Self-driving cars have been automobile industry’s buzzword from past several years when Google started testing its own car project privately, Waymo, back in 2009. Since then, several original equipment manufacturers (OEMs) like BMW, General Motors, Ford, Mercedes Benz have funded significant resources towards building a fully autonomous car. A fully autonomous vehicle will be ready by 2020, and by 2040, it will hold 95% of the world’s automobile market.

What is a self-driving fully autonomous vehicle?

Waymo firefly1
Waymo firefly1

An autonomous vehicle can sense its surroundings, and guide itself, without any human interaction. The sensors can detect traffic, distance from adjacent automobiles, stop signals to predict an optimal path for its movement, with the help of Machine Learning and IoT technologies. While a fully-autonomous car is still a concept, partially automated vehicles already exist in today’s market.

National Highway Traffic Safety Administration (NHTSA) defines an automated vehicle based on the following 5 classifications:

  • Level 0: No Automation – A human driver controls all the driving functionalities. No automated assist.
  • Level 1: Driver Assistance – The driver maintains control over the driving functionalities, but the vehicle provides certain drive assistance like cruise control (steering/Lane keep assist).

Example: 2018 Toyota Corolla (Toyota Safety Sense), 2018 Nissan Sentra (Intelligent Cruise Control).

  • Level 2: Partial automation – The vehicle assist in steering, maintaining distance in traffic or keeping it in the center of a lane, but the driver must always be ready to take control.

Example: Audi Traffic Jam Assist, Cadillac Super Cruise, Tesla Autopilot, Volvo Pilot Assist

  • Level 3: Conditional Automation – Under ideal circumstances, the car can manage most of the driving functions by itself. The human driver is still needed, though they can keep their hands off the steering wheel, and legs off the pedal, but should be ready if the conditions fall below ideal.

Example: 2019 Audi A8(Traffic Jam Pilot)

  • Level 4: High Automation – Under known and ideal circumstances, the vehicle can operate fully, without human interaction. Road and environments use cases must be known before, for it to operate correctly.

Example: Level 4 Waymo cars

  • Level 5: Full Automation – This is a fully-autonomous vehicle, that can drive and maneuver through any road and environment, without any human input, other than entering the destination. Since everything will be controlled by the system itself, there will be no need for pedals and steering wheel.
Audi A8 interior
Audi A8 with Level 3 automation

History of Self-driving cars

While the self-driving car has come to reality in recent years, its inception dates back to 1956, when a version of GM’s Firebird II was built with a guidance system to avoid car accidents. Chrysler Imperial incorporated the cruise control for the first time in 1958.

Stanford Cart: First self-driving car
Stanford Cart: First self-driving car

In 1979, Stanford University built a cart robot (The Stanford Cart) to traverse through obstacles, without any human interaction and was the first self-driving car. VaMP and VITA-2, robot cars by Bundeswehr Universitat, traveled 600 miles safely through traffic, in 1994.

In 2004, the U.S. Defense Advanced Research Projects Administration (DARPA) challenged top research teams to build an autonomous car to travel 142 miles through the Mojave Desert, for 1 million dollars. The challenge was a failure, with the winning team, only covering 8 miles, before setting into a fire.

Google self -driving car project, Waymo, started back in 2009. The testing was conducted privately, and over 800,000 miles was covered by these cars, by 2018. By 2013, Ford, Mercedes Benz, BMW, Tesla, and Uber joined the race to build their own, autonomous vehicle.

Waymo self-driving cars
Waymo self-driving cars

 

The 2019 Audi A8 is expected to be the first Level 3 autonomous vehicle, with Traffic Jam Pilot. Nvidia announced its self-driving chip, Xavier, that is built with artificial capabilities, and partnered with Volkswagen to build self-driving cars.

Tesla developed a software, that is installed in all Tesla models, to provide full self-driving capabilities. The car has 360-degree coverage with the help of 8 surrounding cameras and can monitor up to 250 feet. It also has 12 ultrasound sensors to detect soft and hard objects, and a forward-facing radar to see through heavy rain and fog.

tesla autopilot
Tesla Autopilot

Importance of Self-driving cars

Almost 1.3 million people die every year due to road accidents, caused by poor driving skills, traffic, speeding i.e Human error and emotions. Self-driving cars remove this possibility, by following logistical real-time data with zero distractions, rather than emotions, making it safer than Level 0 autonomous vehicles. Level 5 autonomous car can reduce road fatalities by 90%. Driverless vehicles will cut cost on fuel consumption, and delivery to remote areas will be less prone to accidents.

Self-driving car features and Components

  1. Video cameras — to detect traffic lights, signals, road signs, pedestrians with 360-degree coverage.
  2. Radars (forward, rear) — to see through heavy rain and fog, and calculate the distance from nearby automobiles or objects.
  3. Lidar — a rotatable radar on the roof, to scan and create a 3D map of the area.
  4. Ultrasonic sensors — to detect hard and soft objects
  5. Position sensors — to find vehicle position on the map
  6. GPS navigation — to get the accurate positioning of the automobile.
  7. Central computer — To process the sensor data, and perform actions based on it.
Self driving car features
Self-driving car features

Role of Internet Of Thing(IoT) in self-driving cars

For a car to be fully-automated, it needs to sense its environment, road conditions, and distance to adjacent automobiles and objects. Autonomous car technology is integrated with IoT sensors, to record the data, and send to the server, where it processes and sends back the information to car controls. IoT enables ‘connected vehicles’ through V2X Communication (vehicle-to-everything communication), that allows the vehicles and smart sensors to share data with each other, and process it to build a real-time network grid map, to provide a better insight on how to handle the drive, and avoid obstacles. Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) forms the base for V2X Communication, that enables connectivity.

Today’s connected vehicles require real-time access to high-speed, reliable, low-latency connectivity to analyze the data, and process it. This leads to the risk, during limited or no internet connection, leading to late or no response to impending obstacles. This can be overcome with the help of Edge computing and Edge AI for IoT technologies. Edge computing and AI collects and processes the data on the edge; vehicle in this case, before transmitting to the central server. This allows the system to interpret and act on the data, even if there is no connectivity, making it safer for both the user and the location. Smart traffic systems sensors add on to the real-time data for the vehicle to work on; thus, providing much more sophisticated navigation. Such connected system between vehicles and its surroundings will help in contemplating the condition of the road and traffic and adjusting to it. Smart Parking along with car sensors enables it to park in the least required space. Smart transportation system enables autonomous trucks to stay connected with each other, and help in reaching the destination the earliest.

Volvo XC90 - Park Assist Pilot
Volvo XC90 – Park Assist Pilot

IoT business models will help OEMs to create flexible vehicle system architecture with full autonomous car technology, to allow fast-to-market productions, with increased efficiency, and leading-edge technology. Testing it through Digital Twin will let them test all possible use cases virtually, enabling them to release a safer, reliable and zero-defect autonomous vehicle.

Conclusion

While there is still a social dilemma of autonomous cars, due to several accidents, that have happened, but the automotive industries are working resiliently to create 100% reliable software, and pass legislative and regulatory issues. With a gradual acceptance from society and a better and reliable system, fully autonomous self-driving cars will be available by 2020.

 


Media sources:

Self driving car technology pin