Only visible observations or spy devices were able to track an indoor suspect live. But what techniques are most advanced to track in real-time indoor? Shall we focus on the wireless routers RF parameters but missing the movement from room to room? Who has experience in advanced indoor location tracking? As its important for investigations to devide between private and public areas strictly its a potential to digitally observe with available resources in use.
Worth to consider that indoor tracking is also the domain of smart home solutions and devices on the market. At the end we focus on a timeline of movement to overlay with points of actions (e.g. workspace, office, kitschen, bathroom)
Your experience is worth spreading to learn from you -)
Wifi tracking
Wifi tracking requires multiple APs and to run the obs silently is a space problem e.g. the TL-WR802N is unpacked small but requires 5V microUSB power feed by a slim powerbank. Is OpenWRT ablet to mesh-connect multiple devices and log simultaneously?
Not exactly what you need Rolf, but worth a look to get ideas
https://
The Framework for Internal Navigation and Discovery (FIND) allows you to use your (Android) smartphone or WiFi-enabled computer (laptop or Raspberry Pi or etc.) to determine your position within your home or office. You can easily use this system in place of motion sensors as its resolution will allow your phone to distinguish whether you are in the living room, the kitchen or the bedroom, etc. The position information can then be used in a variety of ways including home automation, way-finding, or tracking!
Simply put, FIND will allow you to replace tons of motion sensors with a single smartphone!
The system is built on two main components - a server and a fingerprinting device. The fingerprinting device (computer program or android app) sends the specified data to the machine learning server which stores the fingerprints and analyzes them. It then returns the result to the device and stores the result on the server for accessing via a web browser or triggering via hooks.
Excellent hint, Greg! Thank you - helps a lot for our obs
Single! Access Point (AP) location algo was presented at USENIX '16 conf from CSAIL Computer Science and Artificial Intelligence Laboratory (of MIT) called 'Chronos'. See algo description/calc and slides (pdf) here
MIT article around see here
Thanks Rolf , interesting design.
This reminds me of another far less sophisticated method that was attempted to be rolled out during the London Olympics. Scanning for users smartphone Mac Addresses using APs placed in recycling bins. It adopted the cookie based approach.
http//
After public concerns over privacy etc. the idea was dropped. The start-up marketing company's website doesn't appear to be available now (unless you use wayback..), but the designers of the technology are still around ( http//
Thanks Greg! We evaluate the Lenovo R2105 in decaped form. Who runs actually these devices? ESD-shielding seems to be poor and highly problematic (I am aware to decap has neg. side effect of ESD) but both 2.4 and 5GHz and PoE capable.
While I find the paper
https://
an extremely interesting work ) , and - strangely enough for a paper of this kind - acctually understandable, it seems to me like it is (as often happens) very problematic to be actually used in practice.
Particularly, I can see it more likely to be developed in an effective geo-fencing system but much less likely as an accurate location detection system.
As well, the use for a personal drone (should actually such a device have any use 😯 ) seems to me very nice.
But a rate like 94.7% of accuracy is - even if at first sight it does not seem so - is very, very low, especially if measured as "accuracy against catastrophic failure".
In a part of the article the theme is "accuracy of the measurement", i.e. how different is the measured distance from the actual distance, said to "localize clients to within tens
of centimeters. ", like
• Chronos computes the time-of-flight with a median er-
ror of 0.47 ns in line-of-sight and 0.69 ns in non-line-
of-sight settings. This corresponds to a median distance
error of 14.1 cm and 20.7 cm respectively.
• Chronos enables a WiFi device (e.g., an AP) to localize
another with a median error of 65 cm in line-of-sight
and 98 cm in non-line-of-sight settings
but the 94.7% accuracy is related to being able to correctly locate a device inside a whole room, which in layman terms means in practice plus or minus 3 meters.
In a geo-fencing system 95 to 97% accuracy, i.e. risking to have 3 or 5 out of 100 customers of the cafe' angry because they cannot connect may even be acceptable (actually enlarging the area allowed will make possible a much lower rate of angry customers with a very minor risk in the corresponding increase of "leechers").
But in a "geo-tracking" use, and particularly for "surveillance" use the similar 94.7% is IMHO mostly meaningless, and the math for conditional probabilities ( Bayes’s theorem) would probably make mincemeat of the findings in a court.
jaclaz