Findings:
- We can feed detectable features to a neural net or some other form of ML algorithm to increase detection accuracy.
- Drones are very obvious on the RF spectrum, when activated, put out a lot of power
- Features we can use include RF frequency, drone vibration, doppler signature, wavelet analysis, temporal consistency, coefficient variances, body shifting patterns, and moving object detection.
- HackRF sweep samples fast enough that we will be able to detect frequency hopping drones without trouble