ECG Acquisition in a Computer - ECG Acquisition and Automated Remote Processing - page 85

Fig. 4.12 Test results for ECG encoding and detection of header using wireless communication.
a Encoded packet structure, b Decoded packet at the receiving end

4.3.6 Error Correction

Other than the common reasons like multipath propagation, ISI etc., the DC-like
nature of the equipotential segments (ST, PQ. and TP) of the ECG signal causes an
inherent source of error. This inherent error is beyond the scope of error detection
schemes since the transmitted data between these segments are not the exact
replica of corresponding portions. Hence, at the receiver end, a distorted ECG will
be reproduced.
To avoid this, bi-phase modulation scheme was adopted as described in
Sect. 4.3.1 . However, data can be corrupted during transmission due to many
factors for which one or more bits of a given data units are altered or wiped out.
The data link layer of this transmission system was so designed that it is capable to
detect any such alteration of one or more bits using the 'checksum' error detection
method. The packaging of ECG data, flow control of this packet to the phone set
using the available band limited bps selection in the transmitting end belongs to
the data link layer. In the receiving end, an addition of decoded data frames from
the received bit stream with this flow rate and error correction principle was also
adopted. The decoded packets were transferred to the receiving-end desktop PC
through serial communication link using RS-232 protocol. The incoming packets
are buffered in a temporary file and the BER and PE are computed by the appli-
cation software using the following formulae:

Bit Error Rate (BER) ¼ Number of missed bits (or faulty packets)
Total number of transmitted bits

100

Number of lost packets
Total number of packets received 100

Packet Error (PE) ¼ ð 4 : 1 Þ