What are autonomous cars?
Autonomous cars, also known as driverless or autonomous cars, are vehicles that can navigate on their own, sensing their surroundings without relying on any human intervention. These cars are equipped with advanced control systems that interpret sensory information to determine navigation, as well as identify obstacles or any relevant signage. However, it's important to note that a self-driving car doesn't necessarily mean it's fully robotic. There are several autonomous functions in a car that operate at different levels and each car on the market today has certain automatic systems.
Figure 1: Autonomous cars
In recent years, we have witnessed many new things introduced by some of the major players in the automobile world. From Google's autonomous car to Tesla's autopilot; Innovative technology has seemingly taken the autonomous concept to a whole new level. As we try to gauge how far we've come on this driverless car journey, here's a concise introduction to self-driving cars, covering all the relevant details of the subject.
History of driverless cars
As far as autonomous vehicle technology goes, Leonardo Da Vinci designed a cart that could move without any push or pull in the 15th century AD, when cars hadn't even been invented yet. But later in history, when cars were invented, experiments to automate cars began in the 1920s, with trails occurring in the 1950s. However, it is believed that truly autonomous cars emerged in the 1980s. In addition, all prominent companies such as Mercedes, Nissan, Audi, General Motors, Toyota and even Google have started working on preparing autonomous vehicle prototypes.
Precisely, the first step towards driverless cars was taken with the introduction of anti-lock brakes in the 1980s. The same system is used today with certain improvements and additional features like traction control and different schemes depending on the number of speed sensors and valves available for the wheels. Cars without an ABS system run the risk of locking the wheels, which can cause the car to skid. To avoid this lock-up, the driver must continue to press the brake pedal. But an ABS takes care of the pumping and prevents skidding much more efficiently than a human being ever could.
Automation levels
As mentioned above, there are pre-described levels of automation that determine the extent to which a car can be called an autonomous vehicle. Let's take a quick look at these levels.
• Level 0 (no automation): At this level, there is little or no assistance provided by the vehicle and the driver is forced to be fully concentrated on the different aspects of driving.
• Level 1 (Driver Assistance): In this classification, the vehicle offers assistance in tasks such as steering or acceleration/deceleration. While the driver is fully responsible for all driving tasks, technology helps in the form of warnings and alerts.
• Level 2 (Partial Automation): These cars use technology to study their driving environment and provide real-time feedback as the driver is required to react to it and take appropriate action.
• Level 3 (Conditional Automation): At this level, the car is capable of making certain decisions based on minor driving conditions. The driver can intervene if desired, but the car can monitor the situation very well.
• Level 4 (High Automation): In high automation, the car takes care of all important aspects of driving, regardless of the driver's response to a specific warning.
• Level 5 (Full Automation): At the highest level of automation, the car is equipped with systems that can handle all aspects of driving on all types of roads or environmental conditions. Driver inputs or intervention are not required.
How does a car drive itself?
To function autonomously, a car needs to be able to understand its driving environment, thus communicating with other elements on the road and responding to specific changes. To this end, these vehicles are equipped with a vast set of infrastructures and technologies, essentially consisting of the following components -:
• LIDAR: It uses laser beams to generate a 360-degree image of the car's surroundings.
• Cameras: They detect traffic lights and other signs and identify moving objects on the road.
• RADAR sensors: They measure the distance between the car and an obstacle or object on the road.
• Main Computer: analyzes data received from sensors and compares stored maps to assess the situation.
• GPS: It uses radars and camera feedback to detect variables and choose the best path to reach the destination
Figure 2: How an autonomous car sees its surroundings
These sensors installed in a car interpret the surroundings in the form of colored boxes and fences to identify various objects and send them to the main brain. After evaluating the situation, the necessary decisions are made by the system, whether they are executed automatically or by the driver when he receives the warning.
Autonomous driving features already in use
• Adaptive cruise control
Figure 3: Demonstration of Adaptive Cruise Control of a Toyota Model
Cars equipped with this technology measure the speed and distance of other cars using long-range radar and laser systems. This maintains a preset distance with other vehicles on the road and is often useful in heavy traffic. It ensures that when a vehicle in front of you slows down, your car also slows down and when a vehicle enters your lane, your car is a safe distance away.
• Anti-collision system
Figure 4: Collision avoidance warning
• Lane departure warning
Figure 5: Lane departure warning alert
• Blind spot monitor
Figure 6: Blind spot monitor alert shown on the car's ORVM
• Parking assistance
Figure 7: Park Assist features shown on the dashboard
The Park Assist system is equipped with sensors and cameras that capture a wide-angle view of the rear of the car and the feedback is projected onto the dashboard or instrument panel. Furthermore, it also steers the car ideally to assist with reverse parking.
