TogetherROS™·Bot is a software-level system, and to use it, you need to find a hardware-level computing platform to install it.
Previously, we mentioned that TogetherROS™·Bot will be based on Horizon’s AI chips, fully utilizing hardware performance. How do we use this chip?
Sunrise X3 Board
No problem, Horizon has prepared for us the robot development board equipped with the Sunrise X3 chip – the Sunrise X3 Board.
At first glance, does anyone think of the Raspberry Pi? That’s right, to adapt to developers’ usage habits and to be compatible with many existing modules on the market, the overall appearance and interface of the Sunrise X3 Board are similar to the Raspberry Pi, but the development targets and the computing power of the board are completely different from the Raspberry Pi.
All common sensors are supported by the Sunrise X3 Board, and more robot kits will appear around us in the future.
You can see the hardware resources of the Sunrise X3 Board in this table. The CPU is a quad-core A53 with a frequency of 1.2Hz, and it can also be overclocked within the system. The BPU is a hardware-level AI engine, which is the biggest difference of the Sunrise X3 Board compared to other development boards on the market, and it is the main source of core performance. Similar to the NPU often heard in the industry, it uses Horizon Robot’s unique Bernoulli 2.0 architecture, which can provide 5 Tops of equivalent computing power support.
-
In terms of memory, you can choose between 2GB or 4GB versions, and the storage uses an SD card, which we can choose as long as it is larger than 8GB.
-
In terms of multimedia, the Sunrise X3 Board supports hardware-level H265 and H264 encoding and decoding, and common images and videos can be played smoothly.
-
In terms of interfaces, it can meet our common external device and development needs. Here is a CSI camera interface, the board has 3 USB ports, including 1 USB 3.0 and 2 USB 2.0. There is an onboard debugging serial port through which we can obtain a lot of debugging information at the system level and communicate with the board.
-
The board also has wired and wireless networks. We can directly connect the wired network port to the computer with a network cable to remotely log in and communicate with the board or connect it to the router to allow the board to go online. The wireless network and Bluetooth are a two-in-one module, and we can also directly use the wiki to connect the board to a certain network.
-
The display interface is mainly the HDMI port. If we want to see the system interface or the recognition effect of many visual examples later, we can display it by connecting an external monitor via HDMI.
-
Finally, there is a crucial expansion interface with 40 pins, consistent with the interface definition of the Raspberry Pi, including GPIO, serial ports, I2C, SPI, etc., all of which are available.
-
When using this board, it is recommended to use a 5V3A power supply, connected via a Type-C cable for power supply. The official system currently provided is Ubuntu 20.04, which does not differ much from the native system.
Overall, the Sunrise X3 Board is an embedded AI development board aimed at ecological developers, compatible with Raspberry Pi interfaces, with 5 Tops of edge-side inference computing power and 4-core ARM A53 processing capability.
For individual developers, the Sunrise X3 Board allows every developer to have their own AI development kit without being limited by the high prices of development boards. For higher education institutions, using the Sunrise X3 Board development kit to explore AI and robotics applications is the beginning of learning, helping students quickly get started with artificial intelligence and robotics development. For the industry, the Sunrise X3 Board and Horizon’s AI toolchain can help customers quickly bring products to market and rapidly implement solutions in the field of intelligent robotics.
Sunrise X3M Chip
Of course, the core of the Sunrise X3 Board is this AI chip. Here is specific information about the chip for your reference.
It is worth mentioning that with the iteration of chip technology, in 2023, Horizon will launch the X5 chip, with significant changes including an upgrade to an 8-core A55 CPU with a frequency increase to 1.8GHz, the BPU’s computing power will also reach 8 Tops, and it will add a GPU and more expandable interfaces.
During the use of the Sunrise X3 Board, you can also look forward to the Sunrise X5 Board.
Sunrise X3 Board Interfaces
Alright, let’s return to the development interfaces of the Sunrise X3 Board. Let’s organize this, these are the most commonly used interfaces during the development process of intelligent robots using the development board.
Currently, most modules you use on Raspberry Pi or other development boards can basically be reused on the Sunrise X3 Board, and there are more intelligent ways waiting for you.
Regarding the pin arrangement of the 40Pin expansion interface, you can see here. The pin numbering starts from the pin closest to the HDMI board’s inner side, which is pin 1, and counts in a serpentine manner up to pin 39 and 40.
The above image shows the detailed definition of the 40PIN expansion interface, with 3.3V and 5V power signals, I2C, I2S, SPI, UART, and other communication interfaces, as well as many dynamically configurable GPIO signals, which can basically meet various robot development needs.
Alright, we have understood the Sunrise X3 Board clearly. How to use it will be discussed in future operations.
“Core Principles and Applications of TogetherROS™·Bot”
This series of courses will continue to be updated with the iteration of the system, from system installation to AI applications, building systematic learning resources. It provides content support for intelligent robot industry applications, maker development, and teaching practice.
(Scan the code to join the learning for free)
