Project Diagnosis
It could be argued that the most valuable asset for any human being is their own body. Replacing a missing human limb, especially a hand, is a challenging task that makes us appreciate the complexity of the human body.
Throughout history, innovators have tried to replace lost limbs with artificial elements similar to the natural ones. These early designs, although orthodox and functional to some extent, were primarily aesthetic and not fully functional.
History has shown us that in ancient civilizations, prosthetic devices have been found that demonstrate the continuous progress of technology. Until recently, prosthetic design had been relatively slow, from a simple wooden prosthesis to the current prostheses that interact with our biological organism, such as the prosthesis presented in this project, which is a brain-machine combination.
Proposal and Development
Direct Brain Interface
The most advanced form of control is a direct brain-computer neural interface. A surgical procedure places electrode array matrices on the surface of the brain that are connected to pedestals implanted in the patient’s skull. When the patient thinks, the movement signals detected in the pedestals are used to control the movement of a robotic arm. This technology is still in its infancy, but has demonstrated that people with disabilities can control bionic devices using only their thoughts.
Physical Design
The complexity of the mechanical and electrical systems determines how well the device mimics the human arm and the level of dexterity it can offer. The design will aim to be as physically advanced as possible.
Control Scheme
Ideally, the prosthetic should be as easy and natural to control as possible. If the user struggles with basic tasks like grasping, the prosthetic is likely not beneficial.
Practical Utility
The device should be useful for amputees. The goal is to develop a prosthetic that can benefit people missing a hand, regardless of whether it is used in uncertain real-world conditions.}
Affordability
We will try to minimize the material cost of the device. Modern commercial myoelectric prosthetics are very expensive. Both the physical design and the control system will be thoroughly discussed.
Plausible Solutions
The objective is to design and build a 3D printed prosthetic arm. Users will control the arm through muscle flexion detected by surface electrodes on the skin. 3D printing allows rapid creation of detailed mechanical components, which would be much more costly with other fabrication methods. To actuate the device, one approach is an artificial tendon network, which is a relatively simple way to control the fingers. Another option is a rigid joint linkage system, but this is more complex and less feasible with weak 3D printed parts.
Socket Design
The primary objective of prosthetic treatment is protecting the residual limb, considering that 90% of upper limb amputations result from physical trauma. Prolonged pressure on damaged soft tissue areas can significantly compromise the remaining appendage. Amputees may experience various issues such as pain, swelling, blisters, skin irritations, edema, and restricted blood flow.
Sockets must be designed ensuring:
– User safety
– Comfort
– Hygiene
– Optimal distribution of prosthesis weight
The most common method of prosthetic fitting is creating a custom socket that fits the amputation stump. This socket can be self-supporting, vacuum-adjusted, or secured with harnesses. Comfort and load distribution can be improved through:
– Prosthetic socks
– Inflatable air bags
– Reduced density and rigidity in sensitive areas
Several low-cost prosthetic arms use thermoformable plastic to create custom sockets. The process involves heating a plastic sheet and molding it around the amputated stump. A prosthetic sock can be used to achieve optimal fit between the user and the device.
Mechanical Design
To create a functional myoelectric prosthesis, it is essential to develop a mechanical system that efficiently mimics human arm functionality. The mechanical design considers joint activation and forces present in the system. The presented bionic arm can be manufactured entirely using a 3D printer and basic tools.
Initial Ideas and Conceptual Designs
After analyzing various actuation methods for prosthetic arms, an artificial tendon design was selected. Based on literature review of Shadow Hand, UB Hand, and InMoov, artificial tendons have proven to be a viable method for activating bionic hands. Tendons can be any type of high-strength line that maintains tension without stretching. These connect to the fingers and are tensioned by motors located in the forearm, thus allowing fingers to open and close.
The electric motors controlling these tendons must be fully integrated within the device to make it portable and adaptable to the amputee. Ideally, these motors should be located as close as possible to the fingers; however, due to their relatively large size, they cannot be housed in the palm section and must be located in the forearm.
Ergonomics
The field of ergonomics studies the interaction between humans and machines. In prosthetics, ergonomics is particularly significant as it deals with the relationship between prostheses and amputees, including physical attachment to the body and sensory feedback. Ergonomic considerations must also extend to interactions between a person’s prosthesis and others. An ideal prosthesis should be physically comfortable for the amputee, easy and natural to control, provide useful sensory feedback, and interact effectively with its environment.
Design Proportions
A universal goal in prosthetic design is to achieve forms and sizes that match average female physiques.
