Tesla Is Proving The Entire Industry Wrong With Its Bold New Self-Driving Tech

Flamboyant entrepreneurs like Elon Musk are hard to come by in this excessively politically correct world. Unlike other billionaires, he doesn’t rely on “experts” to build his persona. The guy is a straight shooter who never shies away from speaking his mind on any platform. He is the bloke who flies cargo for NASA and is planning to colonize mars after-all. He is rich, flawed, and ambitious. And that’s what puts his electric car company Tesla at the forefront of innovation. Tesla single-handedly made the electrical cars cool. And now the same company is our best bet for truly autonomous cars. The idea of a robot car comes from over a century ago. Trials of an Automated Driving Systems (ADS) go back to 1920s, but we still don’t have a fully automated car. However, if everything goes according to the plan, Musk has promised to put a million driver-less cars on the road by the end of 2020. And he is not talking about the current cruise control or autopilot mode that can handle cars in certain situation. Elon is promising full autonomy capable of taking you from point A to B with zero human interference. So, what is it that separates Tesla’s ADS from other companies such as GM, Ford, Audi, BMW, Toyota, Badu, Uber, and Google’s Waymo?

A Bold New Approach

There’s a stark difference between how the current market leader Google Waymo and Tesla approach the driver-less tech. To make its car drive around in a specific area, Google and other brands including Uber scan the entire place to build an intricate 3D model of the locality. It’s like building a track for your hot wheels. On the other hand, Tesla’s cars are more like explorers. These autonomous cars are meant to be street smart. Instead of limited to a designated playground, Tesla’s cars can survive on most roads. The system is designed to help you travel at optimum speed while avoiding collisions using data coming from multiple sensors.

Image for representational purpose.

Waymo is currently testing cars in Austin, Atlanta, Detroit, and many parts of California. To add any new location to this list, Google has to digitally recreate the roads including curbs, lane markers, traffic lights, and even sidewalks to the accuracy of centimeter. Tesla doesn’t have to go through such tedious process. Its cars are built to work on any road with proper infrastructure. Due to this inherent philosophy difference Tesla cars are better at reacting to real word conditions. This bold new approach from Tesla is going to have a huge impact in the autonomous car technology.

Doing Away With LIDAR

Light Detection and Ranging (LIDAR) is one of the most accurate techniques for visualizing any environment. The technology involves continuously firing the pulses of laser light, and then measuring the time it takes to return to the sensors. Since the speed of light is constant, the differences in laser return times and wavelengths provide us with data for detailed mapping.

Back in 1971, NASA used this technology to study the Lunar surface during the Apollo 15 mission. While you read this article, there’s a good chance that a satellite in the low Earth orbit is scanning our planet for accurate topography. So, when Elon Musk called such sophisticated LIDAR instruments appendices, it was natural for journalists to throw a hissy fit. However, a quick glance at the large ugly domes at the top of Waymo, Uber, or Baidu’s autonomous cars, and you will know that he does have a point. And there are many more reasons why Musk isn’t a fan of LIDAR.

To get the 360-degree view of the surroundings, LIDAR equipment needs to constantly spin and beam lasers in all directions. As we all know, moving parts are expensive to manufacturer and maintain. Moreover, there’s a higher chance of failure too. Although LIDAR offers astounding accuracy, it requires good weather to function properly. The technology just doesn’t work properly during fog, dust, or heavy rains. As a workaround, Waymo cars are also loaded with cameras and other sensors. But the biggest problem with this technology is the exorbitant price of the equipment. A single LIDAR unit from Waymo costs around a whopping $75,000. With equipment costing more than many luxury cars, it is no surprise that nobody is mass producing autonomous cars. It has been the biggest hurdle in the development of fully autonomous cars.

Tesla seems to have solved the issue by getting rid of the LIDAR system. The company believes that LIDAR is an overkill for the autonomous car technology. Instead of light ranging, Tesla relies on its robust yet affordable new system. Tesla’s solution comprises of RADAR, cameras, and a bunch of sensors. As opposed to Waymo’s tech that costs more than the car itself, Tesla’s autonomous car system will add $3000 to $4000 to the car’s price. That’s something most people wouldn’t mind paying for.

How Does The New System Work?

Tesla autonomous car system comprises of a front RADAR sensor, 12 ultra-sonic sensors, and eight cameras. The cameras are strategically placed to track everything around the car. The data coming from radio ranging and optical sensors, is then processed by the on-board FSD (full self-driving) computer. These new computers are built by Tesla and are powered by chips fabricated by Samsung. It can perform 144 trillion operations per second (TOPS). As a result, you get a robust system of vision and ultrasonic sensors that can work during the fog, dust, and heavy rains.

Sure, RADAR (RAdio Detection And Ranging) scans are not as accurate as LIDAR. While RADAR merely detects an object and its size, LIDAR can distinguish between a dog and a fire hydrant. But Tesla has solved this issue by complementing RADAR with vision sensors. And the results are quite stunning, check out the amount of data Tesla’s autonomous system manages to capture in just a few seconds.

A recent independent study from the Cornell University concluded that a pair of cheap cameras can produce 3D data as precise as what laser scanners are able to generate. All this at a fraction of a cost. Suddenly, Elon’s claim that any car with LIDAR tech is “doomed” doesn’t sound too scandalous, does it?

Using machine learning, Tesla’s autonomous cars also learn to anticipate how the road will extend at the end of the corner. Tesla claims that its ADS system can even navigate streets that don’t even have lane markings. It is not even afraid of densely packed roads either. Thanks to the current autopilot feature with partial autonomy, Tesla has trained its system with billions of kilometers of experience. This data will help every Tesla car to better react to dynamic road conditions.

So, there’s a very good chance that Tesla will be the first company to mass produce fully autonomous cars by the end of 2020. Just don’t get your hopes too high in India, as no technology in the world is advance enough to deal with the millions of speed breakers and trillions of potholes.

The post Tesla Is Proving The Entire Industry Wrong With Its Bold New Self-Driving Tech appeared first on MySmartPrice.

from MySmartPrice http://bit.ly/2UHxXpt