Weekly Digest #20: Unmanned Taxi, Nougat from Google, and IoT Tools for Developers

August 25th, 2016

IT News / Internet of Things / Big Data // Zee

Unmanned taxis launch and other IT News

Google presented Android 7.0 Nougat, a new version of its mobile platform, Microsoft has granted access to Windows PowerShell code, and car manufacturers have stepped up their efforts to create unmanned vehicles.

IN A NUTSHELL

  • Microsoft made the Windows PowerShell code open-sourced. It is used for the system administration in the Windows OS. In addition, the company made the code available for Linux platform. You can see and get the code on GitHub.
  • Google introduced Android 7.0 Nougat. The new version of the mobile platform has over 250 new features and updates. Now apps can share the screen and the power-saving mode got enhanced as well.
  • Superbook, a docking station, will transform any Android smartphone into a laptop. It will be available next year for $99.
  • Mobile devices are becoming the main way to connect to the Internet. According to Sandvine, 30% of users who are connected to the Internet at home use their mobile gadgets for browsing the web instead of their laptops and desktops.
  • Over 60% of Americans are afraid of the Internet of Things because they see a potential for break-in, especially the connected cars and cameras.
  • Within the last three months, the number of application downloads made thanks to Google mobile advertising has reached 3 billion. In May, this figure was 2 billion. Mobile advertising on Facebook hasn’t been as successful.

Data Pipeline for working with raw data

Parse.ly Platform has introduced a new tool to work with data called Data Pipeline. With its help, app developers can access the data coming into the API service, that is, the data used in statistics. For example, Data Pipeline will give access to the data on the most popular types of content or what was user’s path in the app before placing an order. In other words, Data Pipeline allows you to analyze a great number of events on the basis of the system statistics that Parse.ly hasn’t disclosed before. Data Pipeline’s data is presented in formats like RedShift or Google BigQuery, ready for loading into the SQL database.

Intel Joule - IoT-set for developers

At the IDF 2016 conference, Intel has introduced a new set of tools for developers called Intel Joule. Besides the tools, the set is accompanied by a device - a small circuit board, which allows to develop and test applications in such fields as robotics, augmented and virtual reality, and the Internet of Things.

Two types of Intel Joule devices have been shown. The first model, Intel Joule 550x, runs on Atom T5500 quad-core processor with a frequency of 1.5 GHz, 3GB of RAM and 8GB of internal flash memory. The second model, Intel Joule 570x, has quad-core Atom T5700 with a frequency of 1.7 GHz, 4GB of RAM, and 16GB of internal flash memory. Both boards support Intel RealSense cameras, Wi-Fi (802.11ac), 4K graphics, and OS based on Ubuntu Linux. The first model, according to preliminary information, will cost $369.

Open Registry for the Internet of Things

Chronicled company has created a public registry for IoT-gadgets. The registry is built using Ethereum blockchain technology and can store individual codes that describe a particular IoT-device. This registry will allow private and corporate consumers to be sure that they have bought and use an original device and not a counterfeit one. This is one more way to protect yourself from possible data leakage due to unauthorized devices.

The registry is designed to store a code with built-in NFC and BLE  chips. The creators say that it already contains data about over 10,000 units. Thanks to this registry, all IoT-device developers will be able to register them and protect from counterfeit, at the same time obtaining confirmation about their identity. Over time, a label can be placed on the device signifying that it is an original, backed up by the secure registry.

In additional to the traditional IoT-devices - gadgets for the smart home or car - the registry will allow also to identify, for example, objects of art, ensuring they are original and not fake.

Compressing picture with record-breaking efficiency

Google employees have taught a neural network to compress images with record-breaking results, exceeding the degree of compression that a JPEG format conversion can achieve. The neural network created at Google works on the basis of TensorFlow, a free machine-learning library. This development reminds of the one shown in the “Silicon Valley” TV series.

Over 6 million photos from Kodak database with dimensions of 1280 by 720 pixels and a set of non-overlapping blocks of 32 by 32 pixels have been used for the neural network training. The samples that were used for the training had the least efficient compression, difficult for the algorithm to work. While learning to compress complex samples, the network has learned to compress simpler fragments along the way. To access the level of compression, several standard valuation methods were used, such as Multi-Scale Structural Similarity (MS-SSIM) and Peak Signal to Noise Ratio - Human Visual System (PSNR-HVS).

Developers are planning to improve the algorithms for the large image compression techniques and use video codecs, such as WebP based on VP8, for the task.

Unmanned vehicles are becoming a reality

The testing of an unmanned minibus with automatic control without a driver began in Helsinki. At the same time, the world’s first unmanned taxi began working in Singapore.

Uber began testing of the unmanned taxi and promised that such cars will be a part of their service will be taking on passengers as early as next year. Uber is doing this project in conjunction with Volvo and $300 million have been spent on the implementation of the project already. Talking about futuristic technologies in the field of unmanned vehicles, Airbus is planning to test an unmanned flying taxi by the end of this year.

These unmanned vehicles cannot travel without the appropriate software and Delphi and Mobileye companies are working on its development. The first version of the platform for such cars will be presented in 2017, with their mass adoption expected in 2019. Mobileye is a major player in the computer vision technology market and Delphi has been developing hardware and software for cars.

But that’s not all. Ford plans to establish a serial production of unmanned vehicles under its own brand. To make this happen, Ford has invested in Velodyne company that has been developing LIDAR systems for receiving and processing information about remote objects. In addition, the American car manufacturer bought Saips, the developer of machine vision systems and machine learning.

Who’s up for a ride?

Author: Zee

Zee is in charge of the Marketing at the APP Solutions/Grossum. Her areas of interest include quantum physics, astronomy, new trends in the web & mobile development (especially in the areas of AI and machine learning) and digital marketing instruments.

Tags Neural Networks Blockchain Analytics Google Internet of Things

See all blog

x

Grossum Startup Guide:
Five Things to Consider Before Web & Mobile Development

Get the Guide