'headless' recorder for replay attack with hackrf

[Digitally Colourful Mistifier]

Hi, we have been engaged for a pentest and we would like to build a device that will allow us to

1) drop an SDR in the vicinity of the radio-controlled gate of our client

2) the SDR should be listening for keys constantly, but only record when there really is traffic. For this I think the device should constantly buffer the past say 15sec and store those 15sec only to file if they actually contain significant data.

3) remote control the device to monitor the number of collected samples, battery-level and potential discovery of the device by security personnel

I was thinking of a raspberry pi on a batterypack with a usb gprs-modem. But would the pi's CPU be fast enough to handle the samplerate from the hackrf? How long would a decent batterypack last with a pi, a gprs-modem and an hackrf running on it? Are there any examples in GNURadio that I could take a look at for this scenario?

Anyone ever done such a thing before?

Thanks!

[Digitally Colourful Mistifier]

This blog https://medium.com/@rxseger/sdr-first-project-initial-setup-node-hackrf-gnu-radio-on-linux-os-x-rpi-3-w-fm-tuner-ee16cdc8fd82
suggests that the PI's ARM cpu indeed is too slow to handle the recommended minimum samplerate of 2e6 on the HackRF. He did have some success running on 1e6 samples/sec it seems.

[NotPike]

Never used one but have you ever heard of the BeagleBone Black? Seems to be the more popular choice when it comes to ARM computers and SDR.

[Digitally Colourful Mistifier]

I heard about it before. People in our hackerspace used them for some projects, forgot about it. Will look into it!