Automotive company Tesla has filed a series of patents which proves that the company is working on a new Artificial Intelligence chip that will work as a neural net accelerator for its Autopilot Hardware 3.0 computer upgrade. The company’s upcoming upgrade will substitute Tesla’s existing Autopilot 2.0 platform and allow full autonomous driving capabilities in the car.
According to a patent filed with USPTO, old processors can immediately perform single mathematical operations, but it has a restraint like processors can only run on a small amount of data at a time. Basically, they weren’t developed for AI and Machine Learning, while GPUs can perform multiple tasks with larger datasets. As patent noted, processing for AI and Machine Learning typically require performing mathematical operations on large sets of data and usually involves solving multiple convolution layers and pooling layers. So to conquer this, Tesla had to build a new neural AI which can handle Machine Learning and AI specific processing operations and filed a handful of patents for an accelerated mathematical engine, a Computational Array Microprocessor system with a variable latency memory access, a Computational array microprocessor system using non-consecutive data formatting, and a Vector Computational Unit.
CEO of Tesla, Elon Musk has even urged his workers to test the Autopilot Hardware 3.0. Employees who own a Tesla and serve as a beta tester will have the Full Self-Driving package installed into their vehicle without cost, according to the reports. Furthermore, Musk said that Autopilot Hardware 3.0 will be made available in the next few months and the company expects that the improved capabilities of the next upgrade will develop autonomous cars abilities and make it less apt to accidents.