Page 14 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

3 Case Studies

As part of our studies on use cases, and our detailed discussions with the use case authors,
we have selected certain case studies which bring out the benefits (or lack of it) for increasing/
measuring AI readiness. Especially we focus on those case studies that utilize the readiness
factors mentioned in Section 1 above. In addition, we look for clear metadata, supporting
references, and published research papers, with experimentation that can practically showcase
the benefits of AI readiness on these terms.

Each case study is mapped to the 6 readiness factors listed in clause 2 above and the instances
of the readiness factors are explained for each case study.

3.1 Case Study-1: IoT-based Environment Monitoring Based on
Standard Indices

This case study involves a set of use cases which monitor environment parameters such as soil
sensor, piezometers, and water level sensors etc. and infer standardized indices for specific
use cases e.g. groundwater level (GWL) mapped to drought codes (DC). The area of coverage
may be quite large, for example, multiple hectors of forest land. Verification of sensed data
and inferred data with ground truth in collaboration with experts is an essential characteristic
of such use cases. Communication networks, including data format conversions are important
standard requirements for such use cases.

Net-Peat-Zero [48]: Networked Association of Southeast Asian Nations (ASEAN) Peatland Forest
for Net-Zero delivered by University Putra Malaysia is an excellent example of a use case with
real-world deployment and its application of open data, which is accessible to everyone.
This use case presents the possibility to leverage AI in predicting Forest Fire in peatland areas
in South Asia. An improved tropical peatland fire weather index (FWI) system is proposed, by
combining the groundwater level (GWL) with the drought code (DC). To monitor the peatland,
a LoRa-based IoT system is used, and sensors such as soil sensors, piezometer sensors, water
level sensors, and weather sensors are used, with the expectation that integral meteorological
information could be detected. All the data mentioned above could be cross-checked with the
ones used by the Malaysian Meteorological Department (METMalaysia), which means that the
data collected by the IoT system is authentic and ready to be processed.

In addition, an improved model to apply the GWL is proposed for the FWI formulation in the Fire
Danger Rating System (FDRS). Specifically, DC is formulated using GWL, instead of temperature
and rain in the existing model. From the GWL aggregated from the IoT system, the parameter
is predicted using machine learning based on a neural network. The results show that the DC
calculated from the IoT system has a high correlation with the data released by METMalaysia.
This shows that DC can be calculated using predicted GWL.

The solution has been deployed in Raja Musa Forest Reserve (RMFR) in Kuala Selangor, Malaysia.
Deployment in a natural environment proves the effectiveness and efficiency of the whole
system. It is desirable to extend the project by measuring carbon emissions from the peatland
forest and how net zero can be achieved by managing the peatland better using IoT technology
and community-based management.

7