Editorial Credit: Nadezda Murmakova / Shutterstock.com
What’s the first thing you think about when buying a car? I want wheels that are sexy, speedy, but most importantly, safe. Car wrecks kill around 40,000 people per year in the US, and those numbers may be rising. Low gas prices have more people out on the roads, while bigger and better smartphones have them more distracted. Companies like Tesla are looking to reduce driving deaths with advanced driver assistance systems (ADAS). Their ADAS enabled cars have already reduced fatalities and will continue to save lives as they become more advanced. Eventually, with the aid of multi-sensor fusion and artificial intelligence (AI), Tesla will produce self-driving cars that reduce car crash incidents by up to 90%. At that point, I’m not sure I’ll feel safe driving anything else.
Back in 2013, Tesla’s Model S received the NHTSA’s highest safety rating of 5 stars. The Model S’s batteries and electric motors take up less space than a conventional gas tank and engine, allowing for larger crumple zones. These crumple zones save lives by absorbing impact energy during crashes. One Tesla even saved the life of its driver after falling off a 500-foot cliff. Clearly, they make safe cars, but the safest car is one that never crashes. That’s why Tesla is enabling its cars with ADAS.
Proving itself to be a true pioneer, Tesla has introduced comprehensive ADAS in its cars called Autopilot. This system includes features like autosteering and adaptive cruise control. Sure these things are nice, but do they actually make the car safer? Well, the NHSTA recently released the findings of its federal investigation into Tesla’s safety features. These results include data showing that the Autosteer component of Tesla’s Autopilot reduced wrecks by 40%. That is statistically significant.
Tesla announced last year that all of its cars will now be equipped with an even more advanced Autopilot. This system will use a combination of even more sensors to enable better features and make driving a breeze.
Eventually, Autopilot will make Tesla cars totally hands free. Editorial Credit: Aleksandra Suzi / Shutterstock.com
Tesla’s ADAS Advancements
If Tesla’s autopilot is already reducing accidents by 40%, how much lower could they go? Some experts estimate that high-level ADAS enabled and fully autonomous vehicles could reduce crashes by up to 90%. In order to reach this goal, Tesla is incorporating more sensors into its systems and is eventually looking to go fully autonomous.
Multi-sensor fusion is one of the buzzwords for self-driving cars. Basically, it means using a combination of several different sensors to give the car the complete picture of its environment. In Tesla’s case, these sensors are a combination of passive visual, ultrasonic, and radar. Radar gives a car long range vision in all kinds of weather, passive visual provides detailed images, and ultrasonic is used for close range object detection. By integrating all of these sensors the car can read signs, park itself, and detect traffic all around it. Some have criticized Tesla for its exclusion of LIDAR arrays from its cars. However, these sensors are still too expensive (around $70,000 each) to incorporate in vehicles. Though, I hope that as the price drops, the experts at Tesla might consider incorporating them in later models.
Multi-sensor fusion and AI will make Tesla cars extremely safe.
In addition to the extra sensors, Teslas are getting more processing power. Their website states that their new processors will be 40 times faster than the old ones. The next generation of ADAS needs to be able to interpret what it sees and then act based on that information. This requires a lot of processing power. With a more powerful processor, the new Autopilot will be able to make split second decisions that save lives. In fact, Tesla is so sure of their Autopilot system, they’re considering packaging their cars with lifetime insurance policies. This is one reason why experts are saying that the new Model 3 will be one of the safest cars out there. A more powerful processor is also necessary for the next logical step in cars, AI and fully autonomous driving.
For full autonomy to happen Tesla’s cars will need to act intelligently. It just so happens that Elon Musk is now dabbling in AI, as well as self driving cars and space travel. I doubt this a coincidence as Tesla is already incorporating machine learning into their Autopilot system. This will allow Teslas to learn from their drivers how to react appropriately to strange new situations. Is there a particularly hairy blind corner ahead? The driver will likely brake, and the car will learn that it needs to slow down on this part of the road. As more and more Teslas hit the road, they’ll all learn more quickly. Eventually, millions of Tesla drivers will be teaching the entire fleet how to react to random problems.
Tesla is already making safe cars, but its new autopilot system will make them even more reliable. Soon driving fatalities will be a thing of the past and we’ll have moved on to reducing things like fender benders. The only way that level of safety can be reached is through multiple sensor fusion and AI. Luckily, Tesla is pushing forward in both categories.
As cars start using more sensors and incorporate machine learning, their software will become increasingly complex. That’s where you come in. We’ll need you to program cutting edge ADAS features like Tesla’s autopilot. To do that, you’ll need next generation software development tools. That’s where TASKING comes in. With things like standalone debuggers and great static analysis features, TASKING’s products will help you in each stage of the development process.
Have more questions about ADAS? Call an expert at TASKING.