Detection of Dangerous Situations Using a Smart Internet of Things System - New Contributions in Information Systems and Technologies - page 370

data of this new sample has anomalous values, then the algorithm will verify if these
values are used to occur. If they aren't, it is necessary to see if other sensors, which
are used to come out of the expected range are also anomalous. If the recent sample
has anomalous values in the normally affected data and one more symptom, then this
is an urgent situation and the weight given to this biometric element, that usually isn't
affected, must be higher.
Finally these new weights attributed to each sensor will be used to calculate the
average and classify the situation as fear or illness.

3.4 Algorithm 1

This algorithm attempts to identify situations where an anomalous value, which isn't
used to occur in a certain biometric record, indicates a more dangerous situation. At
some cases of illness the emergence of a certain symptom may have a higher relev-
ance. The input parameters of this function are the arrays tts1 until ttsn, which record
the thresholds of every sensors according to the time of the day, sensor1 until sensor
n, the new sample collected and the array array_heartbeat until array_sensorn, which
have all the records of each sensor.

array sensor1_consult(){
integer array_heartbeat[n]= SELECT heartbeat FROM records.user;
// where n is the total number of records
return array_heartbeat;}
(...)
array sensorn_consult(){
integer array_sensorn[n]= SELECT sensorn FROM records.user;
return array_sensorn;}

integer k=0;
boolean weight_adjustments (integer array_heartbeat, ..., integer ar-
ray_sensorn, integer array_tts1, ..., integer array_ttsn, integer sen-
sor1, ..., integer sensorn){

IF(heartbeat> array_tts1[i]) THEN
k++;
array_heartbeat.add[heartbeat];
FSE

IF(k<m)THEN // m is the maximum number of times that the
// anomaly can occur and be considered unusual
IF(sensor2> array_tts2[i])THEN
integer anomaly= 2;
FSE
(...)
ELSE

IF(sensorn> array_ttsn[i]) THEN

anomaly++;

FSE

IF(anomaly>=3)THEN

boolean flag=1;

return flag;

algorithm2();

FSE