Imagine a day when your car drives you to work and then finds and parks itself into a spot while you focus on starting off your meetings for the day. While you’re hard at work, your car spends its day making more money for you by doubling-up as a self-driving Uber. And when you’re finally done for the day, it’s waiting right outside, ready to drive you home as you catch up on news and prepare for the next item on your social calendar.
Based on Tesla’s announcement on October 19, that day might just be right around the corner.
The auto manufacturer announced that all of its current production cars, including the Model 3, will now be equipped with hardware capable of making the car fully autonomous. Tesla plans to incorporate this feature with a view to improve safety, reduce operational costs and help customers get the most out of their vehicle.
How is this different from Autopilot?
Tesla has been moving steadily towards making its cars fully autonomous, the most significant steps being the addition of a radar, a forward looking camera and ultrasonic sensors. The Tesla Autopilot update that was released in October last year provided the brain for the system. The Autopilot system aimed at improving safety, incorporating features like lane-changing (with the flip of an indicator switch), collision avoidance and parallel parking.
Tesla’s new hardware/software combination now provides an enhanced autopilot capability that allows cars to enter and exit freeways, maintain speeds based on current traffic conditions and automatically change lanes without any driver input.
But the icing on the cake is SAE level 5 autonomy – this means that other than starting the system and setting the destination, no driver intervention would be required. According to CEO Elon Musk, Tesla hopes to achieve this by the end of 2017.
How can the cars become fully autonomous?
Tesla plans to accomplish this by equipping their cars with a suite of sensors that constantly map the surroundings of the vehicle, feeding data to a powerful on-board ‘supercomputer’ capable of fast real-time computations that help the vehicle determine its next action.
The hardware suite that can accomplish Tesla’s self-driving vision includes:
- Three forward facing cameras (up to 250m of visibility)
- Two forward-looking side cameras with 90o visibility (up to 80m visibility)
- Two rearward-looking side cameras (up to 80m visibility)
- One rear-view camera (up to 100 m visibility)
- A forward-facing radar (up to 160m)
- Ultrasonic sensors providing all-round proximity sensing (up to 8m)
Coupled with the powerful on-board computer (40x the power of the previous generation), the system will be capable of truly ‘seeing’ in every direction around the vehicle, gathering a vast amount of information and processing data in real-time using Tesla’s advanced neural network software.
360o Vision and The Autonomous Car:
This 360o vision is critical to achieving complete autonomy, since it enables the vehicle to respond in almost any scenario. At the outset, Tesla vehicles equipped with this hardware will be operated in ‘shadow’ mode, which will enable the sensors to collect important data with a human driver behind the wheel – this would be used to train the autonomous driving model. Once Tesla is satisfied with the accuracy of the driving model, the autonomous driving capability will be pushed via another over the air software release.
Tesla will then most likely look to make the system suitable for extreme weather conditions such as rain and snow, a technology aspect that is also being looked into by Ford. Though competitors like Google and Volvo are already working on functional versions of such vehicles, Tesla has definitely established itself as the forerunner with the introduction of the 360 vision system in its production vehicles.
The future is here, and it definitely saw you coming!
