Google Driverless Car - Self Driving Car

From its invention onwards, cars have always ruled the roads. A major percentage of passenger commutation happens in cars. Car industry has grown from lakhs to uncountable figures. In the meantime, driving has turned out to be a passion that has no barriers of age. In everyone’s life, there has been a day; rushing through destinations, having all those mind calculations until that awful sound of your car hitting reaches your ears even without your mind knowing that accident. In these situations, nothing less than driver’s mind is the real bad guy. Year’s back there was no solution for such a misplaced mind but from early 2000s onwards Google’s engineering team have started working on a mind blowing project of bringing  ‘Driverless Car’ to a reality. It all begun with 2005 DARPA Grand Challenge, where a team from Stanford University led by Dr.Thrun came up with a robotic vehicle, Stanley, and bagged the first position winning $2 million as prize from the US Department of Defense. Dr. Thrun led the Google team which consists of 15 engineers who work on this dream project.

The Self driving car which included a fleet of cars equipped by the project team , which included six Toyota Prius, three Lexus RX450h and an Audi TT making it a total of ten cars. The team also consisted of drivers with excellent driving records, who will occupy the driving seat accompanied by an engineer in the co-driver seat. According to U.S. state laws, driverless cars are not allowed on roads. Google’s efforts brought about a change in law, allowing the permission for autonomous cars to move around in the U.S. state of Nevada by June 29, 2011. Within a year,Google acquired the first license for driverless autonomous car, a Toyota Prius equipped with Google’s driverless technology. This trend continued with states like Florida and California permitting the testing of autonomous driverless cars on public roads by May 2012. Google came up with a new prototype of driverless car with neither steering wheel nor pedals by May 28, 2014.

Earlier Prototype of Google Driverless Car-Self Driving Car

Earlier Prototype of Google Driverless Car

To begin with, Google’s project team worked on a group of ten cars which included 6 Toyota Prius, 3 Lexus RX450h and an Audi TT. Self driving car integrates Google Maps with artificial intelligence software and various hardware sensors including LIDAR, Video Camera, Position Estimator and Distance Sensors. The driver’s seat is occupied by drivers having excellent track records and the co-driver seat is occupied by Google engineer. This car has travelled through the famous Lombard Street in San Francisco which is known for its steep hairpin turns and also through the city traffic.The car drives at a speed that is programmed with the map and the vehicle-to-vehicle distance is maintained by help of sensors build around the vehicle. Anytime human can control over the self driving car by turning the wheel or stepping on the brake, which is very similar to cruise control feature found in luxury cars. It has covered around 700,000 miles in and around California which also includes busy streets and hairpin turns of San Francisco.

Working of Google Driverless Car: Saving Lives and Energy

Google Driverless Car prototype

As already stated, self driving car  integrates many of the Google applications to run efficiently. It works with Google Maps – to provide the minute information about the road, Hardware Sensors – To provide information about the real time environment conditions, Artificial Intelligence – Real time decisions are taken with help of this software.  Google Maps use the Global positioning system or GPS for proper tracking of the location and directions. Hardware sensors gives the real time informations to the software so as to make the car undergo suitable changes in its speed, direction etc. Artificial Intelligence turns out to be the key part with combining all information and controlling the car through Google built software.

 

Google Maps:

Google Maps

GPS (Global Positioning System) is a system developed by U.S military which later on has changed the overall way of reaching out for a destination. It is a navigation system based on satellites to provide ‘the accurate’ information about the preferred location. In Google Driverless Cars, Google Maps interacts with GPS and provides accurate location as well as it acts as the database. The speed limits are integrated with the maps which control the car to the restricted speeds depending on the nature of the roads and terrains. It also provide information to the car about the upcoming intersections and nearby collisions, if any. Traffic report availability will help the car to control its speed, which is also available with the Maps. In overall, Maps will provide the directions for the car to move around without any human interventions.

