To USB or Not to USB

 

World’s First ASIL D Ready EV Processor (Pair it with USB!)

World’s First ASIL D Ready EV Processor (Pair it with USB!)

 

We are announcing the industry’s first ASIL D Ready embedded vision processor: https://www.synopsys.com/company/newsroom/mnr/asil-d-ev6x-vision-processor-sep.html

 

Embedded vision processors allow for rapid, local detection of items in a field of vision for an optical, radar, or other sensor. It’s used in car safety systems and autonomous vehicles because it provides faster feedback for warning or collision avoidance.

 

ASIL D is the Automotive Safety Integrity Level, and there are four levels (A, B, C, D). It measures the ability to find single or multiple faults. ASIL D is the highest, most stringent level of safety. It is required for systems that, if they fail, someone could die. The EV processors are available in ASIL A, B, C or D to fit the application’s requirements.

 

To achieve ASIL D, a system needs 99% coverage of the possible faults.

 

To implement ASIL D support, here are a few things we did:

  1. Implement Error correction code for Caches and Closely coupled memories. This detects and corrects internal data corruption.
  2. Implement a Memory Protection Unit to prevent a process (program) from accessing memory that has not been allocated to the process. (Software stays in the memory it is assigned because it can’t access memory outside its sandbox)
  3. Implement lockstep capabilities to provide redundancy for error detection and correction by running the same set of operations at the same time in parallel
  4. Implement logic built-in self-test (LBIST) ready features to allow hardware/software to test its operation and report back all is well (or not so well).

 

This is just the latest in Synopsys’s broad portfolio of ASIL D Ready IP that we’re delivering to multiple customers. We are excited to be the first to ASIL D Ready EV processors.

 

 

A couple of other great features of our EV offering:

  • Save millions of dollars initially and in reuse by adopting a single programming platform. This is especially useful since cars will have in excess of 20 cameras that can use an EV processor, with varying ASIL requirements.
  • Easy to configure a power and area optimized embedded vision offering targeting different applications including radar/lidar, pedestrian/object detection, distracted driving detection (watching you inside the car), or a 360 view optical camera.

 

 

 

 

What’s the relationship to USB? The 20+ optical cameras, infrared sensors, radar and other devices will likely have a USB connection to communicate with the visual unit to move data streams for processing during development. Most chips in the car will have a USB port as part of the chip at least at the development stage for firmware upload, debut, and diagnostics. It may or may not be accessible in an actual car when finally deployed.

 

Eric’s personal note on Autonomous Vehicles – It seems that every single accident involving a Tesla gets publicized widely.

When I was training to be a Civil Engineer about 192 years ago, we talked about having autonomous vehicles. Back then, we talked about cars getting on a freeway and locking into a track or something. I would have thought we would be there by now.

 

The sensationalism that the internet likes to publicize Tesla accidents (and possible failures) often overlooks or forgets human error is a big cause of accidents. And of course they never mention the lives saved due to safety features.

 

In the US, there are about 40,000 auto fatalities every year. The goal of autonomous vehicles in not luxury, status or coolness, but safety. If autonomous vehicles can reduce the total number of accidents, reduce the severity of accidents, reduce the number of fatalities, then the billions being poured in will be worth it. All the other things, like reduced greenhouse gas emissions from having shared electric cars will be worth it. Maybe our car insurance rates will drop also. Which would be cool.

 

We already have driver assistance in many cars which I would guess is already helping reduce the number of accidents. We have Uber/Lyft* drivers preventing  drunk drivers from driving drunk (probably).  Read the study linked below.

https://www.theverge.com/2017/10/4/16418782/uber-drunk-driving-crashes-study-cities

 

It’s an exciting time to be part of the automotive industry. A lot will be invested to make everyone on the road safer.

 

Oh yeah, and I just need to say this. Flying cars are a stupid idea, no matter how cool they looked in the Jetsons. There is no way I want a 250+ pound aircraft carrying a 100+ pound person above my house or my head.

 

*Comment from Agent K “There’s an argument to be made that Uber/Lyft cause more traffic as they circle around, waiting for a fare.”

https://imgs.xkcd.com/comics/fatal_crash_rate.png

Share and Enjoy:
  • del.icio.us
  • Digg
  • Facebook
  • Google Bookmarks
  • Print
  • Twitter
  • StumbleUpon
  • LinkedIn
  • RSS