Zoom out Search Issue

ManualsBrandsContents Manualsaudio & home theatreZoom in

IEEE SIGNAL PROCESSING MAGAZINE [13] MARCH 2015

Signal processing is also essential for

creating receivers that can cope with an

onslaught of data coming in from large

numbers of ultrasonic sensors floating

inside a body. “Basically, solving some

mathematical optimization problems

gives us the best way to share the chan-

nel between different devices that are try-

ing to transmit at the same time,”

Melodia explains.

Early in their investigation, the re-

searchers proposed basing their ultra-

sonic intrabody network on ultrasonic

wideband (UsWB), a relatively new mul-

tipath-resilient physical and medium ac-

cess control (MAC) layer integrated

protocol. UsWB is the only MAC proto-

col specifically designed for ultrasonic

intrabody sensor networks, while a wide

variety of MAC protocols designed for

traditional RF-based wireless networks

are currently available.

According to the researchers, UsWB is

based on the concept of transmitting short

carrierless ultrasonic pulses following a

pseudorandom adaptive time-hopping pat-

tern, featuring a superimposed adaptive

spreading code. After testing the protocol,

the researchers were able to show that

UsWB enables nodes to flexibly trade data

rate performance for power consumption

while allowing multiple concurrent sen-

sors to coexist by dynamically adapting

their transmission rate to channel and in-

terference conditions.

Recently, the researchers compared

the performance of UsWB with a pair of

existing MAC protocols originally designed

for use with RF-based wireless networks:

ALOHA and carrier sense multiple access

(CSMA). Their tests showed that UsWB

generally outperforms ALOHA in terms of

throughput, although CSMA can achieve

comparable performance under certain

kinds of setups. Additionally, according to

the researchers, ALOHA and CSMA both

exhibit very high packet drop rates com-

pared to UsWB, which always keeps the

packet drop rate below a given threshold.

The tests also found that UsWB performs

better than either ALOHA or CSMA in

terms of short-term fairness, average

packet delay, and delay variation. Finally,

the researchers discovered that CSMA has

the highest energy consumption per bit,

due to long idle listening times. UsWB’s

bit cost, on the other hand, is the lowest

and can be further reduced by trading

throughput for energy consumption

through energy-minimizing rate adapta-

tion, the researchers say.

Testing various intrabody network

system technologies and configurations

required creating an environment that

mimicked real-world conditions. “We

used some devices known as medical

phantoms,” Melodia says. For this par-

ticular project, the phantoms were syn-

thetic devices designed to produce the

basic ultrasound characteristics of real

tissue. “We’ve used mostly kidneys so

far,” Melodia says. “We have a synthetic

kidney that we transmit information

through” (Figure 1). The next step is

building a miniaturized prototype of the

transceiver. “We have a prototype that

works well, but it’s big,” Melodia says.

“We want to build a miniaturized plat-

form that will be able to do what we’re

doing now, but be much smaller and can

be implantable.

Melodia predicts that an intrabody

network system could become available

for general use within a decade. “It will

take a lot of work, but that seems like a

realistic possibility,” he says.

NEURAL RECORDING SENSORS

A research group in the University of

Bath’s Centre for Advanced Sensor

Technologies (CAST) is investigating the

use of implantable devices for electro-

neurogram (ENG) signal recording, po-

tentially increasing the quantity and

quality of received information. An ENG

is used to visualize directly recorded

electrical activity of neurons in the cen-

tral nervous system (brain and spinal

cord) or the peripheral nervous system.

Reliably collecting neural data is a goal

that has eluded numerous researchers for

many years. “Nerves tend to come in bun-

dles of hundreds or thousands and carry

neural traffic to different destinations to

and from the central nervous system,”

says John Taylor, CAST’s head and a pro-

fessor in the University of Bath’s Depart-

ment of Electronic and Electrical

Engineering. Identifying individual path-

ways and the traffic on them is difficult.

“Several years ago we invented a tech-

nique called velocity selective recording

(VSR) that, we believe, goes some way to

solving this problem,” Taylor continues.

Working with project collaborators, in-

cluding University College London, the

University of Cambridge, the University of

Freiburg, and Aalborg University, the Bath

researchers developed a range of implant-

able electrodes and amplifiers to test a

technique that Taylor says is essentially a

simple signal processing concept.

“Our recording technique provides re-

al-time velocity spectral analysis of ac-

tivity on a nerve,” Taylor says. In practice,

[FIG1] A synthetic kidney was used in preliminary tests at Northeastern University

to evaluate various ultrasonic communication technologies and configurations.

(Photo credit: Northeastern University.)

THE WORLD’S NEWSSTAND

THE WORLD’S NEWSSTAND