Be first to read the latest tech news, Industry Leader's Insights, and CIO interviews of medium and large enterprises exclusively from CIO Advisor APAC
Internet of Things (IoT) has made an immense impact on various sectors like agriculture, manufacturing, defense, among the many and is sure to pave a path for more advancement in the future. Some critical aspects Of IoT which are still to be observed are:
Sensors might not be accurate every time
Everything recorded by IoT devices cannot be correct. For example, the temperature sensors might drift in a desert range.
In the oil and gas industry, the temperature sensors might drift due to the range of desert temperature. The cost of sensors meager and replacement is also easy. To mitigate this drifting effect, information from multiple sensors must be considered.
Even when sensors are providing accurate data, there might be various aspects to be considered, which is why there is a need to second subject consideration.
Sensors are designed to support remote monitoring and control of the whole device, not to store-and-forward high-frequency data from each sensor. Control units filter and summarize data collected by the sensor. But, the information filtered by a control unit might be crucial for another. So, the premature summarizing of data should be avoided.
Scoring
Model-building and model-scoring to produce predictions in IoT applications are separate processes. Optimizing end-to-end systems, requiring advanced computation while building the model and tuning it necessitates access to historical data from many devices is needed.
Sensor data can be useless
Sensor data is not useful as a basis for action; it requires integration with other data from the organization to become useful.
For example, if an oil pressure sensor on a train temporarily exceeds a threshold. The Comparison with previous data that preceded failure in the past, using historical and operations data is needed to determine whether the information from sensors is authentic.
And if the gearbox is indicated as failing, decisions have to be taken about how, where and when to conduct the repair, using network operations data, parts inventory data, HR records for qualified engineer-availability, and so on.
Failing to combine sensor data with organizations other data fatally undermines their entire project.