Here we are, in the 21st century. The technology is being innovated so fast that we’d probably be a multi-planet species by the end of next decade. Artificial Intelligence is one of the major breakthroughs in the field of computing and technology. Our lives have begun to see the substantial effect AI has made in our lives. And with the ongoing year it is expected to see a great deal of integration of AI into the devices that we use every day. Digital assistants that manage our home entertainment system or keep track of our activities are already making our everyday job simper. Much of this has been made possible by the fact that finally the technology has caught up with the expectations since the coining of the term “Artificial Intelligence” in 1960’s.
Artificial Intelligence has taken up a major stride in innovation in the Autonomous Vehicle sector. Private firms have put in so much capital in developing the next generation of vehicles that employ Artificial Intelligence to assist driverless functioning. These vehicles have been equipped with huge data collecting and processing capabilities and fitted with sensors to gather data effectively. The data that these sensors gather are processed by the computer and involves self learning algorithms that, simply put, are powered by AI. The main focus that the automakers are pressing on now is to properly manage the data that is gathered by the sensors installed in the car. Radar, Cameras, ultrasound, accelerometers and gyroscopes are some sensors to be named that obtain parameters correlating between the real world scenarios and the models developed by the AI which have relatively limitless number of variables in the equation and are of great complexity. For the true development of autonomous driving of level 4 and 5, the system has to be functional in all weather conditions.
Cars that are autonomous function on two levels: Driver Assist and Fully Autonomous. Driver Assist technology is basically the co-pilot in your car that monitors dozens of sensors, processes the gathered data and then maps out the outcome of the relatively possible scenarios that are most likely to occur on the road ahead. It then alerts the driver, or takes control of the vehicle in predefined emergency situations to avoid a mishap.
Cars that are fully autonomous theoretically do not need a driver. With the data generated by the sensors placed in the car, the Autonomous Driving Platform based in the cloud, processes the information using the algorithms and tries to make accurate decisions based on the data fed to it by the car. The ADP then passes on the required manoeuvre for the car to act upon. The reality is however a little different. Manufacturers have been running tests on autonomous vehicles for quite some time now. In multiple events, the cars were driven across for over 10,000,000 miles without a driver in the driver’s seat but an employee was present to take control when required. The human had to take control of the vehicle on different occasions, more than once.
So we are not quite there yet. But the recent developments and the interests shown by tech giants into developing driverless technology may have good news very soon. The automotive AI market is reported to be valued at $54 billion in 2019 and is expected to exceed $556.67 billion by 2026 with a cumulative annual growth rate of 39.47% according to Allied Market Research estimates. This is probably the reason why every multimillion dollar company wants to tap the huge potential this market has to offer. All of this adds to big business. Let us look at the companies that are leading the pack in the development of Autonomous Vehicles.
Tesla is probably a household name in the automotive sector. And that seems to be the company’s goal in Autonomous Vehicle market.
On April 12, 2019 Tesla announced its Model 3 with standard Autopilot. The pricing for the models have increased. But if you look at the features that have been introduced, you probably wouldn’t feel the pinch on your pockets.
What changed from the previous models is the fact that Tesla bundled all the new models: Model S, Model X and Model 3, with a standard Autopilot feature. The new model comes with an enhanced on-board computer that is 40 times more powerful than the computers on the older models. The newer processor is said to be better equipped to run the Tesla-developed neural net for vision, radar and sonar and feeds more accurate inputs for a better outcome by the algorithm. Tesla claims that the data obtained by numerous tests indicate a lower percentage of accidents when Autopilot was turned on. Customers have tremendously supported the fact that Tesla’s Autopilot improves the overall driving experience.
The Tesla official website says that the new Autopilot system can match traffic speeds, keep within a lane, can automatically change lanes without driver interference, can park itself near a spot and be summoned from your garage. Well it still does require driver supervision and isn’t completely autonomous. Tesla has managed successfully to manufacture vehicles on an autonomy level somewhere between 2 and 3. Also, the Tesla cars have all the hardware needed to operate fully autonomously on levels 4 and 5 in almost all the circumstances, we just have to wait for the achievement of reliable and better algorithms and computing technology.
Alphabet, the parent company of Google, also happens to be the parent company of Waymo. Waymo seems to be Google’s long term plan to dominate autonomous vehicle market. Reports indicate that Waymo has been conducting tests in Phoenix for over 2 years now which started in 2009 and is aiming at launching a public taxi conveyance service. The original plan was to introduce Waymo One, and the commercial cab service could be hailed from a part of the city of Phoenix. But the expansion of the service is slower than expected and the company awaits legislation that permits legit usage of driverless cars.
Since Google dominates the information over the internet and has monopoly over a large share of data, it is relatively easier for Google to train its AI model. The working principle behind Waymo’s vehicle is the same as Tesla’s. The software on the computer crunches and processes data it is fed from the lidars, radars, high-res camera and GPS to name a few sensors and performs complex algorithmic calculations to produce control signals which operate the vehicle effectively. Powerful deep learning algorithms in the cloud server can, most often than not, accurately predict what the objects in the vehicle’s path may do. The most important aspect of AI powered automotive application is that it keeps on learning and adjusting its algorithms to predict a better result every time.
Waymo works exactly like ordinary cab hailing services, except that it does not have a driver. You can hail the cab using an application on your phone. It does not feel different at all from calling an Uber.
Mercedes-Benz plans to debut its first self driving system in its new Mercedes-Benz 2020 S Class model. Diamler’s head of research, Ola Kaellenius, told reporters at Paris auto show in October 2018 that Mercedes is on the verge of jumping to level-3 autonomy. Autonomous level 3 driving does not need the driver to be engaged with the driving environment. At level 2, like the Tesla models, the driver has a necessity to keep track of the surroundings and monitor the drive at all times. The level 3 offers hassle free commute and there’s no need for the driver to keep alert. Although the driver needs to be ready to take control of the vehicle in certain cases as and when alerted by the system.
Diamler is working on its level 2 hands off Driver Assistance Technology that would be incorporated widely into other models. On a collaborative approach with Bosch, it is working on developing a fully autonomous vehicle that would be deployed in commercial cab hailing services.
The new generation Mercedes –Benz S Class is reported to implement the new MBUX infotainment system that is capable of recognising the gestures made by the driver or the passengers. The current plans are to incorporate the AI powered Autonomous Driver Assist technology in just the new S Class that is due in 2020. It would slowly be filtered down to lower range models.
At Las Vegas in 2019, Nvidia announced its commercially available level-2 automated driving system. The Autopilot system by Nvidia, incorporates multiple AI technologies that will allow Driver Assistance Technology to function as a supervised self-driving vehicle to take the road by the year 2020. While it is obvious that it will take an immensely great amount of computational power and intense data crunching in combination with sophisticated software to deliver a better equipped self-driving car that is available on the road today. In 2018, in a report concluded by Reuters, about 300 companies were using Nvidia Drive Computing technology to aid their autonomous vehicles project.
Nvidia is already involved with Uber in providing assistance to its self driving vehicles. Volkswagen also is reported to rely on Nvidia manufactured Drive IX Software development kit to aid its AI feature and improve the processing in its autonomous cars.
As a chipmaker, Nvidia’s technology has been constantly aiding and improving the way automakers approach Autonomous Driving Technology. It sure is a game changer in providing commercially available, powerful solutions to carmakers who can then scale their approach to Autonomous vehicles.