myArm M & C:Advanced Solutions for Diverse Robotic Applications


In recent years, many projects similar to Stanford University’s Alopha robot project have emerged, primarily focusing on learning by mimicking human motion trajectories to achieve human-like artificial intelligence. The Alopha robot, through advanced algorithms and sensor technology, can precisely replicate human actions, collecting data and learning from it to enhance its performance in various tasks. This imitation learning approach not only enables robots to excel in industrial automation but also shows significant potential in educational and research fields. Against this backdrop, Elephant Robotics has introduced the myArm M & C series robotic arms, further advancing humanoid robot technology.

Innovative Solutions for the Future

With the rapid advancement of robotics technology, Elephant Robotics has introduced the groundbreaking myArm 650 M & C series robotic arms, offering unprecedented flexibility and precision for educational, research, and industrial applications. These high-performance robotic arms are not only powerful but also flexible and user-friendly, making them suitable for a variety of applications and providing ideal solutions. Let’s delve into the myArm 650 M & C series products!


myArm C650

The myArm C650 is a versatile 6-degree-of-freedom robotic motion information collection device, equipped with a fingertip controller and two intelligent buttons. The “C” stands for Controller. It can output end coordinates or joint angles at 50 Hz. Designed for education, research, and industrial data collection, its flexibility and highly modular design make it suitable for various complex operations and tasks. It can be extended to a 7-degree-of-freedom data collector.

myArm M750

The myArm M750 is a general-purpose intelligent six-degree-of-freedom robotic arm with a reach of 750 mm and a rated load capacity of 500 g, with a maximum capacity of up to 1 kg. It comes with a 1-degree-of-freedom parallel end effector (gripper). The “M” stands for Master controller. It is suitable for applications requiring complex motion control, precise positioning, and high programmability.

Product Features

● High-Speed Data Output: Both the myArm C650 and M750 can output end coordinates or joint angles at a frequency of 50 Hz, providing real-time motion data to ensure smooth operation.

● Flexibility and Modular Design: The robotic arms feature a highly modular design, allowing users to customize and expand according to their needs, making them suitable for a wide range of applications.

● Robust Load Capacity: The myArm M750 has a reach of 750 mm and a load capacity of 500 g, with a maximum capacity of up to 1 kg, making it suitable for complex industrial tasks.

● Programming and Control: Supporting Python and ROS, the arms offer powerful programming capabilities and flexible control options, making them ideal for education and industrial automation fields.

● Strong Compatibility: The arms support various sensors and end effectors, including cameras and IMUs, further enhancing data collection and processing capabilities.

AI Learning and Training Project

Recently, many projects similar to Stanford University’s Alopha robot project have emerged. These projects mainly involve learning by mimicking human motion trajectories. By recording trajectory data and conducting extensive training, the Alopha robot can autonomously perform certain household chores, such as cooking, washing dishes, and folding clothes.

The key elements of the entire project are the full-body remote control system, human demonstration learning, and collaborative training of datasets. To this end, we have also launched a corresponding humanoid composite suite, equipped with two sets of myArm M & C robotic arms and the Elephant Robotics mobile platform for mobility.

Now, let me demonstrate how to operate using a set of myArm M & C.

The myArm M650 is essentially a 6+1-axis robotic arm: 6 degrees of freedom for the arm itself and 1 for the additional end gripper controller. To ensure better data synchronization, we have also designed the myArm C to be of the same type but different sizes.

You can see that the synchronous remote control operation is quite smooth. This is mainly due to the myArm’s ability to output terminal coordinates and joint angles at a rate of 50 Hz, providing users with real-time motion data. Let’s try grasping some parts to experience its effectiveness.

To achieve functionality similar to the Alopha robot, we have completed two parts: the remote control system and human demonstration learning. The most crucial part remaining is the collaborative training of datasets.

Collaborative Training of Datasets:

Collaborative training (co-training) is a machine learning technique particularly suitable for multi-task learning and ensemble learning. The core idea is to use multiple different datasets or data perspectives to jointly train a model, thereby improving the model’s performance on specific tasks.

Here is a detailed introduction about the dataset:

The entire Aloha project is open source, so we can find how their team conducted the training online. The specific steps are as follows:

  1. Initial Model Training: First, train the initial model using an existing static dataset. This model can perform dual-hand static tasks.

  2. Mobile Operation Data Collection: Then, collect demonstration data of the robot performing tasks in a mobile environment using the full-body remote control system.

  3. Joint Training: Combine static and dynamic datasets for training. By integrating data from both scenarios, the model can learn a broader range of task characteristics, improving its generalization ability and success rate on new tasks. The data involved includes tasks such as grasping and placing objects, tidying up the table, moving speed, opening and closing cabinet doors, camera images, LiDAR data, and IMU data.

To meet the needs of most projects, the myArm M & C series offers significant performance advantages, including high-frequency data transmission at a rate of 50 frames per second. It also has specialized software to collect relevant data, such as the height of objects, the height of the table, and the speed of movement.

Additionally, the myArm is equipped with multiple sensors, such as cameras and IMUs, with the ultimate goal of acquiring precise datasets for training.

Application Scenarios

The project mentioned above is just one example. Our robotic arm can do much more and can be used in conjunction with quadruped robots for tasks like intelligent garbage collection.

Standalone Robotic Arm Applications:

Medical and Rehabilitation

● Surgery Simulation: In medical training, the myArm M & C can simulate surgical procedures, helping surgeons practice and train before actual operations.

● Rehabilitation Assistance: It can assist in the rehabilitation of patients by facilitating hand and arm recovery exercises, with programmable different rehabilitation movement trajectories.

Education and Training

● Robotics Programming Courses: In schools and training institutions, the myArm M & C can be used as a teaching tool, helping students learn robotic programming and control. Through programming tasks, students can master how to control the robot’s movements and perform complex operations.

● Laboratory Experiments: Used in university and research institution laboratories for the study of robotic arm kinematics and dynamics. Students can conduct precise motion control experiments and analyze the robot’s performance.

Combined Applications:

Integration with Mobile Robots

● Automated Warehouse System: Combine the myArm M & C with mobile robots (such as Elephant Robotics mobile platforms) to create an automated warehouse system. The mobile robots handle the transportation of items, while the myArm handles the picking and placing, achieving efficient item management and sorting.

● Logistics and Transportation: In logistics centers, the combination of the myArm and mobile platforms can automate the sorting and transportation of packages, improving logistics efficiency.

● Sentinel Patrol: Paired with quadruped robots, it can perform patrol inspections in parks and clean up trash when detected.


The myArm M & C series robotic arms, with their high flexibility, precision, and modular design, demonstrate broad application potential in education, research, industry, and home environments. When used independently, these robotic arms excel in tasks such as precision assembly, quality inspection, educational training, and medical rehabilitation, showcasing their exceptional performance in complex operations. When combined with other robots, such as mobile robots or other robotic arms, the myArm M & C can build efficient automated systems, enabling more complex task allocation and execution. Additionally, these robotic arms can be applied in artistic creation and smart home setups, further expanding their use cases and creative possibilities.

1 Like