Understanding Apple’s WiFiSLAM acquistion and what’s next

“Apple’s acquisition of WiFiSLAM is one of the first acquisitions of an indoor location positioning start-up,” Bruce Krulwich writes for Seeking Alpha. “Since Apple has a tendency to start (or legitimize) trends, it’s interesting to consider other major companies that are entering the area of indoor location, which are leading in their industry, and what start-up acquisition targets might interest them.”

“Apple was already gathering locations of Wi-Fi access points and using them to estimate location, as Google and others are also doing. But this apparently hasn’t been accurate enough,” Krulwich writes. “WiFiSLAM brings two technologies to the table. One is their very fast & easy method of gathering indoor maps & Wi-Fi fingerprints. Their system uses a cellphone camera to take a picture of an indoor map, and a mobile app to gather fingerprints as a phone user walks around the site. This will make it easy for Apple to add Wi-Fi fingerprinting. In fact, it’s similar in spirit to Google’s mobile app that adds Wi-Fi fingerprinting to Google Maps, although Google’s app doesn’t use cellphone cameras to upload maps (yet).”

Krulwich writes, “WiFiSLAM has also developed very accurate sensor fusion, for tracking phone locations using phone sensors (approach 3). If it really works as claimed, it would be very powerful if added to iOS.”

Read more in the full article here.

14 Comments

  1. You cannot map an indoor facility using just an inboard camera. That would not give you spatial dimension vis-a-vis the interior layout.

    Radio waves do not propagate in straight lines. Firstly radio waves propagate in a conical shape depending on the angle of the antenna. The area within the conical shape is called the main beam where signal to noise levels as measured by S/N dB is highest. Radio waves density fall off in strength the further you are from the radio and where you are positioned with reference to the main lobe, side lobe or radio black spot where there is no reception or the S/N ratio is so low as not to be able to receive a cohesive signal.

    Attenuation from the antenna is also non standard as it depends not only on the transmit strength but also on the number and density of structures it has to pass through. So the unknown are (a) the position of the transmitter, (b) the signal strength which it is transmitting, (c) the sensitivity of the antenna that is measuring signal strength and (d) the direction in which the radio wave is propagating.

    In conclusion, you will need a plan drawing of the facility you wish to map with known transmitters plotted on the map before a derivation of location relative to the floor plan can be found.

    1. You seem awfully sure of yourself, BLN. You may be right, but it is also possible that a little knowledge can be highly misleading.

      I have not read the full article, but i can envision a process that uses triangulation via RF and correlates that position information with optical camera data. Even from a single camera, range and scale information can be inferred fairly well, especially for common types of environments containing items/structures with fairly standard sizes (doors, etc.). In addition, a series of camera images can be processed using motion flow techniques to supplement the position information obtained from RF triangulation. Phones like the iPhone can also provide data from other onboard sensors (e.g., gyros). The resulting indoor maps do not have to be perfectly accurate. They just have to be good enough to get you close to your destination. After than, you can locate out the room number or name placard. This is similar to the manner in which GPS gets you to the parking lot and you visually locate a parking space.

      1. GPS uses trilateration not triangulation. Triangulation would be in effect running a radio detector van to triangulate the source of the transmission. Trilateration measures your geoposition in relation to time delay as measured by the atomic click on board the GPS satellite that takes into account minute fluctuations in space time as defined in the general theory of relativity to plot your position on earth. Position and elevation.

    2. You can also build an effective mapping device by connecting Geordi La Forge’s VISOR to a TRICORDER, and then bombarding the signal with neutrino beams.

    3. Radio waves do not propagate in straight lines

      I don’t know what you are quoting, but this is dead WRONG. Radio waves are ENTIRELY straight, just like ANY OTHER EM radiation. WTF lousy idiot says this bullshite?

      Nothing defines any ‘conical shape’. The radiation is emitted from its source, whatever that happens to be. The waves (at least ‘waves’ is the right term) propagate outward from that point in a straight line. They are altered in direction by wave INTERFERENCE and REFRACTION and REFLECTION.

