Database Reference
In-Depth Information
Table 1.
Sample Query Set
Queries
1. Find total amount of time spent above 250W (Power-measuring)
2. Find Heart Rate for each occurrence above 250W (Power-measuring)
3. Find total amount of time spent above 165BPM (Heart rate-measuring)
4. Calculate average heart rate spent above 200W
5. Find the total amount of time where pedal pivot = 'pivot range 1'
6. Find 'best intervals' for highest '1minute' heart rate and return values
for distance covered
7. Find the average performance factors (Power/Heart Rate/Speed) for each
gradient of type='hill'
8. Find the average Power value when pedal vector magnitude ='peak'
9. Find the average speed when braking activity = 'none'
10. Find all occurences where gradient profile = 'flat' and cycle
cadence ='cadence range 1'
Query Processing
Component
Query/Event
API
Q
Sensor
Data
Integrated
Data
Integrate
rate
Cont
Reposi
Context
Repository
Contextual
Enrichment
Ct
C
En
Enablement
Data Management
Component
MetadataComponent
Hardware Component
Fig. 1.
EventSense System Architecture
involved to close the gap between the requirements of the domain expert and
the initial format of the sensors output. These processors are discussed briefly
and in the following section, we discuss Contextual Enrichment in greater detail
- as this is the main focus of the paper.
Sensor Enablement.
Sensors output data in a simple plaintext format. The
common approach to analysing data requires considerable manual effort using
spreadsheet tools to compare results across multiple sensor files. The common
approach to automation is to build proprietary wrappers for every sensor to
store data in a database format. This requires new wrappers for new sensors
and database reengineering each time the structure of the output differs. Sensor
Enablement is a form of structural enrichment whereby we convert all sensor data
to a standard XML format using simple user defined templates. One can define
how sensor output should be interpreted and structured using sensor profiles,