Hardware Sensors:

Environment is so dynamic that real time updates has to be given to the Self Driving Car to achieve maximum efficiency. This is achieved with the help of hardware sensors build around the car. Sensors equipped with the car attempts in creating a fully observable environment and track the car through all hurdles. The main sensors used are LIDAR, Video Camera, Position Estimator, RADAR and Distance Sensor.

Components in Google driverless car

LiDAR:-

LiDAR

Google driverless car as a self driving car uses LiDAR or Light Detection And Ranging, that is a technology which uses optical remote sensing to measure the distance information of a distant target with the help of the properties of scattered light.  LiDAR is the heart of the system. This is similar to the Radar technology that uses radio waves and the distance to a target is obtained by calculating the time delay that occur during the pulse transmission and the detection of reflected signal. For imaging of objects, LiDAR uses UV, near-infrared or visible light. Materials that are targeted include non-metallic objects, rain, clouds, rocks etc. A laser-beam that is narrow in nature is enough to map the available physical features with high quality resolution. The components that make up a LiDAR include laser, scanner and optics, photo-detector & receiver electronics and position & navigation systems. Laser of bandwidths in the range of 600-1000nm are seen common in applications that are non-scientific and also shorter pulses provide better target resolution. Lasers are inexpensive but to make it eye-safe the maximum power range is limited as it is easily focused and absorbed by the user’s eye. In scanner and optics, the speed taken for the scanning of images affects the developing of the same and also the angular resolution as well as the range detection is affected by the optic choices. Return signals can be collected using a beam splitter or hole mirror. In photo-detector and receiver electronics equips a reliable photo-detector technology and a receiver. Solid state photo-detectors or photomultipliers are photo-detector technologies used in LiDARs. Also the receiver sensitivity is an important parameter which needs to be balanced for the LiDAR design. Position and navigation systems include sensors that are placed on platforms such as satellites or cars need instrumentation to calculate the orientation and absolute position of the sensor. These devices consist of GPS receivers and Inertial Measurement Unit (IMU).

Video Camera:-

Electronic motion pictures are acquired using video camera. Here, in self driving car, video camera is used to capture traffic light. The information about traffic light is used to control the car on the roads especially its behavior in a traffic island. It also keeps out an eye on the pedestrians, other motorists, cyclists and other obstacles. Video cameras used here offers 45 and 90 degree field of view. Line markings are read by video cameras when compared to LiDAR.

Distance Sensor:-

A distance sensor is used in Google Driverless Car to detect the presence of nearby cars or obstacles and inform about the same to the required software so as to control the car accordingly. The sensor emits electromagnetic waves and checks the changes in the return signal. The kind of sensors used change with the type of targets. These types of sensors have long functional life and are highly reliable due to lack of physical contact between the target and the sensor and also due to the absence of mechanical parts. Short and long range radars are used here to cover a maximum range of 150 yards. There are four automotive distance sensors that are used in these cars, that is, one in the rear side and three in the front. This technology already exists in luxury cars with the help of intelligent cruise control.

Position Estimator:-

The position estimator in self driving car is meant for determining the vehicle’s location and also tracks down its movements. Here, rear wheel has an ultrasonic sensor mounted on it to monitor the movements of the car and this information is automatically updated so as to update the position of the vehicle on the map. Orientation sensor is also equipped with this car to provide the accurate orientation.

Artificial Intelligence:-

View from the Google Driverless Car

                                                                    Outside view of  Self Driving Car

Artificial Intelligence (AI) is a kind of intelligence showed by custom-build software or machines. Google Maps and the hardware sensors send the data acquired with the help of various devices to the AI. The AI is the key component of Google driverless car that determines how fast the car should accelerate, determines the moment to slow down/ stop and determines when to take over the steering control of the wheel. The ultimate goal of AI is to take the car to the destined location safely and legally. Processor employed in this car cross-checks all the information obtained from different sensors to make different objects, obstacles differentiated accurately and make driving through them an easy task. The software developed for this is the sole proprietary product of Google.

