Evolving Before Our Eyes: Self Driving Car Technology

Self driving car technology has taken off, as manufacturers and other players such as Google scramble to be the first to revolutionize the industry. The race to develop the first fully autonomous vehicle is on, and it seems that the development driverless car technology is coming in from all facets, from mobile networks to the diligent work in university labs.

What is Self Driving Car Technology?

self driving car technology
Self driving car technology allows for self driving cars to work autonomously. The technology itself consists of various technologies combined together, including lidar, radar, GPS, odometry and motion sensors, computer vision, and cameras. The technology works based on sensory input, thereby allowing the vehicle to navigate itself through various conditions and environments.

Early Beginnings of Self Driving Car Technology

 

Self driving car technology can be dated as far back as 1925 with Houdina Radio Control. Electrical engineer and founder of Houdina Radio Control Co., Francis P. Houdina, worked to make his dream of a radio-operated vehicle a reality. Using a 1926 Chandler, he installed a transmitting antenna that could be controlled and operated by a second vehicle that would follow it. The antenna received radio impulses delivered by the second vehicle. These signals would then be sent to a circuit of breakers that controlled and operated the electric motors. The small electric motors would then dictate the movement of the car. This 1926 Chandler was known as the linrrican Wonder, and made its driving debut on the streets of New York City. Achen Motor took Houdina’s idea and created Phantom Auto, which made its driving debut that December in Milwaukee. Six years later the Phantom was back on the road for a demonstration in Fredericksburg, Virginia.

 

Building on Houdina’s idea of radio-operated vehicles, Norman Bel Geddes depicted a radio-controlled electric car in his Futurama exhibit at the 1939 World’s Fair. The idea was that the vehicle would then be moved by electromagnetic fields from circuits which were embedded in the road.

 

By 1953 RCA Labs was successful in building a miniature vehicle, which was guided and operated through a pattern of wires on the lab’s floor. It inspired a Nebraska Department of Roads traffic engineer, Leland M. Hancock, and state engineer, L. N. Ress, to bring this idea to life on a real roadway. By 1958 a life size system was ready for demonstration. Together with the State of Nebraska, RCA Labs demonstrated its wire guided vehicle on a 400-foot stretch of road. That stretch of road had circuits buried under the pavement, with lights along the edge of the road. These circuits would send impulses to help guide the car along the road.

 

The project was a collaboration with General Motors. The project used two vehicles and added special radio receivers, as well as audio and visual warning devices. Together, it allowed for controlled simulation of automatic steering, braking, and accelerating. During the 1950s and 1960s, General Motors also introduced more in the way of self driving technology with a series of Firebirds, which offered an automated guidance system controlled by a joystick.

 

The 1960s brought on an array of futuristic ideas with driverless car technology. For instance, William Bertelsen’s Aeromobile 35B, an early version of the hovercraft. It was believed that it would lead to self-driving, hovering cars that could reach speeds of up to 1,500 mph. Meanwhile, in the United Kingdom there was testing being done on a driverless Citroen DS, which used magnetic cables beneath the road. This researched carried on into the 1970s, specializing in cruise control devices in conjunction with the magnetic cables. However, the project lost its funding by the mid-1970s.

 

Universities Lend a Hand in Developing Autonomous Technology

 

Universities have had a hand in the development of self driving car technology, with Ohio State University being one of the first in 1960. Ohio State University’s Communication and Control Systems Lab decided to begin developing driverless cars, with the aim of the project was to control the vehicles by electronic devices under the roadway. Heading the project was Dr. Robert L. Cosgriff, who believed that by 1981 the university’s system would be ready for public roads. Stanford joined the game in 1979, with the first computer-controlled autonomous vehicle called the Stanford Cart. Built by Hans Moravec, the Artificial Intelligence Laboratory Cart was remote controlled by a large computer, and able to navigate through obstacle courses.

 

Moving into the 1980s, self driving car technology caught the attention of the manufacturer, Mercedes-Benz. Ernst Dickmanns and his team from the Bundeswehr University Munich designed the van, which reached a speed of 39 mph during a roadway demonstration. He went on to conduct the Eureka PROMETHEUS Project, which paved the way for driverless cars. Collaborating with Daimler-Benz, success was achieved in 1994 when two twin robot vehicles debuted true autonomous driving, demonstrating their capacity for convoy driving, automatic motion tracking of other vehicles, as well as left and right lane changes with autonomous passing of other cars in Paris.

