WEEK ONE
The first week was spent in investigating proof of concept for the project.
There were some unanswered questions before we finalized the project:
Will these sensors work in all conditions? If not, what are the various failure modes.
Are these sensors consistent?
Is the range good enough for the obstacle sensing system? What are the desired parameters? (range, height, angles)
The first question: Will these sensors work in all conditions? If not, what are the various failure modes.
The sensor was connected in a simple circuit involving a battery and multimeter. The KnobBots' base was used to power the sensor and wires were soldered to the signal line of the output to read the voltage. The voltage was dependent on the distance and data was collected at different angles and the height was varied. Different environments affected the sensor but the deviation in sensor behavior was minimal. While surfaces like ground, carpet, wood, cardboard, cloth and wet concrete produced even results, glass surfaces produced uneven readings. After putting on my 'evil thinking cap' to confuse the sensor, I set up the sensor to read values from a glass on a black surface, with light shining onto the sensor. This set up produced interesting results; The voltage values were consistent when the angle from the vertical was below 30 degrees. Larger angles produced distortion that led to inconsistent data values. The above Figure 1 depicts the behavior of the sensor at different angles starting with Column B at 0 degrees until Column J at 60 degrees with 15 degree increments.
Thus, we can infer that the sensor works almost 90% of the time, except in very bight conditions where a light source is facing the sensor.
There were some unanswered questions before we finalized the project:
Will these sensors work in all conditions? If not, what are the various failure modes.
Are these sensors consistent?
Is the range good enough for the obstacle sensing system? What are the desired parameters? (range, height, angles)
The first question: Will these sensors work in all conditions? If not, what are the various failure modes.
The sensor was connected in a simple circuit involving a battery and multimeter. The KnobBots' base was used to power the sensor and wires were soldered to the signal line of the output to read the voltage. The voltage was dependent on the distance and data was collected at different angles and the height was varied. Different environments affected the sensor but the deviation in sensor behavior was minimal. While surfaces like ground, carpet, wood, cardboard, cloth and wet concrete produced even results, glass surfaces produced uneven readings. After putting on my 'evil thinking cap' to confuse the sensor, I set up the sensor to read values from a glass on a black surface, with light shining onto the sensor. This set up produced interesting results; The voltage values were consistent when the angle from the vertical was below 30 degrees. Larger angles produced distortion that led to inconsistent data values. The above Figure 1 depicts the behavior of the sensor at different angles starting with Column B at 0 degrees until Column J at 60 degrees with 15 degree increments.
Thus, we can infer that the sensor works almost 90% of the time, except in very bight conditions where a light source is facing the sensor.
Figure 2
Are these sensors consistent?
Data was collected with four sensors with different values on incidence( angles) and height from the ground.
The graphs obtained suggest that the sensors are consistent up to 45 degrees; angles such at 60 and so on lead to high distortion in data values as seen in Figure 2.
The different columns represent the four different sensors at varied angles. As we can observe from the graphs, the sensors behave consistently at angles up to 45 degrees but we notice a distortion in values around 60 degrees.
Is the range good enough for the obstacle sensing system? What are the desired parameters? (range, height, angles)
Looking at all the sensor data and the robot's physical constraints, we decided to place the sensor around 25cm off the ground; this ensures maximum range for the sensor.
We could push the angle from the vertical to about 45 degrees for maximum range and thus the system includes five sensors on a mount pointing at different points on the ground at angles ranging from 35 degrees to 45 degrees.