Recent Prototype of Google Driverless Car

Recent Prototype of Google Driverless Car

A brand new prototype of self driving car has been unveiled by Google on 27 May 2014, which is the only car available in the market with no steering wheel or pedals. With the launch of first driverless electric car, this is a bigger step to Google’s 5 years long self-driving car project. This car consists of two seats and a space enough to carry a small amount of luggage. The driverless car transits in and around California. The purpose of this car is for the commutation of two passengers between source and destination without the intervention of user. Car gets summoned with the help of a smartphone for pick up from a location to a set destination. As already said, steering wheel & pedals are not present and there is no manual control option in the self driving car. It just consists of a start button and a red emergency stop button. There is a screen in front of the passengers showing the weather, current speed and a countdown animation. The messages are displayed on this screen.

The car has been designed by Google from scratch, since, there is no prototype as such for driverless car or self driving car , starting with the implementation of sensors and interconnecting them using a frame followed by the addition of a cabin without blocking the sensors view or creating blind spots and finally constructing the body shell. The Google Driverless car is an electric motor powered with a range of around 100 miles which uses a combination of sensors as well as software to spot out its location in the real world with the help of highly precise and accurate Google maps. A Global Positioning System (GPS) used for navigation purposes is used to get the location of the Google driverless car which is then monitored using radar, lasers and cameras to get accurate picture of the world in 360 degrees around the car. The software built by Google is efficient enough to recognize obstacles, people, traffic lights, cars and unpredictable hazards. It is capable of detecting collisions, road works and navigates through them safely.

Construction of Google Driverless Car

The newer prototype of Google Driverless car consists of more sensors and is able to see further range i.e., up to 600ft in all respective directions, which is larger in number when compared to previous prototype from Google based on Lexus and Toyota cars.

This Google Driverless Car is limited to 25mph, providing a restriction to certain roads with minimum kinetic energy taking into account the impact of crash if one happens. It has a flexible windscreen and a foam bumper, which is designed keeping in mind a collision with a person so as to absorb energy. As safety measure, seat belts are provided, along with Google’s fault-tolerant architecture for both braking and steering and also red emergency button for stopping the car with passenger hitting the same at any time. Google driverless car has come with a driving style that seems to be defensive and considerate taking into account the feedback from the previous versions to improve the safety of passengers as well as the remaining road users. Compared to new version, the older version driverless car had a problem of blind spots that made sensor’s visibility low.

Google expects to make this version of driverless car road-ready by the beginning of next year but the testing process would take another two years. As far as current expectations, self driving car will take another five years to become mature to come up with a non-prototype product and as per the news, Google is trying to seek partnership with favorite car brands like Toyota, Ford etc., which is yet to be confirmed. Even the legislation in countries need to be looked upon before coming up with driverless cars on to public roads, which is also under work.

Advantages

  • GPS can be used to identify the current location of the car.
  • By reducing the distance between the cars, the roadway capacity can be increased.
  • Cars avoid accidents with quick reaction to real world changes on the road.
  • Due to increased road capacity, traffic flow can be managed.

Disadvantages 

  • In case of using internet with minimum security, there is a high rate of chance of hackers breaking through the system and changing the route which causes serious problems.
  • In any case of system failure, which includes the main sensors, the car can cause severe accidents.
  • Software built for driverless car or self driving car is expected to be an expensive one.

Conclusion

Google Driverless Car, a project led by Google’s 15 member engineer team along with excellent drivers and a fleet of cars is on the way to bring into reality a technology that was just in the dreams of any tech wizard. With the help of Google Maps, Hardware Sensors and Artificial Intelligence , this driverless car will definitely hit the roads and change the whole travelling concepts. Let us hope Self Driving Car will change a mere science fiction to a science fact.

Refer Embedded Systems in Automobiles too and know the innovative trends in automobiles.....