Electronic Circuits and Controls
Bicep Sensor
The input from the voltage divider configuration of the Bicep Sensor goes into the (+) terminal of the LM339 U1A. It then outputs a signal with twice the amplitude into Resistor R28.
High frequency noise was found in this output and caused indistinguishable signals. Discrete signals are used to set conditions and commands in the Arduino’s programming. An oscillating signal could oscillate above and below condition thresholds and cause the Arduino to output random commands. Resistor R28 and Capacitor C1 were implemented into the circuit to form a 1Hz low pass filter. The very low 1Hz cutoff was chosen to overcompensate for much higher frequencies. There are two reasons why this was done. The first was because the exact frequency range was unknown. Ideally, a DC signal characteristic was desired. Therefore, rather than filtering out only the frequency range of the noise, the intent of the low pass filter was to remove as many oscillating components from the output as possible. Only signals approaching a DC characteristic (less than 1Hz in frequency) were allowed to pass through. The second reason for choosing a 1Hz cut-off was that choosing a cut-off too close to the actual frequency of the noise can result in poor filtering. Real life filters do not have perfect signal attenuation at the cut-off region.
High frequency noise was found in this output and caused indistinguishable signals. Discrete signals are used to set conditions and commands in the Arduino’s programming. An oscillating signal could oscillate above and below condition thresholds and cause the Arduino to output random commands. Resistor R28 and Capacitor C1 were implemented into the circuit to form a 1Hz low pass filter. The very low 1Hz cutoff was chosen to overcompensate for much higher frequencies. There are two reasons why this was done. The first was because the exact frequency range was unknown. Ideally, a DC signal characteristic was desired. Therefore, rather than filtering out only the frequency range of the noise, the intent of the low pass filter was to remove as many oscillating components from the output as possible. Only signals approaching a DC characteristic (less than 1Hz in frequency) were allowed to pass through. The second reason for choosing a 1Hz cut-off was that choosing a cut-off too close to the actual frequency of the noise can result in poor filtering. Real life filters do not have perfect signal attenuation at the cut-off region.
Tricep Sensor
Due to its ability to detect large forces this sensor was used to detect forces exerted by tricep contraction while neglecting forces exerted by the weight of the test subject’s arm. The load exerted by the arm was ignored using a comparator. Image above shows a LM339 U2A as a non inverting amplifier with a gain of 2. A low pass filter formed by R32 and C2 remove oscillating noise. A comparator formed by LM339 U4A outputs 5Volts when the voltage input at the (+) terminal is greater than that at the (-) terminal. The voltage input at the (-) terminal is the threshold of the comparator and is controlled by a variable potentiometer labeled Variable Weight Reference.
Monitoring Range of Motion: Flex Sensor
The flex sensor is used to monitor range of motion and set commands accordingly. Its output range is from 0 to 2.5V and it is extended to 5V by the use of a non inverting amplifier as shown in previous sensor set ups. A low pass filter is also used to remove noise. Motors are commanded to stop if the range of motion, denoted by a voltage input 3 shown. The flex sensor is attached to the test subject’s elbow and bends along with their arm by means of an elbow brace shown below. The effectiveness and reproducibility of the bend sensor in detecting changes in angle was determined by statistical analysis.
Monitoring Range Of Motion: Variable Pot
The variable potentiometer is the simplest sensing component used in this device. It is designed to divide voltages by turning a flat knob on its exterior. Its intended used is for easily controlling voltage amplitudes as a voltage divider where the amplitude is divided by a value determined by the ratio of two resistors. The exception in a variable potentiometer is that the ratio of the resistor values is easily adjustable by turning its exterior knob. F. Due to its lack of complexity, its output has unnoticeable noise and has very high reproducibility. It allows a variable potentiometer, whose knob mechanically turns in sync with the rotation of the motor shaft, is used to trace the position of the elbow frame as a change in angle. Although there are sensors on the market that are able to do this, this method eliminates the need for purchasing one and is just as effective in serving its intended purpose. The advantage with using a potentiometer is that it is very cheap (Price range is in cents). It also allows for monitoring of range of motion with very high accuracy and precision using a very simple circuit component. More complex versions of this sensor are not necessary in this application. As the motor shaft turns and the angle of the elbow joint varies, the voltage output of the potentiometer, Variable Pot, changes The voltage-angle relationship is used to monitor position. The output of the Variable Pot is used to set limits to range of motion and program stop commands to be carried out when these limits are reached. Motors will stop upon receiving these stop from the Arduino via the motor controller. The output is a discrete signal and not a range of values as it was for the flex sensor. Hysteresis was, however, found when sensor was not properly secured in place. Statistical tests were conducted to determine reproducibility of this sensor. The results are discussed in the Testing Results section of this website.
Shoulder Actuation
Shoulder extension is detected by movement against an interlink FSR. It is set up in a similar way as for the bicep contraction in a voltage divider configuration and with an amplifier providing a gain of 2. A 1Hz low pass filter is implemented for noise reduction. The output is fed into input 7 of the Arduino. The signal is then converted into a PWM signal that controls speed and direction of the motor. The motor then actuates the arm frame outward away from the body.
Shoulder relaxation takes place passively. When the shoulder extension sensor is not pressed, the gears allow the shoulder to slowly go back down to its resting position by motion caused by its own weight. For the sake of simplicity and due to time constraints, actuation for shoulder relaxation was designed as a passive mechanism. If the test subject wished to slow down or stop while on their way down, lightly touching against the sensor keeps the arm frame suspended. Applying more pressure causes it to extend the shoulder. This motion might be more natural and comfortable for individuals operating the device. Passive shoulder relaxation is the biomechanical mechanism by which the shoulder is usually lowered. During shoulder relaxation, the deltoids are simply relaxed and the arm goes back down to their resting position with little effort. The same concept was implemented in the actuation of the shoulder joint. Also, since there is no active motor actuation in the downward direction, the risk of the arm frame pressing too close against the wheelchair or the test subject while at rest is eliminated.
Shoulder relaxation takes place passively. When the shoulder extension sensor is not pressed, the gears allow the shoulder to slowly go back down to its resting position by motion caused by its own weight. For the sake of simplicity and due to time constraints, actuation for shoulder relaxation was designed as a passive mechanism. If the test subject wished to slow down or stop while on their way down, lightly touching against the sensor keeps the arm frame suspended. Applying more pressure causes it to extend the shoulder. This motion might be more natural and comfortable for individuals operating the device. Passive shoulder relaxation is the biomechanical mechanism by which the shoulder is usually lowered. During shoulder relaxation, the deltoids are simply relaxed and the arm goes back down to their resting position with little effort. The same concept was implemented in the actuation of the shoulder joint. Also, since there is no active motor actuation in the downward direction, the risk of the arm frame pressing too close against the wheelchair or the test subject while at rest is eliminated.