After reading more articles about Intel Euclid, I was surprised by the enthusiasm created around the product. LinuxGizmos, LiLiPuting, The Verge, Engadget, are just some of the online publications that have written about the new product from Intel, a product designed to build robots.
But to make sure there is no false enthusiasm, I did an analysis of the kit to see if is worth it or not to use it in robotics projects. The conclusion is below and it is worth reading to the end.
The first step was to make a list of the advantages and disadvantages of the Euclid kit.
- it has WiFi, 3D sensor, and other sensors in a very small box
- it runs Linux Ubuntu (not a surprise)
- it runs ROS Kinetic
- is Arduino compatible
- Intel has launched tutorials to make the introduction for users
- double processor power and four times more RAM than Raspberry Pi 3
- few ports for accessories
- it does not have GPIO pins to attach other sensors (other than those already integrated)
- inside, the GPS module is useless
- the product is not modular. You use it as it is. If you want to replace a hardware component, you need an expert or a service
I also have some unanswered questions:
I will surely find the answer to the following questions in the period after the official launch, but until then:
- How does RealSense ZR300 work outdoors in bright light?
- Can the internal battery of 3.8V and 2000mAh be replaced?
The price paid for Euclid versus what happens today:
When a robot needs 3D vision, GPS, WiFi, Linux, and a few other sensors, it is used a 3D kit such as INTEL RealSense (~$150), a GPS module (~$30), an Arduino board (~$7), and install Linux Ubuntu and ROS on a Raspberry Pi 3 (~$90). All of these have a cumulative price of about $277. If we add the cost of other sensors and a battery, we will most likely reach $350. An amount quite close to the actual cost of the Euclid kit that can be pre-ordered for about $400.
If we depend heavily on the packaging of the components used in the project and more efficient hardware resources, Euclid is the preferred option. If we want to reuse components in other projects, the modular version is a better choice. Let’s not forget here something very important, namely the Raspberry Pi community.
Intel Euclid is for professionals. Nobody will spend nearly $400 to build a line following robot. Instead, will be spending a lot more on a self-driving robot. With ROS, you can use two or more Euclid modules at the same time. In such a case, you can divide the tasks by running different ROS nodes on each Euclid module individually. Of course, the same thing you can do with Raspberry Pi.
And to don’t have any doubt that Euclid is another Raspberry Pi killer, the first is a niche product that you accept or not as it is, while Raspberry Pi is a development board for everyone.