IoT in Provenance Management of Medical Data
355
Nifti, Minc, DICOM. Addition of semantic data for such images is simply a linking
between file itself and corresponding information from the ontologies and rule base.
3.3 Analysis of the Proposed Solution for Medical Data
Provenance
First, let us state that on the lowest level the source of the erroneous data may be
identified as
1. IoT-device.
2. The patient whose actions has direct effect on the received measurements. Execution that is not compliant with device instructions is an often reason for this
event.
To extend the nomenclature of the data provenance, we consider IoT-device.
Despite being a relatively simple computational device, it combines multiple technologies that should be taken into account. The following list includes general
explanations of errors on the device level:
1. Sensor level (sensing element is not capturing data properly on the declared
sensitivity range with necessary precision or other sources of erroneous data).
2. Hardware level (processing device cannot capture data and transfer them to the
destination).
3. Power supply level (connected with two previous levels as can cause errors for
both of them).
4. Software level (includes various aspects, e.g., protocol failure, software inconsistency, failures due to the specific software environment states, etc.).
5. Communication level (appears during transmission phase and partially dependent
on the selected communication protocol).
This list is sufficient to provide control over most use-cases under consideration.
We would like to devote particular attention to the sensor level malfunction. To
denote sensors state we propose the following classification:
1. Sensor is in normal state under normal circumstances.
2. Sensor acquires data that is biased from the previous history and the patient is
aware that conditions are normal.
3. Sensor is capturing data with output values located out of the limit on the
sensitivity range (applied impact is not suitable for current sensor).
4. Sensor acquires a priori incorrect data under known conditions.
5. Sensor cannot capture a sample.
The second case attracts the most attention, as it is common for the most cases
and relates to both correct and incorrect data. Slight bias from the measured value
might indicate further trends in the patient’s physical state. Hence, this information is helpful for prevention of possible negative impacts. At the same time, small
355
Nifti, Minc, DICOM. Addition of semantic data for such images is simply a linking
between file itself and corresponding information from the ontologies and rule base.
3.3 Analysis of the Proposed Solution for Medical Data
Provenance
First, let us state that on the lowest level the source of the erroneous data may be
identified as
1. IoT-device.
2. The patient whose actions has direct effect on the received measurements. Execution that is not compliant with device instructions is an often reason for this
event.
To extend the nomenclature of the data provenance, we consider IoT-device.
Despite being a relatively simple computational device, it combines multiple technologies that should be taken into account. The following list includes general
explanations of errors on the device level:
1. Sensor level (sensing element is not capturing data properly on the declared
sensitivity range with necessary precision or other sources of erroneous data).
2. Hardware level (processing device cannot capture data and transfer them to the
destination).
3. Power supply level (connected with two previous levels as can cause errors for
both of them).
4. Software level (includes various aspects, e.g., protocol failure, software inconsistency, failures due to the specific software environment states, etc.).
5. Communication level (appears during transmission phase and partially dependent
on the selected communication protocol).
This list is sufficient to provide control over most use-cases under consideration.
We would like to devote particular attention to the sensor level malfunction. To
denote sensors state we propose the following classification:
1. Sensor is in normal state under normal circumstances.
2. Sensor acquires data that is biased from the previous history and the patient is
aware that conditions are normal.
3. Sensor is capturing data with output values located out of the limit on the
sensitivity range (applied impact is not suitable for current sensor).
4. Sensor acquires a priori incorrect data under known conditions.
5. Sensor cannot capture a sample.
The second case attracts the most attention, as it is common for the most cases
and relates to both correct and incorrect data. Slight bias from the measured value
might indicate further trends in the patient’s physical state. Hence, this information is helpful for prevention of possible negative impacts. At the same time, small