The dimensions of a large male hand were used for initial design proportions, as it’s easier to scale down a design than to enlarge it. Scalability has been considered throughout the design process, allowing components to be easily resized in computer modeling software and quickly printed. This enables efficient development of prototypes in various sizes.
Design Software
The prototype was designed using Solidworks, a computer-aided design software package created for modeling mechanical components and assemblies. Solidworks is widely used in engineering for design and analysis of mechanical components.
Fingers
Each finger consists of three individually printed components linked together with polypropylene pins. The artificial tendon wraps around the inner fingertip to create a tendon locking point. This tendon runs through channels within the finger to form a closed loop. When the tendon is pulled, rotational forces are applied to all joints, causing the finger to curve.
The tendon locking point is crucial – when the tendon is tensioned, it pulls the fingertip and causes all joints to rotate. Without this locking mechanism, the tendon would slip when tensioned and the finger wouldn’t move. To open the finger from a closed position, tension is applied to the opposite end of the tendon. High-quality braided fishing line is used as it offers minimal stretch when tensioned. Nylon fishing line would stretch over time, resulting in tension loss that would negatively affect finger movements. While biological tendons in the human hand function similarly, they involve many more tendons attached to different bones, allowing more precise finger control.
Thumb
The thumb follows a similar design principle. While most commercial and research prostheses aim to provide at least two degrees of freedom in the thumb, this design provides a single degree of freedom – it can only open/close in one way. Guide holes have been incorporated into both finger and thumb designs to optimize tendon orientation and prevent tendon lines from catching on sharp edges.
Drive System
The tendons wrap around custom 3D-printed servo hooks creating a closed loop. As the servomotor rotates in one direction, it pulls the tendon and closes the finger. To open the finger, the motor rotates in the opposite direction. The artificial tendon drive unit for the index finger is shown below (other tendons omitted for clarity). The thumb, index, and middle fingers are connected to individual servomotors. Due to limited space within the arm, the ring and little fingers are connected to the same servo, meaning they open and close in tandem.
Modularity
Adaptable Design for Different Amputation Levels
Amputation can occur at any point in the arm and each case is unique. An ideal design must facilitate connection with the stump regardless of its location along the arm. The hand and wrist section is designed to adapt to forearm amputations, while the additional elbow module provides a solution for above-elbow amputations.
Socket Connection
The design of a socket connection requires:
– A mold or CAD representation of the amputee’s stump
– Use of thermoformed plastic to mold around the stump
– Implementation of harnesses or straps to ensure a stable connection
Forearm Design
Although the forearm contains no moving parts, its design presents specific challenges as it must:
– House five servomotors
– Contain a lithium polymer (LiPo) battery
– Allow efficient assembly
The forearm was designed in separate sections that are assembled with screws, as a single 3D print would not allow for the installation of motors and tendons. Special design considerations:
– 3D printed ABS plastic is relatively fragile
– Guide holes were incorporated for screws
– Sufficient material was ensured for firm screw grip
Elbow Design
Main features:
– 110-degree rotation range
– Allows straight position and right-angle flexion
– Gear system requiring 290 degrees of servo rotation
– Necessary modifications to increase standard rotation range of 180 degrees
Manufacturing and Assembly
– All components were printed in ABS plastic using an UP 2 3D printer
– A detailed assembly guide is included in the appendix
Mechanical Calculations for the Elbow
Initial estimates indicate:
– Total arm weight: approximately 1 kilogram
– For simplified calculations, a point load of 1 kg acting at 13.5 cm from the elbow pivot is considered
τ = F d
τrequired = 1 ( 9.81 ) ( 0.135 )
τrequired = 1.32 Nm ( 13,5 Kg cm )
The required torque at the elbow is a crucial factor for lifting the robotic arm. TowerPro servomotors have a maximum torque capacity of 10 kg-cm. Theoretically, a 135% increase in torque would be sufficient for this task. However, it is not recommended for servomotors to operate at their maximum capacity, especially during extended periods. The optimal operation should be maintained at 50% of the maximum torque.
Gear System
To increase the effective torque, a gear system was implemented in the elbow. This system consists of:
– A large driven gear connected to the forearm, designed to maximize the available space in the elbow joint.
– A smaller drive gear, dimensioned to ensure teeth robust enough to support high torque transfer.
Finger Force Analysis
To calculate the theoretical force in the fingers, the following scenario is considered:
1. Index finger fully extended
2. Force applied near the fingertip
3. The tendon generates a moment about each joint
The knuckle joint experiences the greatest moment due to its distance from the force application point. Therefore, the torque at this joint determines the maximum load capacity at the fingertip.
Technical Considerations
At the point of maximum load, moments M1 and M2 are in equilibrium. The calculation begins by determining the tensile force in the tendon, considering that the MG996R Servos have a stall torque of 10 kg-cm (equivalent to 1 N/m).
