Top Three Challenges Facing Autonomous Vehicles
Top Three Challenges Facing Autonomous Vehicles
Automation is popular, even to the point of self-driving cars. However, automating everything is not without its challenges. Here are the biggest challenges to automated vehicles, including radar interference, accident liability, and navigating crowds.
Join the DZone community and get the full member experience.Join For Free
If you have seen Disney-Pixar’s "Wall-E," you will surely remember the self-driving robots and cleaning robots. In one scene, we see the cleaning robot in a dilemma when it sees dirt on the floor. The dilemma is because the dirt is not on the designated route of the robot. After much animated debate with itself, the robot exits the standard path and starts following the dirt, cleaning it.
While the scenarios presented in the film may be too far-fetched, another reality is shaping up on our roads in the form of autonomous cars. Over the years, the area of autonomous vehicles has been a favourite topic of research. Recently, this area has gained a lot of interest due to the work done by Google for its self-driving car.
Google’s Self-driving Cars
Google’s ‘self-driving’ cars (presently) are standard cars equipped with multiple sensors, have logged hundreds of hours and many miles of tests. In all tests, while there has been a person physically present behind the steering wheel (as mandated by US vehicle driving rules and just in case manual intervention is needed), in most situations, the car has driven itself. Initially, a modified Toyota Prius was used for testing, but a recent prototype from Google shows a design that does not have a steering wheel or driving controls.
Each ‘self-driving’ car uses many sensors, to enable it to drive on the road. Each car is mounted with a video camera that detects traffic lights and moving objects; a rotating sensor on the roof that scans the surroundings and creates a dynamic, three-dimensional map of the environment; a position estimator that measures lateral movements and determines the vehicle’s position on the map; and multiple (four) distance sensors (radars) that measure distances to various obstacles.
In situations where the Google car has been involved in accidents, post-facto analysis has indicated that these accidents were caused due to the fault of other cars, driven by humans. Over the years, the technology of self-driving has made dramatic progress. But this technology still has a long way to go before such vehicles become commonplace on the street.
Autonomous vehicles will have to cross many hurdles before they gain wide acceptance. While their acceptance in the developed countries will be easier, they will still face challenges even though drivers have more discipline and follow driving rules. In developing countries, the challenges will be monumental, primarily due to the number of vehicles on the road as well as the non-compliance to road rules by many drivers. While many of the challenges will be overcome with the passage of time, we believe that three of these challenges– elaborated below –are critical to the success of self-driving cars.
Challenge 1: Radar Interference
The Google autonomous car uses lasers and radar for navigation. The lasers are mounted on the roof of the vehicle while the radar sensors (around four in number) are mounted on the body of the vehicle. The principle of radar works by detecting reflections of radio waves from surrounding objects. When on the road, a car will continuously emit radio frequency waves, which get reflected from the surrounding cars and other objects near the road. The time taken for the reflection is measured in order to calculate the distance between the car and the object. Appropriate action is then taken based on the radar readings.
We believe that this principle of reflection may fall prey to interference from other similar vehicles if the density of such radar-enabled vehicles (on the road) is very large. It is likely that radar waves from one car may interfere with radar waves from another car in the same locality. Though the principle works very well for airplanes, the number of airplanes in the sky is quite limited. Additionally, the distance between airplanes is fairly large.
When this technology is used for hundreds of vehicles on the road, will a car be able to distinguish between its own (reflected) signal and the signal (reflected or transmitted) from another vehicle? Even if multiple radio frequencies are available for radar, this frequency range is unlikely to be insufficient for all the vehicles manufactured. A comparative example that may immediately come to mind is that of mobile telephony. Mobile telephony also uses radio signals for communication and hundreds of mobile devices per cell / tracking station / tracking tower are being used simultaneously. Mobile networks use microwaves for communication and the density of mobile devices is usually very high.
The biggest difference between a mobile signal and radar is that radar works on the principle of reflection while mobile devices encode additional information and send it to the tower. In reality, a mobile device never sends the same signal that it receives. It sends different information. In other words, each mobile device is also a transmitter in its own right. The information sent by mobile devices is used to distinguish between multiple mobile devices.
The principle of radar works on the principle of reflection. There is no transmitter involved in the transaction. The sender emits a signal and then waits for the wave to be reflected from an object. It is important to note that the signal received by the sender is essentially the same as the one it had transmitted/generated. A self-driving car uses radar, in order to find out the distance between itself and other objects and/or vehicles on the road. Even if we assume that other autonomous vehicles send information to each other like mobile devices, it will not be possible to calculate the distance between the vehicles unless this information is supported by accurate GPS (Global Positioning System) coordinates. But what of objects like walls and trees, which do not have any transmission mechanism to inform self-driving vehicles of their position?
Challenge 2: Accident Liability
In the case of self-driving cars, one of the most important aspects that will delay their wide adoption is relating to the question of accident liability. Who is liable for accidents caused by a self-driving car? In the famous Toyota case, a few drivers faced problems when their Toyota Prius accelerated even when the drivers pressed the brake pedal. While some of the accidents were minor, a few were fatal for the drivers. While some drivers faced such issues because the floor carpet got pulled into the pedal hole cut-out, some drivers faced problems due to the software used to control the behaviour of the car. During the investigation, the appointed expert had many comments about the quality of the software itself, in addition to the software development process used for development and testing.
In the case of autonomous cars, the software will be the main component that will drive the car and will make all the important decisions. While the initial designs have a person physically placed behind the steering wheel, newer designs showcased by Google, do not have a dashboard and a steering wheel! In such designs, where the car does not have any controls like a steering wheel, a brake pedal, an accelerator pedal, how is the person in the car supposed to control the car in case of an untoward incident? Additionally, due to the nature of autonomous cars, the occupants will mostly be in a relaxed state and may not be paying close attention to the traffic conditions. In situations where their attention is needed, by the time they need to act, it may be too late to avert the situation.
In the case of an accident, who is going to be held liable for damages? Today, the driver of the vehicle is supposed to pay the damages. But, in an autonomous car, if the driver does not have control over the car, how can we hold her responsible for the actions of the automobile? If the car is manufactured by Google, will Google be held liable for the damage caused?
Challenge 3: Crowd Navigation
People crossing the road does not present a serious issue in developed countries (in most cities) but is a serious problem in crowded cities and most developing countries. We believe that crowds will present significant problems for autonomous cars as well. While is perfectly acceptable to wait for people to cross the road at a zebra crossing, will the autonomous car keep waiting even when the signal turns green and people continue to cross the road? If this occurs in cities like Mumbai, an autonomous car will keep probably waiting for an eternity.
The development of self-driving / autonomous cars /vehicles has been an area of active research for many years. Recently, a lot of interest has been generated by Google’s self-driving cars, which have logged many thousands of miles in various scenarios – city driving as well as highway driving, to name a few. While there have been a few accidents, the cause of these accidents have been the humans driving other vehicles, which have collided with the self-driving car. A self-driving car by itself has never been involved in an accident initiated by it. While the technology behind self-driving cars seems unbeatable; it is yet to stand up to many real world challenges. One of the challenges is that of driving where proper roads do not exist. Another challenge lies with the performance of these vehicles in extreme weather conditions like heavy rain and/or heavy snow.
Given that the technologies being used will mature over time, it is only a matter of time, where the performance of self-driving cars will surpass the expectations of any vehicle in the city. Even with the maturity of the technology behind self-driving cars, we believe that the wider acceptance of such vehicles needs to address three important issues, namely sensing technology interference, accident liability responsibility and crowd navigation.
Opinions expressed by DZone contributors are their own.