Progress for cloud-enabled robots

robot-dog

Robots have become smart, but when you use cloud resources to run them, they can become even smarter – and faster. While cloud-enabled robotics has been around in some form since 2011, some companies have recently taken the field from theory to practice.

PerceptIn, for example, is prepping cutting-edge surveillance robots that incorporate proprietary object-recognition machine-learning algorithms, which are in turn powered by a cloud architecture.

PreceptIn’s Dragonfly robots, still in pre-order, will patrol users’ properties intelligently and send an alert when abnormalities are detected. Besides on-demand video streaming and alerting functions, these robots will also be able to recognize objects. You can then search, identify and initiate playback for specific objects like, say, the guy in the homemade ninja costume trying to sneak in through the window.

PerceptIn’s plan is to use this technology for both companies and consumers. On the consumer side, for example, if you have a cat or dog, the robot will be able to recognize Fuzzie when it sees it. Users can search the stored videos to see what their pet is up to when they are away. If they’re like my dog and cat, this means they’ll be playing with the robot.

Now, we usually think of robots as having their “smarts” within them. Perceptin and other companies, however, are finding that they can do more with robots by keeping the bulk of their data and processing in the cloud. Another example of this methodology can be found with the Amazon Alexa Echo and the Google Home.

In PerceptIn’s case they’re is designing a cloud architecture using Alluxio (formerly Tachyon), an open-source, memory speed, virtual distributed storage program to handle its robot’s data and make sense of it. Its unified namespace provides applications with file-system APIs. These are used to access data in storage systems such as storage-area networks, distributed file systems or object stores. This makes the technology very useful for big-data projects.

Specifically, PerceptIn is using it to address:

  • Customers’ real-time demands. The cloud solution must deliver high throughput and low latency for writing and retrieving video feeds.
  • Large-scale data distributed across disparate storage systems. On-demand video streaming generates enormous amounts of data, and not all data is created equal. Depending on the age of data and whether it is being used by the company’s custom applications (such as its on-demand video streaming application and its object detection application), data is stored across disparate storage systems to optimize for resource allocation efficiency.

All this means the company needs a storage engine that not only can handle enormous amount of incoming data, which will end up in different storage systems, but also provide high throughput and low latency for writing and retrieving video feeds.

Therefore, PerceptIn uses Alluxio as a software-defined network that can deliver high performance and unification across disparate storage systems on premises and in the cloud. Specifically, it runs its business analytics, object recognition, query engines and key value store using data stored in Amazon S3, Ceph, and Hadoop HDFS storage systems. The most valuable data is kept in memory for memory speed access.

Perceptin claims its data-write throughput can reach more than 650 MB/s, which is faster than a local hard-drive, native file system with 120 MB/s.

Even more impressive, when they compared the video retrieval latency, using Alluxio, a video can be retrieved within 500 ms. However, when the video is stored in remote machines, the latency can be as high as 20 seconds. Therefore, using Alluxio to buffer “hot” video data could reduce retrieval latencies dramatically.

That’s all well and good, but at this point the storage throughput isn’t the bottleneck, it’s the network. As C2RO, another cloud robotics company pointed out in a recent IEEE robotics paper, latency and network response time are real problems for cloud-enabled robots. That said, both companies are making progress in making practical cloud robotic services and products.

I am waiting to see what happens next, as are my dog and cat.

RELATED LINKS

The potential benefit of robots in the workplace

The power of code: Building a robot

How serverless computing serves the enterprise

Trackbacks

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: