Raspberry pi bluetooth sniffer

jkeefe / raspberry-pi-ble-sniffer.md

in order to get pyserial to talk with the usb port, apparently need to do this, to add the «pi» user to the dialout group:

sudo usermod -a -G dialout pi 

ok, so now I need a port to listen to . where the sniffer device is. so apparently I do this:

which lists a whole bunch of things I don’t understand, except that the last lines are:

crw------- 1 root root 4, 7 Dec 31 1969 /dev/tty7 crw------- 1 root root 4, 8 Dec 31 1969 /dev/tty8 crw------- 1 root root 4, 9 Dec 31 1969 /dev/tty9 crw-rw---- 1 root tty 204, 64 May 16 13:09 /dev/ttyAMA0 crw-rw---T 1 root dialout 5, 3 Dec 31 1969 /dev/ttyprintk crw-rw---T 1 root dialout 188, 0 Dec 31 1969 /dev/ttyUSB0 

Tried the AMA0 one, which didn’t work. The USB0 one, did!

sudo python sniffer.py /dev/ttyUSB0 

Scanning for BLE devices (5s) . Found 2 BLE devices: [1] "" (B8:78:2E:1C:2F:10, RSSI = -104) [2] "MetaWear" (EA:A1:88:31:45:21, RSSI = -88) Select a device to sniff, or '0' to scan again > 

Can this be used with ubertooth one?

I have tried this on pi3, but the process hung. I don’t have Adafruit’s BLE sniffer device, just use onboard bluetooth.

crw—w—- 1 root tty 4, 8 May 4 03:17 /dev/tty8
crw—w—- 1 root tty 4, 9 May 4 03:17 /dev/tty9
crw-rw—- 1 root dialout 204, 64 May 4 03:17 /dev/ttyAMA0
crw——- 1 root root 5, 3 May 4 03:17 /dev/ttyprintk

pi@raspberrypi:~/Adafruit_BLESniffer_Python $ sudo python sniffer.py /dev/ttyAMA0
Capturing data to logs/capture.pcap
Connecting to sniffer on /dev/ttyAMA0
Scanning for BLE devices (5s) .

Still having the same issue in 2021. Here is what came from the log:
12-Jan-2021 11:13:00 (+0000) INFO: Software version: SVN rev. 1111
12-Jan-2021 11:13:00 (+0000) INFO: args: ()
12-Jan-2021 11:13:00 (+0000) INFO: kwargs: >)]>
12-Jan-2021 11:13:00 (+0000) INFO: board ID (random): 120
12-Jan-2021 11:13:00 (+0000) INFO: Error on port /dev/ttyAMA0. file: Packet.py, line 257: Port is already open.
12-Jan-2021 11:13:02 (+0000) INFO: Error on port /dev/ttyAMA0. file: Packet.py, line 258: Write timeout
.

Источник

Sniffing Bluetooth Devices With A Raspberry Pi

Hackaday was at HOPE last weekend, and that means we got the goods from what is possibly the best security conference on the east coast. Some of us, however, were trapped in the vendor area being accosted by people wearing an improbable amount of Mr. Robot merch asking, ‘so what is Hackaday?’. We’ve all seen The Merchants Of Cool, but that doesn’t mean everyone was a vapid expression of modern marketing. Some people even brought some of their projects to show off. [Jeff] of reelyActive stopped by the booth and showed off what his team has been working on. It’s a software platform that turns all your wireless mice, Fitbits, and phones into a smart sensor platform using off the shelf hardware and a connection to the Internet.

[Jeff]’s demo unit (shown above) is simply a Raspberry Pi 3 with WiFi and Bluetooth, and an SD card loaded up with reelyActive’s software. Connect the Pi to the Internet, and you have a smart space that listens for local Bluetooth devices and relays the identity and MAC address of all Bluetooth devices in range up to the Internet.

The ability to set up a hub and detect Bluetooth devices solves the problem Bluetooth beacons solves — identifying when people enter a space, leave a space, and with a little bit of logic where people are located in a space — simply by using what they’re already wearing. Judging from what [Jeff] showed with his portable reelyActive hub (a Pi and a battery pack) a lot of people at HOPE are wearing Fitbits, wireless headphones, and leaving the Bluetooth on the phone on all the time. That’s a great way to tell where people are, providing a bridge between the physical world and the digital.

• 
• 
• 
• 

Posted in Raspberry Pi, Wireless Hacks Tagged ble, bluetooth, bluetooth LE, raspberry pi, reelyActive

18 thoughts on “ Sniffing Bluetooth Devices With A Raspberry Pi ”

Any _well implemented_ personal Bluetooth device will use private, random + temporary, MAC addresses so this won’t work for identifying people.

Источник

Saved searches

Use saved searches to filter your results more quickly

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Sniffer for Bluetooth LE on Raspberry Pi

License

IanMercer/pi-sniffer

This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Sign In Required

Please sign in to use Codespaces.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching Xcode

If nothing happens, download Xcode and try again.

Launching Visual Studio Code

Your codespace will open once ready.

There was a problem preparing your codespace, please try again.

Latest commit

Git stats

Files

Failed to load latest commit information.

README.md

pi-sniffer aka Crowd Alert

This project counts people! It can figure out how many people are present in a given area using a handful of very cheap sensors. It achieves this by counting cellphones. Since most people carry one these days, and since iPhones automatically turn Bluetooth back on every day, there are enough people in most populations with active Bluetooth signals that we can calculate a total number of people with a high statistical significance.

This is not without challenges as cell phones randomize their mac address and a single person may have multiple bluetooth devices on them.

Using the count of people present you can power displays that warn people when it’s too crowded. This capability can be used to reduce crowding to help prevent the spread of COVID-19. It also gives people confidence that the store, restroom, railway carriage or other enclosed space they are about to enter is not crowded, helping us all feel confident again going about our daily lives.

NOTE: A major revision is in the works splitting all code concerned with sending data to MQTT / WebHooks / Azure IoT etc. into separate services communicating over DBUS to the main ‘sniffer’ service. The DBUS interface definition can be found in `src/dbus`. This will allow other languages (Python, Rust, . ) to be used more easily to consume the person data produced by this service.

The code here recently won first place in the global BetterHealth Hackathon organized by HCL and Microsoft.

But this system isn’t limited to crowding signs. You can use the data it collects in many other ways: as an input to a smart home controlling lighting and heating based on how many people are home and which areas of the home are occupied, or as a feed to a marketing analysis system, .

Not only does this project count how many phones are present, it tracks every other device that comes into range and can send data back over MQTT about them. For devices other than cellphones, i.e ones with either defined names or public (unchanging) mac addresses you can use this information to trigger actions. For example, put an iBeacon on your car and trigger an action when it comes home or after it’s been gone for a set time.

It uses the built-in Linux BLUEZ libraries and the Bluetooth antenna on any Raspberry Pi (Pi3 is recommended but also runs on PiW and Pi4) to scan for nearby BLE devices. But it’s also written to be as portable as possible using only the C language and avoiding dependencies unless absolutely necessary.

It reports all BLE devices found (Mac address, name, type, UUIDs, . ) and their approximate distance to an MQTT endpoint. It applies a simple Kalman filter to smooth the distance values. It also handles iPhones and other Apple devices that randomize their mac addresses periodically and can give a reliable count of how many phones/watches/. are in-range.

The system is also designed to be fault-tolerant. It runs just fine even when the internet is down. In a system with multiple sensors and multiple displays the loss of any one component should have minimal impact on the whole system.

• Power displays that warn when a confined space is getting too crowded to encourage people to ‘socially distance’ by coming back later or picking a less crowded store / room / railway carriage / bus / .
• Detect cell phones entering your home, garden, barn, .
• Put the heating or air conditioning on when there are two or more cellphones in the house and off otherwise
• Locate cars, dogs, . using iBeacons attached to moving objects (reverse of iBeacon normal usage)
• Gather other advertised data and transmit to MQTT (temperature, fitbit, cycleops, . )

• Scan for BLE devices nearby a Raspberry Pi using the built-in Bluetooth adapter
• No external dependencies: no Python, no Node.js, no fragile package dependencies, easy portability
• Simplicity: do one thing well, no frills

iOS MAC address randomization

Tracking iOS (and many Android) devices is complicated by the fact that they switch Mac addresses unpredictably: sometimes after a few seconds, sometimes after many minutes. You can see two MAC address swaps in this example:

There is no (easy) way to distinguish a MAC-flipping event from a new device arrival event. Until the old mac address pings again they could be the same device. The pi-sniffer code includes an algorithm to calculate the minimum possible number of devices present assuming that any overlap in time means two sequences are different devices, but otherwise packing them all together like events on a calendar to find the minimum possible number of devices present.

Pi-sniffer transmits this min-count every time it changes so you can easily see how many devices are in range of any of your RPi devices.

It actually transmits a count for each range (1, 2, 5, . )

The MQTT packet is now JSON encoded. It includes the property that has changed and a timestamp. Properties may include name , distance , alias , power , type , uuids , serviceData , manufacturer , manufdata , temperature , humidity .

The MQTT topic is of the form: BLF//messages/events/

For a summary of all devices seen by the access point the following topic is sent:

BLF//summary/dist_hist -- an array of bytes containing the count of devices at each range 

An up message is also sent on startup.

Given delays in MQTT transmit, receive and re-transmit to the receiving application it’s a good idea to use the timestamp passed in the packet. Make sure all your Pis are synchronized to the same time.

• Recently updated to use the Eclipse PAHO MQTT C library which support SSL and Async code. This is now the only dependency.
• This is a work in progress and is still changing fairly rapidly.
• There is a build.sh file that builds and runs the code.
• There is a log.sh file that tails the log while it’s running.
• The MQTT topic prefix, server IP (or FQDN) and port are configurable.
• Environment variables are also used to configure the RSSI to distance conversion parameters for indoor/outdoor settings.
• Multiple instances communicate over UDP on port 7779 (also configurable).
• The total number of people present x 10.0 x a configurable scale factor is sent over UDP 7778 (also configurable). (It’s actually the expectation of the probability curve for the number of people present so after a period of inactivity it goes down to zero following a curve.)
• A separate ESP8266 project is available that listens to the port and displays a crowd warning indication

• Decode advertised data for common iBeacons that also send environmental data (e.g. Sensoro)
• Gather other advertised data and transmit to MQTT including temperature, battery, steps, heart rate, .
• Combine multiple distance values to do trilateration and approximate location, simple ML model

Please see the FAQ for more information.

Источник