📷 Image Sensors World: ResearchInChina: Automotive Thermal Vision is of Little Value Negatives: “It’s… https://t.co/G652xgnvBV

from Twitter https://twitter.com/news_driving

📷 TE raises profile on Triad operations’ silent safety partner role | Business | https://t.co/VUVRMqyKlx https://t.co/TAFXI3V3Mw

from Twitter https://twitter.com/news_driving

China’s 5G remote-control cars – China Mobile showcased teleoperations over 5G controlling a car more than… https://t.co/r5ffvAMZYr

from Twitter https://twitter.com/news_driving

Tactile Mobility adds missing sense of touch to autonomous vehicles

Tactile Mobility adds missing sense of touch to autonomous vehicles:

Using big data and ML  generate real-time insights regarding physical factors that impact any ride and augment existing sensory data.

“There is always a compromise between safety and user experience in autonomous vehicles. They rely on cameras and lidar sensors, know the vehicle speed, the speed of the vehicle in front, its distance from your vehicle, and calculate the safe distance of the vehicle given relative velocities, but they don’t know the road grip level,” said Nisenbaum.

“They need to assume that there is a low grip level in order to be safe. But that’s sub-par from the user experience, because a user is used to driving in a certain way and at a certain speed.”

Tactile Mobility is currently working to implement its technology with six leading car manufacturers in Europe and North America, including American automobile giant Ford. While the company’s software can be embedded in existing vehicles, it hopes to see its technology included in mass-market production as soon as the 2020 and 2021 model years.


Power Allocation Between Redundant Systems on Autonomous Vehicles

https%3A%2F%2Fmedium.com%2F%40LyftLevel5%2Fhttps-medium-com-lyftlevel5-power-allocation-between-redundant-systems-on-autonomous-vehicles-a68f5d80f061&title=Power Allocation Between Redundant Systems on Autonomous Vehicles&selection

  • Perception, planning, or control insufficiencies are likely to be responsible for more faults than plain hardware faults.
  • Additional power to each function monotonically reduces, with diminishing returns, the frequency of its faults. This is the simplest behavior we could choose.
  • Increasing sophistication of each strategy can be thought of moving its respective line down or to the left.


Vodafone Group and Continental powers AI with Netrounds |…

Vodafone Group and Continental powers AI with Netrounds | Telematics News

Using Multi-access Edge Computing (MEC) Vodafone and Continental stream pre-processed sensor data over the mobile network for real-time analytics and instant decision-making processes. AI on edge cloud nodes applies determines if an action is required. 


Volkswagen’s self-driving cars are about to hit a new…

Volkswagen’s self-driving cars are about to hit a new city

The pilot is for five e-Golf vehicles in Hambourg

These are L4 vehicles with safety drivers. 

V2I is included with 37 traffic lights and a bridge which will send information out.

Sensors include 11 LiDARs, 7 radar sensors, and 14 cameras. The cars transmit 5 gigabytes of data per minute while driving, or 666Mbps.

VW already tests in California and other parts of Germany — and has plans to enter China. 


Wayve claims ‘world first’ in driving a car autonomously with…

Wayve claims ‘world first’ in driving a car autonomously with only its AI and a SatNav | TechCrunch

The traditional approach used by all of Wayve’s competitors relies on HD-maps, expensive sensor suites and hand-coded rules that tell the car how to drive. Wayve have built a system that learns end-to-end with machine learning. It is the first in the world to drive on urban roads it has never been on before. 

Their model learns both lateral and longitudinal control (steering and acceleration) of the vehicle with end-to-end deep learning. Uncertainty is propogated throughout the model, which allows it to learn features from the input data which are most relevant for control, making computation very efficient.

This massively reduces the sensor & compute cost (and power requirements) to less than 10% of traditional approaches.

Assuming other independent observers can confirm these claims, it looks like a UK startup just leap-frogged the entire autonomous car space.