Our journey with Alt-Bionics, which began with a university project’s ambition to create an affordable bionic hand, has evolved into a technical venture of sophisticated design and innovation. Building on the foundation established by our previous blog post, which provided a functional overview of the Alt-Bionics prosthetic hand, we’re now set to take you behind the scenes of its technical realization.

We’ll delve into the technical nuances of the Alt-Bionics project, highlighting our design of the Printed Circuit Board (PCB) architecture, which was subsequently manufactured. Additionally, we’ll discuss the firmware improvements and the AI module that streamlines grip control for users.

Overview of the solution

Let’s quickly overview the solution to provide you with the essential context. In collaboration with Alt-Bionics, we’ve created a transformative prosthetic system to improve the lives of amputees.
The system integrates a bionic hand with an array of components and features, and includes a mobile app for seamless control and customization. Each component plays a critical role in replicating the movements, sensations, and capabilities of a real hand, making it a vital tool for those in need of a prosthetic solution. The following list outlines these remarkable features in detail:

  • Bluetooth – With Bluetooth connectivity, the hand wirelessly communicates with the app, providing a user-friendly experience.
  • Motors – At the heart of the bionic hand are motors (actuators), working as the hand’s muscles.
  • Force-sensitive resistors (FSRs) – These embedded sensors provide a sense of touch to the hand.
  • PCB – The PCB is the hand’s brain, it orchestrates control and connectivity of the entire system.
  • Electronic Quick Disconnect (EQD) Wrist Unit – This wrist unit ensures a simple disconnection and reconnection method, allowing for simple integration into existing prosthetic sockets.
  • EMG sensors – EMG sensors inside the prosthetic socket detect muscle movements to control the bionic hand. These connect via a coaxial plug to the PCB for precise muscle signal readings.
  • Prosthetic Socket – The prosthetic socket ensures a stable attachment to the limb, while a rechargeable battery powers the hand, maintaining its readiness for action.

Meeting the Hardware Requirements

The hardware component is crucial in the success of the Alt-Bionics prosthetic hand project, serving as the foundational backbone and ensuring the device fulfills the client’s specific functional requirements.

At the beginning of the project, our engineering team conducted a thorough evaluation of the prototype’s hardware. It quickly became apparent that the existing hardware lacked the necessary computational power and memory capacity to meet functional requirements. This realization led us to embark on a journey of hardware upgrades and optimization.

Custom Printed Circuit Board Development

One of the client’s specific demands was related to the form factor of the printed circuit board. With a vision to reduce the size of the bionic hand, it was imperative for us to correspondingly shrink the size of the PCB. Off-the-shelf PCBs were unable to simultaneously meet the functional requirements (such as the motor control interface and Bluetooth module), performance needs, and form factor specifications.

Our solution? Designing a custom PCB that would seamlessly fit within the prosthetic’s internal structure, making optimal use of the available space. Our skilled engineer rose to the challenge, meticulously crafting a 6-layer architectural design for the PCB. This design seamlessly integrates a Bluetooth module for connectivity, while also incorporating a high-performance processor and sufficient memory to meet the device’s functional requirements. Following this detailed design process, we proceeded to place an order for the manufacture of the newly designed printed circuit board at a factory, ensuring we oversee the entire hardware production cycle. See what our engineer designed PCB looks like in the prosthetic housing.

Firmware Development for the Bionic Hand

First and foremost, it should be noted that our team adheres to an agile firmware development process. The evolution of both functional and hardware aspects underwent several iterations, which is typical for such projects with a research component. With each modification to the prosthetic’s design or new functional requirements from the client or hardware upgrade inevitably cascaded into the need for firmware updates.

  • Firmware Core Functions: Our team developed firmware that introduces three innovative grip performance modes: Sequential, Proportional, and AI. The Sequential mode, enhanced from its original iteration, now permits the bionic hand to cycle through grip patterns in both directions, allowing users to access any of the six installed grips with a maximum of three switches, a significant improvement over the previous one-way navigation system. Our Proportional Mode, on the other hand, enables users to finely adjust the grip in real-time; a slider-like control is achieved by interpreting muscle sensor signals to increment flex and release directions with precision. In AI mode patient activates the most appropriate grip based on the specific muscle activity, a topic that will be explored in greater depth in the subsequent chapter.
  • Bluetooth Component Integration: Addressing the initial separation of the printed circuit board and the Bluetooth module, our innovative design integrated the Bluetooth functionality directly onto the PCB. We implemented the communication logic within the firmware, enabling robust and intuitive interaction between the bionic hand and the mobile application.
  • Safety and Diagnostics: We prioritized safety by embedding comprehensive fault detection within the firmware. It swiftly identifies faults, categorizing them to inform appropriate responses, ranging from user alerts to immediate hand deactivation in critical scenarios. Additionally, we designed the firmware to monitor EMG signals, halting operations if it detects potential overload, thus safeguarding the user.

AI Mode

In AI mode, the prosthesis uses machine learning algorithms to recognize and memorize unique patterns of muscle activity in patients, allowing for the adjustment of prosthetic grips for specific muscle movements.

The technical execution involves the following steps:

  1. Two EMG sensors are affixed to antagonist muscles: one sensor is attached to the Palmar muscle, which flexes the wrist and fingers toward the palm, and the second sensor is connected to the Dorsal muscle, responsible for extending the wrist and fingers away from the palm. Muscle activity is thus transformed into electrical signals that are transmitted to the printed circuit board (PCB).
  2. At the initial activation of AI mode, the patient is prompted to perform a minimum of six exercises targeting the hand muscles. During these exercises, a data flow from the EMG sensors is captured.
  3. The AI module processes the data retrieved from the sensors for each exercise, pinpointing distinctive patterns within the signal stream. These patterns represent the electrical activity of the two EMG sensors during specific muscle movements.

The rationale behind the patient performing precisely six exercises is intricately linked to the design of the current PCB. It contains six memory slots at the logic level, each designed to store a particular grip configuration. These grips may be either pre-set core grips or custom grips fashioned by the user. The allocation of these grips to the memory slots, as well as the creation of custom grips, is managed via a mobile application. Consequently, executing the six exercises yields six unique signal patterns specific to the patient, each derived from the dual EMG sensors.
4. When the user later replicates any of these six muscle movements, the AI identifies the corresponding signal pattern and actuates the grip associated with that pattern, which is stored in the designated memory slot.

Our artificial intelligence module, currently in the throes of development as this post is crafted, showcases promising capabilities in pinpointing these unique patterns with notable precision, a fact borne out by the initial results from our proof of concept. Discover more about our project-specific AI technology in the dedicated blog post, which delves into the technical implementation of the AI module within this project.

Conclusion

As we continue to refine the development of the Alt-Bionics hand, our path has taken us from an initial concept to a cost-effective solution nearing commercial production readiness. We’ve engineered a custom PCB tailored to the unique form factor of the device, and crafted firmware that introduces user-centric grip modes, notably incorporating an AI module for patient experience enhancement.

At the time of writing this blog, the hand is undergoing field trials with actual patients. These trials are crucial in verifying the practicality and user satisfaction of the device. By closely collaborating with our client throughout this phase, we’re dedicated to their vision of democratizing sophisticated prosthetic technology. Our collective aim is to deliver a transformative, economically viable prosthetic hand that has the potential to significantly enhance the quality of life for amputees.

Ready to start the conversation?