Tesla is testing Autopilot with traffic lights, stop signs and roundabouts, full self-driving ‘coming soon’
Dec 10, 2018 | by Iqtidar Ali
Elon Musk is once again leaking details of Tesla’s driver assistance software called the ‘Autopilot’ — Tesla Model S/3/X built in the last two years with Hardware 2.x will soon be able to recognize turn signals and stop signs, in addition, will be able to navigate roundabouts.
Yesterday Elon Musk took on Twitter to express his excitement about the upcoming new features of the Autopilot software (AP) — He first mentioned the latest feature ‘Navigate on Autopilot’ which was recently introduced in the AP Version 9.0 Update, he said:
If you have a Tesla built in past 2 years, definitely try Navigate on Autopilot. It will blow your mind. Automatically passes slow cars & takes highway interchanges & off-ramps.
Autopilot’s Navigate on Autopilot option basically assists in lane-change on highways and takes exits on its own if possible. Navigate on Autopilot feature has 3 modes to select from ‘Mild’, ‘Average’ and ‘Mad Max’ — at the moment the most aggressive Mad Max mode even requires you to confirm lane changes using the turn signal or cruise stalk (see Mad Max mode in action in video below on US highways).
According to Elon Musk, ‘Navigate on Autopilot’ feature is on hold in the European countries due to hurdles in getting regulatory approval, he said:
Video: Navigate on Autopilot (Mad Max mode) in Action on the United States highways
Talking about the Full Self-Driving features Elon Musk said that Tesla Autopilot software’s development versions are already testing traffic lights, stop signs & navigating the roundabouts in the following tweet:
In Elon Musk‘s time estimates ‘soon’ has no definite deadline, but this means that the software team at Tesla is aggressively pursuing the goal defined by their CEO.
In a follow-up tweet, Elon Musk clarified that “go from your garage at home to parking at work with no driver input at all (full self-driving)” option will require retrofitting the existing onboard Nvidia computer (Nvidia Drive PX2) with a new chip that Tesla has been working on which will increase the image processing power of the vehicles 10X.
With this processor upgrade, Tesla Autopilot Hardware 2.x will be called ‘Hardware 3.0’ capable of processing data from 12 ultrasonic sensors,1 radar and 8 cameras that provide a 360° vision at 10 times faster than the current rate.
Tesla CEO Elon Musk explained this in the following tweet:
Based in Silicon Valley, Tesla Inc. recently announced that Tesla owners have driven at least 1 billion miles with Autopilot engaged, what this means is Tesla neural-net has the largest amount of self-driving data compared to any vehicle manufacturer or software firm on the entire globe.
Even Google’s Waymo accumulated only ~8 million autonomous miles through their extensive testing till July 2018 (see graph below0), in actuality a lot of these autonomous miles are acquired on test tracks and limited environments but Tesla has gained their autonomous miles in real-world city roads and highways.
Starting off Q4, 2018 Tesla removed the ‘Full Self-Driving Capability’ option from their online design studio, which was priced at an additional $3,000 at the time, although Tesla says the option is still available off-menu (visit a Tesla Store to get the upgrade).
Let’s see how fast Tesla is able to roll-out the software update enabling traffic signals & stop sign recognition and navigating roundabouts, Elon Musk has recently been more accurate regarding his timelines like the Navigate on Autopilot feature, he was spot-on.
Tells us your thoughts in the comments section below, we love your ideas!
- Seeing the world through Tesla Autopilot V9 – 3D bounding boxes, human detection and beyond
- Tesla Model X avoids a side collision with Autopilot assistance
- Tesla Smart Summon is coming to Model 3, S and X cars in ~6 weeks
- Tesla Model 3 V9.0 software update, get the full release notes
- A real Tesla challenger has arrived — Waymo’s Self Driving Jaguar I-Pace
- Audi wins the license for testing self-driving cars in New York,