Robot Hand Project Begins: Cycloid-Based Micro Actuator

Robot Hand Project Begins: Cycloid-Based Micro Actuator [BCSA Micro Series]

It is often said that the human hand contains approximately 17,000 tactile receptors.

Many experts argue that reproducing all of these sensory capabilities in a robot is, in practical terms, impossible. Yet despite this limitation, the Robot Hand continues to attract significant attention. Why is that?

In industrial environments, the objects robots interact with are rarely ideal. Shapes vary widely, surface materials differ, and packaging conditions are often inconsistent. Even when two boxes appear identical, a slight change in contact position can cause slipping or deformation. Small components are even more sensitive; a minor misalignment can interrupt an entire downstream process.

In these situations, the hand is no longer just a tool for gripping objects. It must determine how much force to apply and detect contact conditions in real time. This means that a Robot Hand is not defined by a single motion, but by a complex interaction of multiple decisions and actions occurring simultaneously. At the core of this challenge lies the role of compact, high-performance drive solutions such as the Micro Actuator, which directly influences how effectively a robotic hand can respond to real-world variability.

The development of a Robot Hand inevitably faces a range of structural limitations.

The hand itself is small, the required motions are numerous, and at the same time, it must be capable of delivering sufficient force. Within these conflicting constraints of limited space and required output, one of the first components robotic engineers must carefully consider is the actuation system.

Those working in the robotics industry are likely already aware of this reality. The range of actuators available on the market today is quite limited, and the situation becomes even more challenging in the Micro Actuator segment. Aside from a few products offered by specific companies, there are very few practical alternatives. As a result, many engineers encounter significant obstacles during real-world product development.

To address these challenges, we have continuously invested in research and development. In 2025, we introduced the BCSA V4 Series, a drive solution designed for humanoid robots. Today, we are pleased to present the next step in this journey.

Introducing the BCSA Micro Series, a cycloid-based Micro Actuator developed to meet the demanding requirements of compact robotic systems, including advanced Robot Hand applications.

BCSA Micro Series

While the BCSA V4 Series serves as a core module for humanoid robot upper and lower body systems, the BCSA Micro Series has been specifically engineered for Robot Hand actuation.

This model features an 18 mm diameter and a 25:1 reduction ratio. It is a rotary Micro Actuator based on a cycloid gear mechanism, designed to transmit rotational force directly within robotic finger joints. This structure enables fast response while maintaining durability suited for repetitive motion.

This project was developed in collaboration with our partner, Company P, with a clear focus on real-world Robot Hand integration. The actuator was designed to be directly embedded inside the joint, reducing internal layout constraints and providing greater flexibility in overall hand architecture.

The design of a Robot Hand involves severe spatial limitations. For this reason, simplifying the mechanical structure while ensuring stable torque transmission and fast response under repetitive operation is critical. The challenge is not merely making components smaller, but creating a structure that can reliably operate within the confined space of a robotic hand.

To address this challenge, we defined one key objective from the beginning: torque density.

What Is Torque Density?

Torque density is a metric that describes how much force an actuator can generate relative to its size or weight. It does not simply refer to the absolute level of torque, but rather how efficiently that torque is produced within a limited volume and mass. For designers of compact robotic systems such as a Robot Hand, torque density is a critical indicator of practical performance.

A simple analogy helps clarify this concept. Torque density is typically calculated by comparing the maximum torque an actuator can deliver to its weight. Consider two scenarios: a 100 kg person lifting a 100 kg barbell, and a 50 kg person lifting an 80 kg barbell. In terms of absolute force, the first case may appear stronger. However, when evaluated relative to body weight, the second case demonstrates far greater efficiency. Actuators that exhibit this type of performance are precisely what robotic engineers look for when selecting a Micro Actuator for space-constrained applications.

This difference has a significant impact on the overall robot architecture. As actuators placed within the Robot Hand become heavier, the load they impose must be supported by larger actuators in upstream joints. This leads to reinforced frames and increased battery capacity, creating a chain reaction that unnecessarily enlarges the entire system.

By contrast, applying a lightweight Micro Actuator capable of delivering sufficient torque changes the equation. The load burden at the robotic hand is reduced, enabling a simpler mechanical structure and faster, more stable motion. As a result, both system efficiency and design flexibility are improved. For these reasons, torque density serves as a key benchmark that defines how far the capabilities of robotic technology, especially in Robot Hand design, can be extended.

