| Concluding Remarks: |
| A system for finding the range and azimuth of a sound signal was designed and
implemented in software. This was done in the near-field, where the angles at which the
signal reaches the microphones can be resolved. Specifically, the differences in the times
at which the signal reaches two microphones can be used to find the azimuth of a signal.
However, in a two-microphone system, only the azimuth can be found. To determine range,
more microphones must be added. In this particular project, a four microphone system was
implemented to determine range in addition to azimuth. Hardware for this system could not be developed primarily because of economic factors that prevented an actual implementation. However, the software for the system is fully functional and can determine range and azimuth for a signal. Only an interface is required to implement SODAR. In the two microphone system, we experienced the endfire phenomenon, a loss of resolution collinear to the microphones. This was resolved by adding more microphones. As a result, resolution remained consistent around the entire array. Another limitation that was not resolved was the tradeoff between frequency sensitivity and resolution. We could improve frequency sensitivity at the cost of resolution or improve resolution at the cost of frequency sensitivity. This is a fundamental limitation of any sort of similar system. The system worked as designed, however, the frequency/resolution tradeoff was irksome. Although this tradeoff becomes acceptable in different media (such as underground or underwater), it makes a physical implementation in air infeasible. However, it does confirm the use of SONAR systems where they are implemented in actuality. |