Companies involved in manufacturing driverless cars
Figure 8: Interior of Google's self-driving car
Numerous car manufacturers like Tesla, Audi, Mercedes Benz and other companies including Uber and Google are passionately working on launching fully autonomous cars by the year 2021. Apparently, it is no longer a concept; Driverless cars are gradually becoming part of the real world.
• In September 2016, Uber teamed up with Volvo and launched 100 self-driving cars in Pittsburgh. Additionally, it recently announced another alliance with Daimler AG to build autonomous Mercedes-Benz models.
• Google's driverless car project began in 2009 and in 2014 a prototype was presented without steering or control pedals. In December 2016, the project was transformed into a new company called Waymo. It has already passed the testing phase and is ready to be commercialized.
• Tesla Motors, the company introduced the autopilot feature (Hardware 1) in 2014, which was recently updated (Hardware 2) in October 2016. As far as the software package is concerned, Tesla released version 7.0 in October 2016. 2015, followed by version 7.1 in December 2015 and finally Firmware 8.0 in 2016.
• BMW recently announced its partnership with Intel and Mobileye, which aims to launch 40 autonomous BMWs by the end of this year. It tested the driverless feature in 2011 with the 330i model, followed by the 6 Series Gran Coupe model in 2014. Almost all high-end BMW cars are equipped with intelligent parking, driving and vision systems.
• Audi has teamed up with graphics card manufacturer Nvidia to develop level 4 autonomous cars by 2020. The Q7 SUV model was equipped with AI technology for experimental purposes, so it learned to drive itself in just 4 days.
• Other car manufacturers such as Honda, Nissan, Ford and Fiat are also taking steps in their respective projects, independently or in alliance with other companies, to develop fully autonomous cars.
Safety Issues Related to Driverless Cars
From the moment companies first had the idea of developing driverless cars, there has been an endless debate about autonomous technology. On the one hand, there are several benefits, such as making it easier for the elderly, sick or children to travel, as well as reducing the number of accidents. On the other hand, it still seems like a bad idea to rely on a machine that drives autonomously on a busy road and it could also lead to an increase in the number of cars, which means more traffic.
Figure 9: Safety concerns for self-driving cars
One of the most important aspects of using a driverless car is the safety of the occupants and others on the road. Even though technology has reduced the number of accidents that occur due to human error; There are some incidents that lead to accidents and casualties. Mentioned below are some driverless car accidents.
Tesla
In June 2016, a Tesla Model S failed to stop in front of an 18-wheeler, resulting in an accident that sent the car off the road, hitting a fence and then a power pole. The Model S occupant named Joshua Brown died in the crash and there were serious concerns regarding the safety of the Tesla's autopilot system. However, it was later stated by automotive safety regulators that no fault was found in the autopilot system.
In February 2016, a self-driving Google Lexus model collided with a bus while trying to turn right. Fortunately, no one was injured, but the car's radar was destroyed and the tire was punctured. According to Google reports, their cars have been in 14 minor accidents and it was only once that the car was at fault.
Uber
A computer-controlled Uber was caught on camera as it passed a traffic light after four seconds of the light turning red. However, the company claimed that the driver was at fault, after which he was suspended.
NuTonomia
One of NuTonomy's (Singapore-based startup) autonomous taxis collided with a truck, damaging the car's right bumper and leaving a dent in the side of the truck. It was later discovered that this was a software failure and services were resumed after the defect was corrected.
AeroMobil
AeroMobil is a startup that has developed a “flying car” that functions both as a car and as a flying plane. In May 2015, the flying car crashed during a test flight while the occupants suffered minor injuries. The company subsequently carried out several successful tests and will now soon accept pre-orders for the car.
Driverless cars: how far are we from conventional autonomous driving technology?
Figure 10: Nissan IDS Concept
Developing a driverless car and marketing a driverless car are two completely different things. There are many moral, ethical and legal issues that need to be addressed before making autonomous driving technology mainstream. Currently, states such as Michigan, California, Florida, Nevada and the District of Columbia have legalized the testing of self-driving cars, while others have rejected or considered the option. Especially after Tesla's fatal accident, laws related to self-driving cars are being carefully reviewed.
Although many automakers are already promising to bring fully autonomous cars by the year 2020, we are still a long way from the practicality of having a level 5 autonomous car. One of the most basic reasons for this gap between plans and reality is that who will be the guilty if something goes wrong? You can't blame the car because it's a robot and you can't blame the driver because he's not controlling the car.
Adequate legislation and clarity of roles and responsibilities are necessary so that loss of life and property can be avoided. Despite so many innovations and developments, there are still some gaps that need to be addressed. For example, some self-driving cars cannot detect cyclists and pedestrians on the road.
So the bottom line is that you can never be sure about the efficiency and behavior of a machine, but you should not stop innovating. No matter how far we are from the idea of driverless cars being commonplace, we are still on the path to a future that hopes to make a big change; this has already happened to some extent.
You can read our Blog and Article section for more topics on electronic engineering, industry and technology.
(tagsToTranslate)pic