Unboxing Yahboom Robotic Arm with Jetson Orin Nano
A few weeks ago, we decided to start our journey for learning robotics as a Sunday hobby project and we decided the best way to learn is to build a robot ourselves. We decided to order this developer kit from Amazon. It took us approximately 3-4 weeks as the shipment came from China. The reason we chose this was that it has most of the sensors we wanted to experiment with, and it offers an integration solution. As our first robot, we want to learn from it but not experiment with too much integration.
It took us a week to decide on Yahboom and our research process included looking at robots and DIY kits, we looked at robots that had a chip that could support LLM-based intelligence on the device, So that we could have low latency. We discovered it on one of the few robots that have support for Jetson Orion. We used Chat GPT and YouTube to discover kits that matched our requirements. There were limited options like the Yahboom kit we also looked at Adeept RaspTank Smart Robot Car, but they didn't have compatibility with Jetson chipsets.
Now, let's move towards unboxing this kit. If you want to quickly watch a video you can check this ~5 minutes video as well or read the detailed text explanation below.
It came with a detailed manual that includes all the information on how to build the robot.
Jetson Orin Nano 8GB Developer Kit :
The first thing we found in the package was the Jetson Orin nano 8 GB developer kit which is going to the brain of our robot. Why? The first advantage is it's a powerful computer that helps us run AI models like object detection, image segmentation, or speech recognition directly on edge. It consumes low power (~7-15W) which matters when you are running off the battery ( it also came in the kit image in the below section). It also helps process large amounts of sensor data (camera, lidar, etc) in real-time. It has 1024 Cuda cores to help process in parallel - faster than an Arduino or Raspberry Pi. You also pre-installed support for TensorRT, CUDA, and Robot operating systems so you can focus more time building. Also, it's small in size and easy to fit on your robot.
Antenna:
Since our GPU doesn't have Wifi or Bluetooth by default, the Yahboom kit comes with a pre-attached version of it in the expansion board. The antenna connects to this module. It will help the robot connect with our local network, stream data, pair with mobile devices, useful for short-range control.
Solid state drive:
We got a 256GB SSD which is the hard disk of our robot. It will store everything, including AI models, logs, OS, training data, and more. 256 GB gives us enough space to experiment and build stuff without worrying about the space.
Hand
It came with 6 DOF meaning 6 different joints for flexible movements controlled by a 6X PWM servo motors made with aluminum alloy fully compatible with ROS + MoveIt. The package comes with motion planning, inverse kinematics, and simulated movement tests. Honestly, I was expecting it to be with more joints but let's see.
Gaming Controller
It's like the robot remote controller which will help us test and drive the robotic arm.
Blocks for visual testing
These are visual markers that will help us test vision, object detection, pick and place skills, and more for our robot.
LED Screen
It comes with a 7-inch LED screen with cables that will help us run GUI without the need to connect to an external monitor. It will run a camera feed, MoveIt interfaces and comes with pre-installed UI tools. More to learn when we run it.
Lidar Camera
This is the eye of our robot, a USB-powered 2D laser scanner that will map 360° around our robot and help with mapping rooms, top-down view of the environment comes with multiple sensors attached.
Robot
This is ROSMASTER X3 PLUS version which is approximately 15 cm (L) × 12 cm (W) × 17 cm (H) built with aluminum alloy. It has a Suspension chassis with 65mm Mecanum wheels and four 520 high-power Hall encoder motors. Will share more about it as we learn.
HDMI Cables, USB, screwdriver, screws, and Batteries
It came with x HDMI and x USB cables that will help us connect for setup development and debugging. The battery and power adaptor will help us power the robot motors.
Summary:
So far, we are excited to build with the kit. You can learn more about it here.
Next Steps:
Over the next couple of Sundays, we will first connect all these pieces and run some standard tests. Then, we will look at the interface of various sensors and how we can use them, and how to run AI locally.
Stay tuned!