      This is Physics 101 stuff dude.

      The decrease in EM radiation intensity follows the Inverse Square Law which you can read about and see illustrated HERE:

      http://en.wikipedia.org/wiki/Inverse-square_law

      Your ‘Attenuation’ paragraph is useful and correct, EXCEPT: drop the stupid term ‘non standard’. This is ALL ‘standard’ EM radiation behavior.

      Your conclusion is correct. However, this remains an inexact system of prediction:

      One factor: Attenuation isn’t going to be perfectly predictable because: Metal chairs move, people density in a room changes, someone parks a metal vehicle between the source and the receiver, etc. These are unpredictable factors.

      Another factor: The effectiveness of the pickup antenna by any receiving device. It has its own attenuation limitations. This problem is NOT going to be predictable in situ. It changes with different devices and even the angle at which the receiving device is located relative to the sending device. It remains a variable.

      Conclusion: You predict the ideal situation then build in contingency. The percent of contingency to use is never exact and requires an entirely difference set of factor predictions. I like to state an ideal, a contingency and a worst case situation.

      1. You started off well but got it wrong on the application of the inverse square law. This applies to radiation from point sources. A radio antenna does not radiate equally in all directions. The fall off with intensity is usually not one over R squared and changes a lot with the angle from the axis of the antenna. I think this is what the earlier poster referred to as the radiated cone.

        1. I left room in my explanation for various point sources.

          The inverse square law remains in place. Follow the line from the source to your chosen location away. The measurement difference between source and destination will fit the inverse square law. Kind of obvious.

          I at no point talked about differences in intensity according to the type of antenna or angle from an antenna. I speak strictly about point source intensity traveling in a straight line to the destination with an inverse square intensity at the destination.

          Yes, you have to know the intensity of the radiation at the source, which I believe is your point.

  2. So will we eventually end up with an electronic surveillance system that constantly tracks everyone? Such a system would be totally benign, of course. It would only hand over tracking data to the “authorities” if they asked for it.

    1. Back during the noise about Apple iPhone’s “tracking” users we learned that all Apple was doing was transmitting from your device to them the location at which wifi signals were picked up. They weren’t correlating that to *you* — which was the focus of all the noise (other than the file on the iPhone file system that stored locations of wifi signals).

      I see the same thing happening here: Apple won’t care *who* is sending the pictures of where in the mall they are and what wifi signals are present there, they just care *what* is being shown in that picture: Store name and perhaps mall store number if visible. Whereas Google StreetView uses Google cars driving around, this new strategy could crowdsource the mall layout.

  3. There’s a ton of potential here.

    Apple has identified mapping as an area where they can leverage their end to end advantage over competitors. (IE: Apple can simultaneously employ the necessary hardware *and* software in their iPhones, Airports, Macs, AppleTVs, etc, to achieve their goals.) It is within the realm of possibility that Apple could choose to broadcast WiFi telemetry, allowing mobile devices to echo-locate the way GPS currently works. With a simple update every Apple device on earth can become a location beacon, and future Apple devices could also employ extra antennas or whatever is needed hardware-wise to improve accuracy.

    Google and MicroSoft can only dream of being able to do this. Unlike Apple, they must rely on third parties to find ways to make their software-based solutions work better.

    Sadly, any advantage won’t last, they’ll simply press to have manufacturers (like Samsung) copy Apple’s innovations, knowing that patents are notoriously hard to win and Apple can’t afford to sue everyone.

  4. And there’s this:

    Location Data Can Uniquely Identify Cellphone Users

    And this:

    Anonymized Phone Location Data Not So Anonymous, Researchers Find

    Based on hourly updates of a user’s location, tracked by pings from their mobile phone to nearby cell towers as they moved about or made and received calls and text messages, the researchers could identify the individual from just four “data points.” With just two data points, they could identify about 50 percent of users.

    Gotcha.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.