nificant power requirements, current harvesting tech-
nologies can only aid in relaxing a power require-
ments of the primary powering source rather than of-
fering an autonomous solution (J. Olivo and Micheli,
2011). Inductive power links can be used, and most of
the existing implant devices use inductive links with
frequencies on the order of a few MHz or less, primar-
ily due to the low absorption rates in tissues at these
frequencies. The primary limitations of these conven-
tional wireless powering technologies, however, are
low efficiency and large form factor. Use of high-
frequency inductive links allows a 100-fold reduction
in the size of the implanted coil (Poon, 2009). How-
ever, high-frequency inductive links also suffer from
low efficiency, in addition to high transmit power
(A. S. Y. Poon and Meng, 2007). A critical issue in
high-frequency inductive coils is the amount of power
received at the implanted secondary coil. It has been
shown that power on the order of a couple hundred
µW can be provided ((Poon, 2009), (A. Yakovlev and
Poon, 2012)), which can possibly be extended to sub-
mW or mW ranges. Such power levels are safe for op-
eration in tissue and and obey specific absorption rate
(SAR) requirements. To power an implanted device
with state-of-the-art high-frequency inductive links it
is imperative to optimize the link. Designing an im-
plantable wireless system within the limits of current
technology therefore mandates a careful review of a
number of factors: link power requirements, modula-
tion, system and circuit design, and system operation.
4 PROPOSED METHODOLOGY
Transmission scenarios primarily dictate the topol-
ogy of the complete link and the architectural design
choices of required electronics and thus the design
methodology. BAN applications target two primary
transmission scenarios: (i) body surface, extending
up to 2 cm above the skin, and (ii) external- up to
5 m (Chavez-Santiago et al., 2013). An evaluation
of approximate link budgets for these topologies re-
veals a need for a two-pronged approach, wherein the
first system would use magnetically coupled antennas
while the second uses RF-based transmission. A key
design objective is thus to enable reliable transmis-
sion with minimum energy requirements.
Use of spread spectrum signaling can not only
aid to enhancing the reliability of the proposed im-
plantable telemetry but can also help in supporting
communication between multiple implantable nodes.
Also, since transmissions in these applications are in-
termittent and erratic, a preamble-based detection sys-
tem suits the application scenario. A low-complexity
transmission architecture can further be supported
by using the proposed asynchronous packet-based
transmission system based on direct-sequence spread-
spectrum modulation (S. Nagaraj and Burnashev,
2009) due to its inherent ability to reliably support
uncoordinated data transmission with low power con-
sumption. The basic modulation scheme is differ-
ential binary phase shift keying (DBPSK) in addi-
tion to spread spectrum. While DBPSK allows for a
simple modulation with superior bit-error-rate (BER)
performance as compared to low-complexity modu-
lation schemes typically used in biomedical applica-
tions (such as on-off keying, frequency shift keying
etc.), the use of spread spectrum signaling offers sys-
tem reliability due to inherent robustness to interfer-
ence and multipath propagation, in addition to en-
abling scalability to support multi-implant communi-
cations. The proposed scheme embeds the transmit
data into a packet structure marked with a preamble
sequence, which when detected by the receiver can
be used to extract timing information for synchro-
nization. This allows for reduced power consumption
since it alleviates the need for continuously operat-
ing a synchronization stage, otherwise typically used
in conventional spread spectrum systems. Addition-
ally, it has been shown in (S. Nagaraj and Burnashev,
2009) that due to differential encoding, the proposed
preamble detection method does not require phase
and frequency offset estimates for channel coherence
time larger than the block duration, thereby, contribut-
ing to a significant reduction in system complexity.
Also, without the need for exact timing information,
a detector operating with two samples per chip essen-
tially performs close to that of a synchronous system
and is robust in frequency-selective channels.
The design methodology is dictated primarily by
the application scenario or link topology. In the case
of body surface transmission, data transfer can typ-
ically be enabled without the need for high transmit
power levels since the amount of loss encountered in
the channel is low for small transmission distances. A
first estimate of link losses at an implantation depth of
2 cm of tissue reveals that losses are about 20-30 dB
less in the 2.4 GHz ISM transmission band as com-
pared to the conventional understanding (N. R. Sar-
choghaei and Schlegel, 2013). It is observed that
to be able to support acceptable transmission using
the proposed communication scheme, the energy re-
quired to transmit every data bit over 2 cm of tissue
is about 64 fJ. Table-1 lists first hand worst-case es-
timates of energy requirements (using upper-bounds
on loss margins) for data transmission through a pro-
posed link for body surface transmission at an im-
plantation depth of 2 cm at low kbps data rates. With
DesigningNext-generationImplantableWirelessTelemetry
273