XGO Mini Pro robot dog features Kendryte K210 AI processor (Crowdfunding)

[Update: Initially published on July 30th, XGO Mini Pro robot dog is now up on Kickstarter. While it was listed on the Kendryte K510 product page, it is only offered with the earlier Kendryte K210 in the crowdfunding campaign]

Kendryte K510 RISC-V AI processor is the successor of Kendryte K210 with much more AI processing power (2.5/3 TOPS) that makes it suitable for robotics projects demanding much lower latency and quasi-real-time inference.

One of the first products to be compatible with both Kendryte K510/K210 capabilities is Luwu Intelligence Technology’s XGO Mini Pro, a four-legged robot that looks like a dog with a tiny head providing the “smarts” for the system including face detection, image recognition, object tracking, voice recognition, and so on.

XGO Mini Pro Kendryte K510 Robot Dog

We first found the description of the robot on Canaan’s website is rather short with an explanation that when equipped with is equipped with a 9-axis IMU and joint positioning sensors, providing behavioral and joint feedback. The head is exchangeable and when using the earlier Kendryte K210, the robot is capable of omnidirectional movements, six-dimensional navigation control, and stabilization support, as well as multiple movement modes. Kendryte K510 would potentially enable more features with object recognition functions, human posture recognition, dynamic gestures, and guide dog traffic sign recognition.

But the robot has now launched on Kickstarter with Kendryte K210 only. You can watch a video showing the capabilities of the robot both in terms of movement (12 degrees of freedom) and AI processing.

YouTube video player

XGO Mini Pro comes fully assembled and can be controlled with an app. The company also says the AI module can run locally and the internet/cloud is not needed.  The robot can also be programmed by the user either using Python programming or CocoBlocky visual programming interface.

XGO Mini Pro key features and specifications:

  • SoC – Kendryte K210 RISC-V AI processor
  • MCU – STM32 for motor control
  • Storage – 4GB MicroSD card
  • Display – 240×240 full-color LCD
  • Camera – 0.3 MP camera using OV2640 sensor
  • Battery – 7.4V/2,500 mAh Lithium battery
  • Power Supply – 8.4V/1A charger
  • Dimensions – 250 x 150 x 60 mm (Materials: aviation aluminum legs, silicone & nylon)
  • Weight – 700 grams

A lite version of XGO Mini was introduced last year without AI module, and XGO Mini Pro builds on that model, but features a Kendryte K210/K510 AI module, increased joint torque (4.5kg.cm vs 2.3kg.cm), comes with steel gears instead of copper ones, as well as magnetic encoders, and coreless servos.

XGO Mini Pro four-legged robot

The smaller & lighter XGO Mini Lite, without AI module, is offered for $199 on Kickstarter, while XGO Mini Pro robot dog reward starts at $549 for the super earlier bird rewards, and will be $599 afterward. The retail price is expected to be $1098. Shipping is expected to start in November 2021. If the campaign can raise over $500,000, XGO Mini Ultra version will be offered as a stretch goal with a robotic arm.

Share this:
FacebookTwitterHacker NewsSlashdotRedditLinkedInPinterestFlipboardMeWeLineEmailShare

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress

ROCK 5 ITX RK3588 mini-ITX motherboard

6 Replies to “XGO Mini Pro robot dog features Kendryte K210 AI processor (Crowdfunding)”

  1. Reminds me of Nova Spot Micro 3 and that hand sized plastic one CNX-Software covered a while back, I cannot recall the name of.

    1. Canaan lists the robot on their K510 page. But then there are mentions of K210 as well. So it’s unclear.

  2. Part of me REALLY wants it but the chances of a crowdfunding to be successful are basically 0, let’s be honest.

Leave a Reply

Your email address will not be published. Required fields are marked *

Boardcon Rockchip and Allwinner SoM and SBC products
Boardcon Rockchip and Allwinner SoM and SBC products