τservo d = F2
F2 = 1 (N/m) 10 mm
F2 = 10 N (Tensión en el tendón)
F1 D1 = F2 D2
F1 = ( 10 (N) 4.5 (mm) ) / ( 65 (mm) )
F1 ≈ 0.7 N
Mass1 ≈ 70 g
When the finger is fully extended, a force of 0.7 N can be applied to each fingertip, equivalent to lifting a mass of 70 g. Although this figure may seem modest, it is crucial to consider that it does not necessarily represent the maximum potential force.
Due to the modification of the perpendicular distance between joints, the fingertip can increase its force application capacity. In this scenario, each finger can support approximately 150 g, providing the entire hand with a lifting or retention capacity close to 600 g.
Finger Actuation Speed
MG996R Servo Characteristics:
– Operating speed: 0.15 seconds per 60 degrees
– Complete wrist rotation (180°): 0.45 seconds
– Calculation: (0.15 s * 180°) / 60° = 0.45 s
Tendon Movement:
– Required displacement: 2 cm (from full extension to total flexion)
– Servo radius: 7 mm
– Rotation angle to open/close a finger: 160°
– Maximum opening/closing time: 0.4 seconds
Electrical Design: Signal Flow
Control Process:
1. User muscle flexion
2. Analog signal generation
3. Amplification, rectification, and smoothing via EMG sensor board
4. Conversion of analog signal to pulse-width modulated (PWM) signal
5. Servo motor activation to tension tendons and curl fingers
Electromyography (EMG) Detection
System Components:
– Single-channel EMG sensor board
– Small-sized PCB
– Three surface electrodes:
- Two electrodes to measure muscle voltage potential
- One ground reference electrode, located on a bony structure
Software and Programming
Development Environment:
– Tool: MPLABX (provided by Microchip)
– Connection: USB programming
Programming Languages:
– Main language: Standard C (Arduino)
– Automatic conversion to assembly language via compiler
Microprocessor Functionality:
– EMG sensor input signal monitoring
– Calculation of required actions
– Signal generation for motor control
Servo Motor Control Signals and Electromyography (EMG)
Servo Motor Control
Servo motor control is performed using a pulse-width modulation (PWM) signal. Every 20 milliseconds, the Arduino microcontroller sends a pulse to the servo’s internal control circuit, with a duration between 1 and 2 milliseconds.
EMG Signal Detection
Each muscle sensor board generates an analog signal in a range of 0 to 3 volts, which is connected to an analog pin of the microcontroller. This device converts the analog signal to digital through a conversion process that generates a 10-bit binary value, used to control servo positioning.
EMG Control Algorithm
System Characteristics:
– Exploration of basic EMG control algorithms
– Use of two pairs of electrodes to generate analog signals
– Activation of specific commands by exceeding predefined thresholds
Control Method:
1. First Set of Electrodes:
– Location: Muscle region (e.g., biceps)
– Function: Monitor myoelectric signals
– Objective: Change between different grip states
* Precision grip
* Force grip
* Wrist/elbow rotation configuration
2. Second Set of Electrodes:
– Location: Alternative muscle region (e.g., forearm)
– Function: Activate specific device states
* Finger closure
* Joint rotation to predetermined positions
System Limitations:
– Execution of only one command at a time
– State transition time
– Impossibility of simultaneous movements
Improvement: Proportional Control
A proportional control was implemented that allows:
– Finger closure proportional to muscle flexion intensity
– Linear variation of PWM signal pulse widths according to EMG signal magnitude
Control Equation:
WPWM(t) = a + k | emg(t) |
Where:
– WPWM(t): Servo signal pulse width
– a: Arbitrary offset (PWM start at 1 ms)
– k: Scale factor
– emg: EMG signal magnitude
Future Perspectives
Current State:
– Low-cost bionic arm prototype
– Not immediately suitable for amputees
– Potential for improvement through collaboration with medical institutes
Contributions and Implications:
– Platform for advanced prosthesis research
– Exploration of:
* Sophisticated EMG control algorithms
* Integrated pressure feedback
* Advanced bio-mechatronic concepts
Potential Developments:
– Leveraging 3D printing technologies
– Improving robustness and durability
– Objective: More functional and ergonomic devices
Technical Characteristics
Weight:
– Total weight: 970 grams
– Comparison:
* Average human arm: 2.5 kg
* Commercial prosthetic arms: Varied
Performance:
– Current limitations:
* Low strength
* Rapid actuation speeds
– Long-term objective: Approach biological human arm capabilities