News

Desktop Level Humanoid Robot can Perform hurdle crossing, grabbing and sorting, more AI gameplay?

Desktop Level Humanoid Robot is a type of robot designed for use in a home, office, or education setting. These robots are typically larger in size and have a broader range of functions than mobile or handheld robots. Some common tasks that desktop-level humanoid robots can perform include assisting with daily tasks, providing companionship, and serving as an educational tool. Examples of desktop-level humanoid robots include the RoboTV, Pepper, and embodied intelligence platforms such as Nao or Bossa Nova. These robots use artificial intelligence and are designed to interact with humans in a natural and intuitive way.

Desktop Level Humanoid Robot

Do you want to have such an AI Humanoid Robot toy? Can understand your instructions and imitate your actions:

Desktop Level Humanoid Robot

Walk, climb, hurdle, grip, and complete many complex movements:

Desktop Level Humanoid Robot

Can he help you deliver something or something?

This small humanoid robot is  a desktop level educational humanoid robot based on the Robot Operating System (ROS).

According to the official introduction, this Humanoid Robot is powered by Raspberry Pi and has self stable inverse kinematics. It can walk, cross obstacles, automatically patrol, kick the ball, recognize objects, and perform facial recognition.
High performance hardware structure supports flexibility in motion
In terms of hardware,This Humanoid Robot is made of aluminum alloy and has up to 24 degrees of freedom (DOF) throughout its body, capable of performing a wide range of movements and ensuring flexibility in movement.

Desktop Level Humanoid Robot

This Humanoid Robot adopts a high-performance intelligent serial bus servo system. Compared with the traditional 7.4V servo system, this servo system greatly improves the accuracy and linearity of the robot by using high-precision potentiometers, providing precise angle control. The plug and play wiring method makes robot assembly simpler and wiring more convenient. The use of metal gears can greatly extend the service life and provide greater power for robots.

This Humanoid Robot looks like a cute one eyed monster, with a head that can rotate in all directions. Don’t underestimate this “eye”. It is a 120 ° high-definition wide-angle camera, which gives the robot a wide field of view. The camera also supports manual focusing. In addition, using OpenCV image processing technology, robots can effectively recognize and locate target objects.

It also features an innovative dual hip joint structure that overlaps the X, Y, and Z axes to ensure orthogonal alignment, allowing the legs to rotate on all X, Y, and Z axes. This integration not only enhances the bionic quality of This Humanoid Robot, but also significantly improves the robot’s motion flexibility.

At the same time, the hip joint is equipped with a 12 bit high-precision magnetic encoder (HX-35HM servo), which can expand the angle range to 360 degrees and maintain 4096 bit absolute position accuracy; In addition, it can achieve a locked rotor torque of 35kg.cm at 12V voltage, providing more power for the robot.

Unlike the traditional position feedback mechanism using potentiometers, the HX-35HM servo adopts non-contact magnetic feedback, which can prolong service life, reduce size, and improve impact resistance.

This Humanoid Robot’s robotic arm can be fully opened up to 72mm, allowing for flexible gripping and handling of small items. The servo provides angle and temperature feedback, effectively preventing damage caused by obstacles.

Self stable gait algorithm+AI visual recognition to complete various complex tasks
In terms of software and algorithms,This Humanoid Robot is supported by Raspberry Pi, which uses a semi closed loop self stable gait algorithm. It can not only dynamically adjust the robot’s gait in real-time, but also adjust its height, speed, and turning radius, allowing it to turn while walking.

Desktop Level Humanoid Robot

In addition, This Humanoid Robot uses OpenCV as an image processing library to identify target items. With superior gait algorithms and AI visual recognition, This Humanoid Robot can complete multiple complex tasks, such as route tracking, target tracking, ball shooting, intelligent picking and sorting, transportation, climbing stairs, etc.
Color recognition and tracking: This Humanoid Robot utilizes OpenCV to process images for color recognition and localization. When This Humanoid Robot detects red, it will nod. When green or blue is detected, the robot will shake its head.
Using OpenCV to load pre trained face detection models, This Humanoid Robot can perform greeting actions when recognizing faces.

Desktop Level Humanoid Robot

Identify AprilTags and make various reactions based on the recognition results.

Desktop Level Humanoid Robot

Using PID algorithm for calculation and camera based recognition, This Humanoid Robot can change its gait to successfully travel along the route and also track the movement of target colors.

Desktop Level Humanoid Robot

When playing football, This Humanoid Robot will walk to the position of the ball and kick it again. It can calculate the distance to the target by determining the target coordinates and adjust its gait autonomously to kick the ball.

Desktop Level Humanoid Robot

In addition to motion gait and target tracking, This Humanoid Robot has recently seen an upgrade in AI interaction. Developed based on the MediaPipe algorithm, it can recognize human features and achieve functions such as face detection, gesture recognition, and human recognition.

Desktop Level Humanoid Robot

Sensory control, ainex can perform appropriate actions after recognizing the human body

Fingertip trajectory control, This Humanoid Robot can adjust its height accordingly based on the distance between the detected fingertips
Fully open source code, capable of secondary programming, supporting multiple international competitions

The entire robot kit includes: robot body, charger, card reader, small props used for AI training, toolbox, wireless controller, and electronic manual, etc. The official tutorials cover up to 18 topics.
If the game is still not enough, users can further develop the robot to expand more functions. It is reported that This Humanoid Robot is built on the Robot Operating System (ROS), which provides code reuse support for robot research and development, enabling flexible deployment and expansion of robot functions. In addition, the official also provides comprehensive source code and ROS simulation models.

They also provide Gazebo simulation support, which is a novel method of controlling This Humanoid Robot and validating algorithms in a simulation environment. During secondary development, the development results can be directly verified without physical experiments, greatly improving efficiency.


So far, This Humanoid Robot as assisted in multiple international robot competitions and attracted many well-known domestic and foreign universities and research institutions.
In addition to This Humanoid Robot, this company has also developed multiple biomimetic educational robots such as multi legged robots and intelligent visual robotic arms. 

You are welcome to  https://www.youtube.com/@tallmanrobotics to watch our video centre for more projects or visit our website to check other series or load down e-catalogues for further technical data. 

Prev:

Next:

Leave a message