160 likes | 174 Views
This document explores possible uses cases for multimedia applications involving motion capture and gesture analysis with a BAN network, and presents the related technical requirements.
E N D
Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Possible BAN uses case for multimedia application] Date Submitted: [March 15, 2008] Source: [Jean Schwoerer] Company [France Telecom R&D] Voice: [+33 4 76 76 44 83] E-Mail: [jean.schwoerer@orange-ftgroup.com] Re: [] Abstract: [Showing possible uses case implying motion capture or gesture analysis with a BAN network and present the related technical requirements] Purpose: [To consider the requirement of thoses application] Notice: This document has been prepared to assist the IEEE P802.15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P802.15.
Gesture analysis and motion capture : related requirement for BAN networks Jean Schwoerer France Telecom R&D
Purpose and Contents • Motion capture and gesture analysis are possible BAN Applications • The purpose of this presentation is to discuss uses case and corresponding requirements for BAN networks to permit those applications. • Contents • Motion capture and gesture analysis • Uses cases • Concerned BAN devices • Technical requirements • Conclusion
Gesture Analysis : it consist of.. Capture relative information on a motion Analyze those information to identify either the gesture or the person who do the gesture (pattern matching) The goals is not to get the representation of a motion but just to identify a gesture (by ex : "hand was raised") Gesture Analysis : How to do it ? Most of the time it's done through 2D or 3D accelerometer It doesn't require a fixed spatial reference Gesture identification is done through contextual filtering and patter matching Gesture Analysis & Motion capture
Motion Capture : it consist of Capture the trajectory (2D or 3D) of one or several point on the body Uses those information to re-construct the motion. This requires more complete information than gesture analysis Motion capture give full knowledge of the motion (direction, amplitude, speed, duration..) Motion Capture : How to do it ? It do require a fixed spatial reference Accelerometer is not sufficient: magnetometer is also required Video analysis techniques also exist (more accurate) In the future, low consumption gyroscope will be helpful Some sensors integrated into clothes can also be useful Gesture Analysis & Motion capture
Gesture Analysis application • Interface navigation and browsing • Finger or mouse gestures are already used in some web browser & handheld devices • Limited handwriting recognition • Some gaming applications • Medical, Healthcare & biometric • Some gesture can be used as a kind of signature. • Activity monitoring • Sports training and monitoring (podometer for ex.) • Fall detector
Hardware • Existing hardware..
Hardware • In a BAN.. • Video glasses and headset • Dataglove for hand and finger motion • Motion sensor in watch or shoes • Optical or resistive bend-sensor can also be included in clothes.
Motion capture applications • Sports and bio-dynamic applications • Studies of technical gestures • Training and equipment improvements • Medical applications • Re-habilitation • New interface for communication • Virtual reality (for gaming or communication) • Remote manipulations
Latency : delay between user's action and system reaction Asynchronous interaction (user interact in open loop, like gesture navigation): 100ms max Synchronous interaction : 20 ms max (user interact in closed loop, like moving a pointer on a screen) Data rate : depend on accuracy : usually 8 bits/axis/measured data sampling rate : depend on how fast the body is moving Network requirements
Gesture navigation Gesture identification is enough for this 2D or 3D Accelerometer : 2 or 3 axis (8 bits/axis) Sampling rate : 100 Hz Usually used for asynchronous interaction : max latency : 100 ms Data rate : between 1,6 to 2,4 kbit/s Network requirements
Handwriting recognition : Fingers are one of the fastest parts of the body Motion bandwidth : at least 30 Hz. Required sampling rate : at least 100 Hz. Required data : XY coordinates (2D Motion capture) at fingertips + Optional pressure sensor Hand motion : Basic dataglove include at least 5 sensors More complete models can include up to 22 sensors Sampling rate : 100 Hz Data rate from 4 to 17,6 Kbit/s Network requirements
Arm or Leg motion capture (arm) Require to compute 3D positions of several points on the body This require 3D accelerometers (8 bits / axis) : provide the dynamics of the motion This also require 3D magnetometers (8 bits / axis) provide a reference to a fixed reference. Required amount of data : 48 bits per measurement Arm motion capture require 3 to 5 measurement points, each at 60 Hz (at least..100 Hz is more comfortable) Datarate from 8,6 Kbit/s to 24 Kbit/s Max latency : 20ms (motion capture is usually synchronous) Network requirements
Technical requirements • Network mechanisms • We need a very flexible network with automatic subscription / unsubscription mechanism : you don't want to reconfigure your network each time you change your shoes. • Physical Layer • Maximal ranges is about 2 m • Need high flexibility in • Data rates : (from a few kbit/s for sensor to 10 Mbit/s for video) • Energy consumption : is a constraint for sensor integration • Every devices (LDR, HDR and implant) need to be able to communicate on a unique network with a kind of "universal" air interface.
Conclusion • Motion / Gesture capture is a key feature for various BAN applications • Key requirements are • Ability to support important number of nodes • With guaranteed latency for each of those nodes • A BAN include • A lot of sensors (medical, motion, environmental) • Personal objects and interface (watches, video glasses) • or PDA A cell phone is the BAN gateway to the outer world • Scalability, in data rate as well as consumption if a mandatory feature