Long distance commute, traffic congestion or any daily routine that causes physical and mental stress automatically hampers hand-eye coordination, concentration and thus driving ability that results in elevated risk from road accident. That being the case for motorway operators at large, automobile manufacturers and technology companies worldwide have been testing driver-friendly methods to enhance roadway safety, from the development of various machanic engineering mechanisms all the way to today’s IoT-powered systems with the application of cameras, sensors and softwares that talk to each other via artificial intelligence (AI). What we have now is like another set of eyes have been added to broaden our perception and elevate our intelligence from the vehicle’s perspective.
You now have 13 months a year: rough guesstimate show ball park figures of an entire month freed up every year for any self-driving car urban owner from time otherwise spent stuck behind the wheel, despite the congested DVD player. With today’s technology seeming more than enough by any standard and owners eager for driverlessness (mainly for parking and cruise control), why then has widespread implementation not happen, yet?
1. Infinite feasibility … potentially happening on the roads
AI can be achieved through lessons taught or self-learning essentially as a result of basic human input. Although human beings can define speed, braking distance and traffic as well as other signs, the most difficult but very necessary task is learning human behaviors. Over the course of the next few years before driverless cars emerge to share the roads with human-driven cars, we might want to teach them what the right things are and how cars should be driven. However, who knows what would happen if AI were confronted by humans who fail to comply with traffic laws, found its route situated in a construction zone or came across hand signs or signals or body language coming from pedestrians at a zebra crossing? How were AI to react in situations such as these? In real life, there are many more incidents while a case of infinite feasibility could occur. How would AI learn about them unless it was actually running on the road, not to mention a situation where the GPS or a sensor malfunctions?
2. Cost of new technologies
One of the core technologies regarding driverless cars is a surveying (remote sensing) method called light detection and ranging (LIDAR) which – similar to radar – can create 3D representations of the target based on the vehicle’s surroundings. LIDAR used for tests or otherwise costs as much as US$ 70,000 each, equivalent to the price of a car. Despite information that the price of a modified version is merely US$ 4,000, people still doubt whether its efficiency is up to standard. At any rate, given the ever increasing number of driverless car manufacturers, LIDAR price is expected to adjust downward resembling the price of the world’s first mobile phones at US$ 3,995 (equivalent to approximately 130,000 Thai baht) at the time they were launched.
3. Weather condition: Same routes, same place but not as easy
Driverless cars are usually tested when the weather is favorable, i.e. with dry smooth road surface, sunshine and clear visibility, tests that are almost never conducted under severe or extreme weather conditions, e.g. driving at nighttime during snowstorms on slippery icy road or a case where snow accumulation hampers the sensor’s functioning or electronics equipment are frozen. At present, technology manufacturers in Israel have successfully come up with sensors that resist both hot and cold weather and can even function well in areas with thick fog or heavy rain. Thus, it is just a matter of time when these cutting edge technologies will be studied and developed for use with driverless cars.
Despite the ability to automatically give commands, AI still has to rely on the ability to link with information sources be it in the area of GPS, weather conditions, traffic conditions or system updating which requires linking through the internet, a process with a weakness enabling hackers to manage to have a control on the vehicles, listen in on a conversation or collect data concerning the trip history. While manufacturers are busy developing a software to prevent hacking, hackers are also hard at work developing their skills.
5. Acceptance on the part of consumers
The American Automobile Association revealed in 2016 that over 75 percent of drivers had no confidence in driverless cars while the latest survey conducted in January 2018 showed a better sign, albeit a high percentage still, as only 63 percent did despite the fact that over 90 percent of accidents stem mainly from human errors. Automobile developers expect that an increase in the number of driverless cars and their ever clearer presence on the roads will lead to consumer familiarity and more confidence just like in the case of someone who has never used a lift before. The first experience could be frightful. But once he has come into contact with it, he can become totally besotted by the comfort and convenience. But then, there’s also complications with the insurance: “I didn’t do it, the machine took over!”.
Photos: newsecuritybeat.org, Unsplash, Getty Image