Next Generation gaming: no more physical attachments, and no more camera range r…

By | June 6, 2013
Next Generation gaming: no more physical attachments, and no more camera range restrictions!

This is quite neat: based on WiFi Signal Analysis, Basic gestures can be recognized. I can't wait for the next Generation of gaming consoles coming out – could you imagine a Wii without having to hold a band, stick, or other control Device? Or a Kinect that works even from behind a wall? And not even optical recognition needed to Analyse the Situation, meaning potentially closer to real-time reactions and more processing Power available for physics models and other effects.

At the same time, it's also quite invasive. Triangulating signals from a router at a known position in order to analyse movements and guestures on the other side of a wall means that someone parked outside of your home – where WiFi is typically still reachable – could reconstruct exactly what you are doing. Without a video camera. That's a scary thought. But it also opens amazing possibilities if used as input for augmented reality devices, such as on-wall monitors, or perhaps also as an input source for google glass.

/via +Paul Armstrong 

8 thoughts on “Next Generation gaming: no more physical attachments, and no more camera range r…

  1. Valdis Kletnieks

    +Sophie Wrobel Good luck on getting cross-coverage in a single home from a neighbor's network.

    There's some major security issues involved – you have a system that doesn't have much actual knowledge of where the target is, and you're asking your neighbor's system for info on what you're doing, without allowing spying on what the neighbors are doing.  In addition, there's the communications issue – if you're using different WPA2 keys (for instance), how do you securely send data from the remote system to yours, without allowing you each to see other traffic, or a third party to see any traffic?  (Note that this is a generic problem in computer security – trusted communication between two endpoints in different administrative domains that have reason to be distrustful of each other is always hard).

    Oh, and there's that whole "the neighbor changed his SSID to DontStealMyWireless" – which is I guess a 7-word summary of the preceeding paragraph 🙂

    So you're probably better off just biting the bullet and using multiple sensors of your own. (Oh – and I stand corrected.  You can't do it reliably with 2 sensors, as motion perpendicular to a line drawn connecting the sensors won't generate a Doppler signal at either on.  You need 3, preferably as close to an equilateral triangle as possible.)

  2. Sophie Wrobel

    +Valdis Kletnieks part of my observation on why it would work: the technology is ubiquitous. And in particular multiple-generation buildings typically have multiple overlapping wifi-access points already – even single homes have multiple overlapping wifi access points, thanks to neighbour networks. Thus the infrastructure is already there, which is a major advantage to uptake when deploying a new technology and may make the investment in the extra effort worthwhile, in particularly for technology-adverse senior groups.

    (Thanks for the explanation of your cynicism. If you're up for it, I'd like to send you an exclusive invitation in the next month to help make stuff go wrong for a project I'm on. 🙂 )

  3. Valdis Kletnieks

    I'm not sure if it actually helps all that much for the "I've fallen" case – what this technology actually does is measure a Doppler pattern, and correlate various patterns with different gestures. This has several implications:  First, it will have a very hard time processing new and novel Doppler patterns to figure out what human geometry changes produced the pattern (which goes back to my original "will this be any less processing than current schemes").  It also means that if the motion is perpendicular to the line-of-sight to the transmitter, there's almost no signal.  Which means if somebody takes a spill perpendicular to the sensor, there's low signal, and difficult to recognize "fall".

    What might work is a pair of sensors with overlapping coverage fields, and basically set a rule that says "If this 85 year old manages to generate any movement over 3 meters/second, signal an alert".  Just make sure that nobody startles Whiskers the kitty. 😉

    (And yes, before you ask, part of my job is to sit around and think of ways that stuff can possibly go wrong 🙂

  4. Sophie Wrobel

    +John Said agree – I wonder whether smart carpets/floors or WiFi signals or something else is the best way to go with respect to non-invasive precautionary applications for the elderly. There's lots of room for improvement there!

    +Valdis Kletnieks my assumption is that due to a lower bitstream throughput rate, algorithms should be able to be optimized for better performance. But as you mention, location detection is limited, meaning that the potential applications, in particular things requiring high resolution location-specific input data, are still limited. I suppose you could develop some modulation techniques following the radar principle to overcome this… but as the video hints, it's still a research prototype and commercialization is some ways off.

  5. Valdis Kletnieks

    "And not even optical recognition needed to Analyse the Situation, meaning potentially closer to real-time reactions and more processing Power available for physics models and other effects."

    Umm.. is there any indication that the signal processing needed for this is significantly less than what traditional optical tracking methods need?  Especially considering that there's only basic gesture analysis and (as far as I can tell) no really good location detection.  Sot it can't tell if you're swinging a virtual tennis racket, or trying to wave off some smoke/steam from the pan you're cooking dinner in.

    So yes, "analyze the situation" is still needed.

  6. John Said

    if this could be used to know if someone has fallen over.  If an elderly relative fell on the floor in the bathroom, it could warn you but it doesn't invade their privacy like a camera would.

    could be very useful for retirement complexes and hospitals. 


Leave a Reply

Your email address will not be published.