Autopilot (FSD) News Software Updates Videos

Tesla is migrating from programming logic to Neural Net for FSD decision making


Tesla firmware hacker and researcher who goes by the name of ‘green’ in cyberspace has just revealed that Tesla is migrating towards Neural Nets (NNs) for Full Self-Driving (FSD) decision making. Currently, Tesla vehicles are taking decisions such as ‘right of way’ with programming logic built in C++.

Green usually reveals his findings via his Twitter account and in the recent past has been under Tesla’s wrath resulting in a lawsuit. He posted the latest finding on Twitter in his tweet:

The Tesla hacker thinks that the Silicon Valley-based automaker is implementing this migration bit-by-bit, probably starting from the ‘right of way guessing’ function to other more complex decisions that the car needs to take.

Currently, Tesla is improving the Autopilot (FSD Beta) decision making with each firmware update it is pushing to the select few cars in the United States. With the NNs in the loop, the cars will be able to instantly get feedback from the Tesla Mothership which has tons and tons of machine learning data to help make a better decision.

– Sponsored –

How Computer / Autopilot Vision Works

Tesla Autopilot engineer explains how computers recognize objects via Traditional Programming vs. Machine Learning + AI. Tap/Click to load the HD version of the infographic (full video below). Source: / Tesla.

Tesla Autopilot engineer Kate Park explains how computer vision works in the following video. I have included one infographic above from the video that explains how decision making or object detection is limited to the use of traditional programming. With the use of Machine Learning + AI + Neural Nets, the whole process provides limitless possibilities.

For example, traditional programming cannot recognize “X” if it is not exactly drawn with the defined parameters (see image above). But the combination of machine learning, AI, and Neural Networks enable the computer to learn multiple patterns and possibilities an “X” can appear in front.

How Tesla Neural Net vision works, explained by Tesla AI director, Andrej Karpathy on Tesla Autonomy Day 2019. Tap/Click to load the HD version. Source: Tesla.

To make the computer vision work perfectly, it requires from thousands to millions of images to correctly define an object. In the case of Tesla, every car of the fleet sends back video data from all 8 cameras to train the Tesla Neural Net. Tesla has billions of miles of real-world driving data gathered from its worldwide fleet of more than a million cars. This gives Tesla an edge over any company working on self-driving cars and car vision neural nets.

Earlier this year, Tesla CEO Elon Musk said that the company’s powerful next-gen neural net called Dojo is being built and its version 1.0 should be up and running next year — interesting indeed.

Autopilot driving decisions migrating towards NNs is a giant leap forward in Tesla FSD development and the new year should be more interesting with the results.

Stay tuned for more, follow us on: Google News | Flipboard | RSS (Feedly).


By Iqtidar Ali

Iqtidar has been writing about Tesla, Elon Musk, and EVs for more than 3 years on, many of his articles have been republished on CleanTechnica and InsideEVs, maintains a healthy relationship with the Tesla community across the Social Media sphere. You can reach him on Twitter @IqtidarAlii