 

Meanwhile, thanks to self driving car technology developments by the University of Maryland, Carnegie Mellon University, the Environmental Research Institute of Michigan, Martin Marietta and SRI International, the United States began working on an autonomous land vehicle (ALV). The project marked the first time lidar, computer vision and autonomous robotic control were utilized. Come the end of the decade, in 1989, Carnegie Mellon University began using neural networks to maneuver and control driverless cars. This was the foundation for self driving car technology, and how autonomous vehicles would be controlled. Artificial neural networks were inspired by how the brain transmits and receives signals. Each of these “neurons” have numeric weights, which can be tuned based on its experience and thereby create adaptive neural nets that are able to learn.

 

Carnegie Mellon University’s Navlab project did its USA tour, known as the “No Hands Across America” tour. Driving across the country 98.2% autonomously was made possible using the revolutionary RALPH computer program. RALPH (Rapidly Adapting Lateral Position Handler) utilized video images to help figure out the vehicle’s location and location of the road ahead to maintain the optimized steering direction to keep the vehicle on the road. In fact, RALPH’s neural networks controlled the steering wheel. The 1.8% of driver control was an accumulation of researchers handling throttle and brake.

 

The ARGO Project was launched in 1996 at the University of Parma, using a modified Lancia Thema that would follow the painted road lines. The modified car was demonstrated over the course of a 1,200 mile journey. Modifications included two inexpensive, black-and-white video cameras and utilized stereoscopic vision algorithms to maintain a better sense of the environment. It was operated in automatic mode for 94% of the trip.

 

Self Driving Car Technology Meets Manufacturers

 

Self driving car technology finally made its way to the manufacturer in 1995. Using saccadic computer vision and transputers that react in real time, the autonomous S-Class by Mercedes-Benz was re-engineered. It was demonstrated driving completely autonomous for 95% of the journey, on a 990 mile trek from Munich to Copenhagen. While long distances of the vehicle’s autonomous driving were not emphasized, the vehicle still managed drive up to 98 miles without human intervention. By August 2013, Mercedes-Benz S-class featured close-to-production stereo cameras and radars, which allowed it to drive fully autonomously for close to 100 km. With the success of these results, the 2014 Mercedes S-Class offered various self driving options, which was embedded in safety features. DISTRONIC PLUS with Steering Assist. This is a semi-autonomous, using the vehicle ahead for orientation. Meanwhile, PRE-SAFE® Brake allows for autonomous braking and can even be paired with pedestrian detection. Adaptive Highbeam Assist PLUS autonomously adjusts the vehicle’s LED headlamps for oncoming traffic and/or vehicles ahead.

 

In 1998, Toyota transformed the industry with mass producing self driving car technology by introducing Adaptive Cruise Control. By 2013, Toyota’s driverless car technology was ready for demonstration with a partially self driving car. The vehicle featured a plethora of sensors and communication systems.

 

BMW began developing its self driving car technology in 2005. Collaborating with the Chinese search company Baidu, the companies plan to produce a self driving car that will be launched in China. With successful preliminary testing, the self driving car will be fitted with a reconfigured BMW 3 Series. This partnership began in late 2014. BMW’s focus will be on the automotive aspects, while Baidu will be responsible for providing AutoBrain software. This software will provide artificial intelligence, image comprehension, voice recognition, automated driving maps, positioning, detection and more.

 

By 2011 General Motors had entered the driverless car technology race with its EN-V (Electric Networked Vehicle), which was designed to be a self-driving vehicle for urban settings. The following year Volkswagen developed and began testing its TAP system. TAP, short for Temporary Autopilot, was designed for vehicles to drive up to 80 mph without human intervention.

 

Ford Motor Company has also done a plethora of research into autonomous technology and its communication systems. The manufacturer announced it was ready to begin testing its self driving cars, enrolling in the California Autonomous Vehicle Testing Program in 2015. In December 2015, Ford Motor Company and Google announced that they would be collaborating in making self driving vehicles.

 

Nissan’s Japan plant installed its self driving car technology in the 2014 Nissan Leaf as part of a demonstration, demonstrated at the 2013 Nissan 360 event in California. In the same year, the Nissan Leaf was fitted with Nissan’s Advanced Driver Assistance System and began public road testing in Japan. Part of the Nissan’s family was the 2014 Infiniti Q50, which was released in late 2013. The vehicle utilized cameras, radar and various other technologies to enable lane-keeping, collision avoidance and adaptive cruise control.