The Importance of the Robot Hand and the Role of the Micro Actuator

Many people envision humanoid robots as machines that closely resemble the human form, with the expectation that they will eventually be capable of performing tasks traditionally handled by humans. Since most industrial environments have been designed around human workers, adapting robots to existing environments is far more effective than redesigning entire facilities to suit robots. While this may sound straightforward, it represents one of the most critical challenges in real-world manufacturing automation.

The automotive industry provides a clear example. Despite its high level of automation, many tasks are still performed by human workers. This is not due to a lack of technology, but rather the structural complexity of the industry. A wide variety of parts, frequent changes in vehicle models and options, and ongoing model updates make it difficult to rely solely on dedicated automation equipment. In such environments, automating each individual task with specialized machinery quickly reaches its practical limits.

Traditional robotic automation performs extremely well when tasks are fixed and repetitive. However, it is inherently weak when it comes to change. For a single operation, robotic arms, dedicated tools, feeding systems, and inspection equipment must all be custom designed. Once the task changes, the value of that equipment drops sharply.

So what happens when humanoid robots are introduced into these processes? The burden of redesigning production lines and building dedicated equipment can be significantly reduced. Because humanoids share a human-like structure, they can be deployed into existing work environments with minimal modification. Instead of designing or replacing end effectors for each process, humanoids can use the same tools and components designed for human hands. This is where the performance of the Robot Hand, supported by compact and capable drive solutions such as the Micro Actuator, becomes a defining factor in practical humanoid deployment.

Of course, it would be difficult to say that current technology has already reached this stage. Expectations surrounding humanoid robots are less about immediate deployment on the factory floor and more about the direction in which automation is expected to evolve. These expectations serve as a reference point for anticipating how the structure and approach of today’s manufacturing automation systems may change over time.

In this context, the Robot Hand emerges as a decisive factor in defining the practical scope of humanoid robots. In assembly processes that require delicate manipulation, or in environments where tasks change frequently, the dexterity and drive performance of the hand directly determine whether automation is feasible. This is precisely why interest in Robot Hand technology has been rapidly increasing across the humanoid robotics industry.

To realize these capabilities in practice, one technology is essential: the Micro Actuator. Fitting multiple joints within the limited size of a robotic hand requires not only compact dimensions and simplified mechanical structures, but also the ability to support multi-degree-of-motion configurations and sufficient durability for repetitive operation under real working conditions. At the same time, the actuator must still deliver the output torque demanded by industrial applications.

The human hand is small and lightweight, yet it is capable of generating significant force while maintaining delicate control. Each finger moves independently, while still achieving highly coordinated motion. The challenge lies in fitting all of this into a confined space. Actuation systems must be housed inside the fingers, force transmission mechanisms must be accommodated, and sensors and wiring must be integrated alongside them. Even a slight increase in size or weight can drastically reduce the overall usability of the Robot Hand, making compact and efficient Micro Actuator design a critical requirement rather than an option.

<BONSYSTEMS Robot Hand Project Concept Design>

For these reasons, the Robot Hand is an area that cannot be solved by hardware alone. Fingertip sensor performance, control stability for adjusting force based on contact feedback, and gripping strategies capable of handling a wide range of objects must all work together as a unified system.

Decisions such as how to grasp an object at a specific angle, how to regrip when slipping occurs, or how to disengage safely when excessive force is detected during insertion ultimately extend into the domain of data and learning. In this sense, the Robot Hand is not just a mechanical component, but a critical interface that directly connects the goal of general-purpose manipulation with learning-based control architectures.

How efficiently force can be generated within a limited space will become a key criterion that defines the future competitiveness of Robot Hand technology. Compact and high-performance drive solutions such as the Micro Actuator play a central role in this challenge. This Robot Hand project marks the starting point of that journey, and we plan to share more detailed insights through upcoming demonstrations and real-world application cases.

For those interested in robotic manipulation and compact actuation technologies, we invite you to follow the progress of this project by subscribing to our YouTube channel. Thank you.

Frequently Asked Questions (FAQ)

※ References

1. The Humanoid Mission in Manufacturing | Boston Dynamics Tech Talk | Boston Dynamicd (2025.12.17.)

2. PIM KOREA invests in Bonsystems to boost humanoid robot components | ChosunBiz (2025.12.02.)