of science fiction, but in a little over a decade McKinsey predicts that autonomous cars will account for around 15% of all new cars sold. The future of the automotive industry is arriving, technology has evolved enough to support the sensors and complex algorithms that are used to drive vehicles without human intervention, using machine learning to integrate incoming data about the surroundings, from GPS, radar, and laser light/LIDAR.
Data from both external and internal sources are processed by the car to allow it to make autonomous decisions about how to conduct the journey.
The commercialization of autonomous vehicles carries numerous benefits, from less traffic and potential increases to safety to helping drivers enjoy more free time and boosting their health by eliminating the health risks related to driving, such as increased blood pressure and anxiety. However, the automotive industry faces several major challenges with commercializing autonomous vehicles.
Factors Surrounding Self-Driving Cars
When we consider a future where self-driving cars are the norm, safety is one of the first and most important factors that come to mind. Multiple studies have sought consumer opinions on these vehicles, and results have shown that anywhere from a half to two-thirds of Americans would feel scared to travel in a self-driving car. This presents a major barrier to the adoption of autonomous vehicles, if the consumer is afraid of using the product then their popularity will not grow. Public opinion has worsened recently, with the two widely publicized fatal accidents that happened with partially autonomous vehicles. The automotive industry is faced with raising the vehicles’ reputation by proving beyond doubt that they are safe. Because in fact, the cars can increase road safety, but consumers need to be shown this.
Therefore, devising thorough safety testing for self-driving vehicles is paramount. This is a complex task in comparison to testing conventional vehicles. A thorough test of a driverless vehicle requires both test drives as well as computer simulations and mathematical modeling to test how the vehicle responds in a variety of scenarios and environment. Testing becomes even more challenging with the nature of the machine learning that powers the vehicles. Analysis cannot delve into the reasons why certain behaviors manifested as outcomes. What’s positive is that medical testing often works in a similar way, with experts not always being certain on how the medication may be working, but the outcomes show that it is.
Next, the cornerstone to this technology is its reliance on complex software, forcing the automotive industry to put itself in the hands of outside industries. Meaning the sector is relying on outside help to power their goals. Fortunately, investment in this area is snowballing, with a predicted annual growth rate for the global software market for autonomous cars to be 76.1% from 2017 to 2021. Further to needing the expertise of software companies, the automotive industry is also facing competition from these same companies. The software industry will be creating the code to make the cars work, they have the ability to go up against the car manufacturers to create their own self-driving vehicles. To thrive in this environment, the big industry players are forming alliances now to ensure they are kept in the loop.
Finally, cybersecurity is another major factor for the automotive industry to consider ahead of commercializing self-driving vehicles. A regular modern car has as many as 100 million lines of code to make all its features work, in an autonomous vehicle the complexity of code is significantly increased, and with this comes increased risk from hackers. Famously, two researchers recently demonstrated the relative simplicity of hacking into a car’s system and its impact. They remotely hacked into a Jeep Cherokee’s connection, stopping it on the highway. In further demonstrations, they were able to manipulate the car’s brakes, steering wheel, and speed. Fortunately, the industry is already remedying future cyber-attack risks by developing machine learning systems within the cars that can learn to protect themselves against hackers, and alert riders of imminent risk.
Before the widespread adoption of autonomous vehicles, the automotive industry must overcome major challenges relating to consumer perception of safety, the guarantee of safety, reliance on the software industry and competition with this industry, and the risk of hackers. Fortunately, the industry is sitting in a beneficial position, having already made moves to address these issues.
Sources and Further Reading