 

Tesla Motors joined the autonomous trend in 2014 when it announced its first version of AutoPilot. Meanwhile, the manufacturer’s 2015 Model S vehicles that are equipped with AutoPilot offer an array of autonomous features, including autonomous steering, braking and speed limit adjustments, and autonomous parking. The system has the ability for software updates, allowing its technology improvements.

 

As of 2015, major manufacturers have started to produce various autonomous safety features such as collision avoidance, adaptive cruise control, lane departure warning, blind spot monitoring, electronic stability control, automatic parking, 360 degree radar, and GPS. Meanwhile, other companies such as Apple, Google and Uber try to make a play in the race to produce the first fully autonomous vehicle. Uber, the company known for its taxi app, uses GPS technology to help users find the nearest Uber taxi. It is rumoured that the company has recruited experts from the field to collaborate on creating a self driving taxi company to coincide with its app. Apple has been working on what it calls Project Titan. While it has been a rather secretive project, rumours have surfaced that the company is scouting for testing space.

 

 

A Prime Example: Google’s Self Driving Car Technology and How it Works

 

The real player that stepped up to the game was Google, which largely popularized the development of self driving car technology with its self driving car. In 2009 the company began privately developing its vehicles. By 2012, Google’s self driving car was ready for testing. Its test route spanned 14 miles in the city of Las Vegas. The test was successful; however, roundabouts, railroad crossings and school zones were not included in the test. In 2014 Google made the announcement that the powerhouse company would unveil 100 self driving car prototypes from the powerhouse’s secret X lab. By mid-2015, Google’s self driving cars had been in 14 minor accidents. Each accident is claimed to be caused by human error in other vehicles, with 11 of the accidents resulting in rear-end collisions. Since 2009, close to 2 million miles worth of testing had been logged.

 

Although Google’s vehicles are still considered prototypes, the vehicles have been designed to be used with no steering wheel or pedals; however, they’re currently equipped with removable pedals and steering wheel. With these features, humans can still intervene if needed.

 

These prototypes utilize various detection technologies, which include sonar devices, stereo cameras, lasers, and radar. These technologies combine together to offer different ranges and fields of view. Google utilized lidar technology to enable an accurate map of the vehicle’s surroundings. Radar technology is utilized to adjust acceleration and braking in real time. This idea is similar to adaptive cruise control, although it is assumed by many enthusiasts that Google has paired its radar technology with sonar systems to improve this feature. Google’s software integrates each respective technology’s data to continuously adapt to road and environment conditions.

 

Big Data and Self Driving Cars

 

The term big data accounts for very large sets of data which are analyzed computationally. These sets of data reveal patterns, trends and associations, allowing self driving cars to adapt to their current environment. Therefore, since self driving cars rely on big data collected by its sensors to properly monitor its surroundings and environment, there are many opportunities for data providers. Each manufacturer has established its own software, and/or outsourced various aspects of it. A great example of this could be BMW’s collaboration with Baidu and AutoBrain, as previously mentioned. However, there are a vast amount of players that are shaping the industry, with other companies such as ArcGIS which can help with geo-spatial analytics such as traffic patterns; whereas SAS Enterprise Miner allows for the vehicle to learn driving routes.

 

Mobile Networks Come Into Play

 

After mobile phone technology grew, automakers began to focus around connecting drivers to their accustomed lifestyle. The technology grew from there, utilizing Advanced Driver System Technology. These systems are computerized, utilizing many of autonomous technologies to provide a safer ride. For instance, it helps with user data and analytics, transition data, field data, mapping, and more. Features such as adaptive cruise control and blind spot monitoring are just a few examples. Manufacturers such as BMW believe that mobile networks are the secret to bettering self driving car technology. The company predicts that 5G mobile networks will be the key to reliability, believing that it will allow vehicles to communicate with one another in order to be successful in the self driving car game.

 

What’s Next for Driverless Car Technology?

 

Provided manufacturers can overcome certain technical barriers, self driving cars could become a reality in the near future. As new players like Apple and Uber enter the game, it will be interesting to see the evolution of self driving car technology. With driverless car technology advancements and more data becoming available, autonomous driving features will become more prevalent. It is a matter of overcoming barriers such as sensors recognizing certain weather patterns and public opinion. Needless to say, certain features such as adaptive cruise control prove that self driving car technology is evolving and certain tasks, such as blind spot monitoring, can be left up to the vehicle. In the meantime, more real road testing in different conditions is needed to better develop and overcome these technical hurdles in order to advance self driving car technology.