The DOFBOT AI vision manipulator uses jetsonnanoB01 as the main control, OpenSourceCV as the image processing library, the mainstream JupyterLab as the development tool, and Python3 as the primary programming language. The integrated design of the camera and the body of the robotic arm, the visual recognition of “hand-eye integration” makes the robotic arm more agile; it can not only realize color recognition, tracking, and grasping but also recognize and interact with human body features, and even carry out garbage classification model training garbage Sorting; through the ROS robot control system, the complex motion control of the 6-DOF serial bus steering gear is simplified. The official provides a guided quick-start tutorial to help you quickly control the DofbotAl vision robotic arm.
Powerful hardware configuration:
Expansion board function distribution:
| 1 | Micro USB interface | 2 | 12C interface | 3 | Cooling fan interface |
| 4 | Button: K1+K2 key+RESET key | 5 | Status Indicator | 6 | RGB light |
| 7 | Ultrasonic interface | 8 | Serial port function selection | 9 | PWM servo interface |
| 10 | Buzzer | 11 | Bus servo interface | 12 | PS2 handle receiver holder |
| 13 | T-type power supply interface | 14 | switch | 15 | Serial interface |
Various remote control methods:
Product specifications:
Product Size:
Product parameters:
| Microprocessor: | Quad-core ARM A57+128-core NVIDIA Maxwell |
| Al computing power: | 472GFLOPs |
| Operating system: | Ubuntu 18.04 LTS +ROS Melodic |
| Programming language: | Python |
| Steering gear solution: | 15kg*5+6kg*1 serial bus intelligent steering gear |
| Enter: | Wide-angle camera, emergency stop button, network distribution button, reset button, 2*1IC interface, 5V fan interface, voice recognition module |
| Output: | RGB indicator light, OLED display, buzzer, 6-way bus steering gear interface, 6-way PWM steering gear interface, voice broadcast module |
| Power plan: | 12V5A power adapter |
| Power interface: | T socket |
| Remote control method: | Mobile APP, PC host computer, PS2 controller (PC side) |
| Way of communication: | WiFi network |
| Safety protection: | Overcurrent protection, reverse connection protection |
| Robotic arm material: | Anodized aluminum |
| Dimensions after assembly: | 303*135*473mm |
| Assembled weight: | 1256g |
| Robotic arm degrees of freedom (DOF): | 5 degrees of freedom + gripper |
| Payload: | 200g (stretch can be clamped weight) |
| Load: | 500g (clamp handling weight) |
| Wingspan: | 350mm |
| Effective crawling range: | Area with a radius ≤30cm and the central axis as a semicircle |
| Repeatability: | ±0.5mm |
| Voltage: | 100~240V50/60Hz |
| Camera | The 300,000-pixel 110-degree wide-angle camera |
| Resolution: | 480p(640*480) |
| Frame rate: | 30fps |
| Focus method: | Focus can be adjusted manually |
Product features:
1) Based on the powerful AI computing capability of JETSON NANO B01;
Jetson NanoB01 is a small but powerful computer that allows you to run multiple neural networks, object detection, segmentation, and speech processing applications in parallel. Jetson Nano is equipped with a quad-core cortex-A57 processor, 128-core Maxwell GPU and 4GB LPDDR memory, brings enough Al computing power, provides 472GFLOP computing power, and supports a series of popular AI frameworks and algorithms, such as TensorFlow, Pytorch, caffe/caffe2, Keras, MXNET, etc.
2) Artificial intelligence eye tracking;
Control the robotic arm from a first-person perspective, making visual recognition more flexible.
a. Artificial intelligence, hand-eye integration
b.FPV HD video control
3) Cross-platform interconnection control
a. Multifunctional mobile APP (IOS/Android)
Built-in a variety of AI gameplay, you can watch the recognition screen of the robotic arm in real-time and support gesture interaction, movement tracking, garbage sorting, and other gameplay.
b. PC host computer control
In addition to the FPV camera video, the host computer adds a 3D simulation model of the robotic arm. The 3D model rotates synchronously with the entity, combining the robotic arm control theory with practice.
c. USB game controller remote control
In addition to mobile apps and computer control, it comes standard with a USB game controller, allowing you to experience the fun of combining multiple degrees of freedom and keys.
4) Color interactive recognition
a. Color recognition tracking
In addition to standard colors, custom colors can also be extracted for robotic arm color recognition and tracking, enabling eye-following manual operation.
b. You let go, I catch
Put the wooden block in the map recognition area, and the robot arm will automatically distinguish the current color and clamp the wooden block to the corresponding color area of the map (standard configuration).
c. Color capture
Put the wooden block in front of the camera, and after the robot arm recognizes the color of the wooden block, it will clamp the wooden block in the corresponding color area to the middle area.
d. Color interaction
Choose the favorite color of DOFBOT, which will follow the target color back and forth: you advance and retreat, you retreat, and I advance.
5) Model training garbage sorting
Basic functions: The kit has multiple wooden blocks with garbage patterns as standard. By loading the garbage classification model, it can identify the types of garbage on the wooden blocks, return the garbage name and classification data to the APP, and the robotic arm will sort the “garbage” to the corresponding classified area on the map.
Developer function: You can also calibrate the map identification area, place multiple “garbage” wooden blocks in the identification area for real-time identification, and continuously sort various “garbage” to the corresponding classification area on the map.
6) Human body feature recognition interaction
DOFBOT can recognize human body features such as gestures and faces through camera images and complete interactive actions such as grabbing, recognition, and tracking.
a. Gesture interaction
DOFBOT can recognize a variety of gestures and perform corresponding actions for interaction.
b. Gesture capture
Recognize the number gestures, stack the corresponding number of layers, and finally push down the stacked wooden blocks with the “fist” gesture.
c. Face recognition and tracking
Detect whether there is a face in the current picture, mark the face after it is recognized, and follow the movement.
7) Speech recognition + voice broadcast
Combining voice recognition control and broadcast interaction with the AI visual image processing of the DOFBOT robotic arm, the AI visual + voice interactive gameplay is realized between humans and the robotic arm.
Reviews
There are no reviews yet.