UPM showed a farm digital twin in 2024. Here is what its sensors were meant to help growers decide

A digital twin in agriculture is not a science-fiction copy of a farm. In UPM's 2024 demonstration, it meant moving field data—from sensors and crop records—into a digital system that growers could inspect on a computer or phone.[1][2]

The university presented the work during the national AI for the People launch at UPM in Serdang. Its article highlighted water use, soil quality, fertiliser absorption and tree height as examples of planting-site data that could be monitored more regularly.[1][3]

The more detailed showcase guide named the prototype MyAgriDT: Digital Twin System for Integrated Multi-Crop Intelligence. It described a research system funded through an RMC UPM Strategic Grant for 2022 to 2024—not a national farm rollout or a commercial product already used by every grower.[2]

UPM said sensors could be placed at selected positions to record conditions and help farmers monitor tree growth, soil, fertiliser and water needs. The greenhouse photograph published with the university's report shows monitoring equipment positioned among potted plants, grounding the project in physical data collection.[1]

The showcase guide added a second layer: MyAgriDT was intended to integrate several farm IoT systems with crop-growth records in a central crop data lake. That integration matters because readings from separate devices are less useful when their formats, timestamps or crop identities do not line up.[2]

The digital twin was therefore both a representation and a data pipeline. The useful chain ran from a real plant and sensor, through cleaned and linked records, to a status, trend or recommendation that a grower could understand. A missing or poorly calibrated sensor would weaken every step after it.[1][2]

UPM's article said regular monitoring could help a grower respond when a tree lacked fertiliser, needed water or showed signs of disease. The showcase described outputs that reported crop status and trends, highlighted crops at risk and recommended nutrient recovery.[1][2]

Those functions are decision support, not automatic proof that the advice is correct. A grower still needs context: crop variety, growth stage, recent weather, irrigation history, soil conditions and the confidence behind a recommendation. The system should also show which measurement triggered an alert.[2]

A useful operational test would compare the alert with what happened in the field. Did the grower act sooner? Was the diagnosis correct? Did water or fertiliser use improve without harming growth? Launch descriptions establish the intended workflow, while field trials and repeated seasons establish whether it works reliably.[1][2]

UPM's event report said 1,000 people attended the wider AI for the People launch. That number belongs to the event, not to MyAgriDT. It does not mean 1,000 farmers used the prototype, 1,000 farms were connected or 1,000 recommendations were validated.[3]

The digital twin article and showcase guide did not publish a sample size, farm count, commercial deployment figure or before-and-after result for yield, water, fertiliser, cost, disease losses or income. That does not invalidate the research; it defines what the 2024 evidence can and cannot support.[1][2]

A Bernama report four months later mentioned digital twin farming among UPM's smart-agriculture initiatives, showing that the work remained part of the university's agriculture programme. The later reference still did not convert the prototype into a proven nationwide outcome.[4]