Title: Tripod Localization and Sensing Testbed
1Tripod Localization and Sensing Testbed
2Goals of this effort
- Develop a system to calibrate locations of
field-deployed motes - Turnkey system
- Deploy, use, and take down in a day
- Minimize restrictions on deployment
- Provide a convenient platform for
- testing and evaluating new ranging techniques
- Acoustic channel measurement?
- Other kinds of sensing experiments?
3Mote Localization Problem
Motes
Tripods
4Tripod Design
- Resource un-constrained design
- Runs on batteries for 6 hours
- Four audio channels
- 4 fixed microphones
- Omni-directional, 15Hz-30KHz
- 1 fixed speaker with a reflector
- 96 KHz / 24 bit samples
- CPU
- 1 GHz P3, 256 MB, 20GB Hard drive
- Runs on 12V
- Network
- 802.11 for inter-tripod network
- MoteNIC to talk to mote network
- 3-axis orientation sensor
5More detail
USB 802.11 adaptor
Mica2 Mote
Reflector
Car stereo speaker
3-axis orientation sensor from MicroStrain. May
be dropped due to cost if found unnecessary.
Behringer ECM8000 microphone, intended for
acoustic testing. Very flat response from 15Hz
to 20KHz
2x Audiobuddy 2 channel microphone preamp,
feeding to M-Audio Delta-44 4 channel 96/24 PCI
sound card.
30 cm inter-mic spacing
CPU Currently a ruggedized industrial PC from
littlePC.com
6Questions
- Would this kind of platform be helpful to your
group? - Are there additional / unnecessary features?
- Is 96 KHz / 24 bit useful to you?
- Microphone spacing?
- Microphone preamps
- Need phantom power
- The one I selected runs off 9V AC ?
- Rather have it run off a (noisy) DC supply
7Mote Localization in More Detail
Relative locations of motes more poorly
constrained
- Inter-mote ranging.. Why not?
- Additional hardware and software constraints
- Resource consumption
- Emitter only is low-cost
- How important is it?
- Bi-partite geometry may leave locations of motes
poorly determined - Other constraints
- Bearing estimates?
- Mote connectivity?
8Tripod System in more Detail
After motes in blue region are located, move
tripods to cover green region
- Grounding the coordinate system
- Locate one tripod at origin position, define it
as the origin - Will another fixed point be needed?
- System feedback to user
- System can ask tripods to be moved to improve
constraints - System indicates when all motes in region are
located - Cover more ground by moving tripods
Origin
9Experimental Ranging TechniquesLeveraging Four
Channels
- Bearing estimates
- Add constraints to multilateration
- May be helpful to address bi-partite topology
- Angular inconsistencies across nodes might detect
obstructed signals - Computed based on detected ranges to each
microphone - More sophisticated analysis might correlate mic
signals
10Multipath Bearing Estimation
- Bearing estimates of multipath components?
- Distribution of incoming paths (angle and
intensity) - Might provide clue about whether path is
obstructed - Cases with LOS should have much stronger direct
path - Deduce location of sender in some cases?
- Estimation using model of reflectors