Aroma Intelligence Systems (AIS)

Aroma Intelligence Systems (AIS) originates from more than three decades of professional engagement with olfaction, sensory science, botanical natural products and ecological practice, combined with over a decade of qualitative field observations at EcoPlanetFarm.

Established in 2012, EcoPlanetFarm has evolved from an aromatherapy garden into a regeneratively managed agricultural landscape and living ecological observatory. Its mango orchards, perennial vegetation, native flora, soil communities and associated biodiversity provide a field setting for observing relationships among plant development, climate, biological interactions and seasonal ecological processes.

Repeated observations suggested variations in the perceived intensity, character and persistence of plant aromas, particularly during mango flowering, seasonal transitions, elevated temperatures and periods of changing orchard condition.

These qualitative observations form the experiential origin of AIS. They have not yet been verified through instrumental chemical measurements or controlled experiments.

Mango development and productivity reflect interacting physiological, environmental and ecological processes, including temperature, water availability, nutrient status, developmental history, pests, pathogens and reproductive physiology.

A severe crop-loss episode occurring alongside adverse climatic conditions and pest pressure raised a central scientific question:

Can changes in plant-emitted volatile compounds reveal measurable physiological responses before substantial visible damage or reproductive failure occurs?

Conventional orchard monitoring combines visual assessment, meteorological observations, pest surveillance and selected physiological or agronomic measurements. AIS proposes investigating plant volatile emissions as an additional measurement domain that may complement these approaches.

The scientific task is to determine whether selected chemical signals contain reproducible biological information and provide useful information beyond established monitoring methods.

Physiological transitions and biotic or abiotic stress may alter plant volatile metabolism, producing measurable changes in VOC composition, abundance and temporal dynamics. When integrated with independently assessed phenological, physiological and environmental data, selected compounds or multivariate volatile signatures may provide reproducible indicators of plant condition and potentially support earlier detection of defined stress events.

This hypothesis remains untested within AIS.

Plants produce and release chemically diverse VOCs, including:

• Terpenoids and their derivatives.
• Green-leaf volatiles.
• Benzenoids and phenylpropanoids.
• Other low-molecular-weight volatile metabolites.

These compounds may arise from constitutive metabolism, developmental processes or responses to environmental stress, herbivory, pathogens and plant–microorganism interactions.

Measured VOC concentrations also depend on temperature-dependent volatilisation, humidity, atmospheric dispersion, chemical transformation, surrounding vegetation, soil emissions and sampling methodology. A difference in volatile composition therefore does not, by itself, establish physiological stress or disease.

AIS approaches plant-emitted VOCs as dynamic chemical phenotypes shaped by metabolism, environmental conditions and ecological interactions.

The central experimental challenge is to distinguish biologically meaningful changes from normal phenological variation, environmental interference and analytical artefacts.

AIS proposes a research architecture connecting biological observation, chemical measurement, computational interpretation and practical field guidance.

Establish independently characterised biological and environmental reference conditions through structured observations of phenology, physiological condition, pest incidence, reproductive development, soil conditions and local microclimate.

This layer provides the biological context needed to interpret chemical measurements.

Characterise plant volatile emissions using standardised headspace sampling and gas chromatography–mass spectrometry (GC-MS).

Develop and evaluate an electronic-nose system for measuring patterns in complex volatile mixtures. GC-MS will provide molecular-level analytical information, while electronic-nose arrays will be investigated as complementary, potentially lower-cost sensing tools.

Integrate chemical profiles, electronic-nose responses, biological observations and environmental data using chemometrics, statistical modelling and AI/ML.

Investigate associations, temporal patterns, anomalies and potential predictive relationships. Evaluate models against independent observations and conventional monitoring methods.

Translate independently validated findings into practical suggestions for orchard monitoring.

Depending on the evidence, future guidance could include inspecting particular trees, checking water status, conducting targeted pest or disease assessment, or obtaining confirmatory measurements when a validated signal departs from its established baseline.

Recommendations must be proportionate to the signal's specificity and uncertainty. A chemical-pattern alert would require appropriate agronomic assessment before any intervention is proposed.

This layer is a prospective application of validated findings. AIS does not currently provide operational alerts or treatment recommendations.

Determine whether independently characterised developmental, physiological or stress conditions correspond to measurable differences in plant volatile profiles.

Assess whether observed differences can be consistently measured through standardised sampling, GC-MS and electronic-nose sensing.

Investigate whether statistical and AI/ML methods identify reproducible, biologically interpretable patterns while accounting for environmental variability and uncertainty.

Determine whether validated volatile measurements provide useful information beyond conventional monitoring, including any measurable early-warning advantage.

1. How do mango VOC profiles change across flowering, fruit development and vegetative growth?
2. Which compounds or multivariate patterns, if any, are associated with independently characterised physiological conditions?
3. Can stress-associated responses be distinguished from normal phenological variability?
4. How do temperature, humidity, soil moisture and background emissions influence measured profiles?
5. Are candidate signatures reproducible across trees, cultivars, sampling periods and seasons?
6. Can electronic-nose arrays consistently discriminate biologically relevant patterns supported by reference measurements?
7. Do computational models improve on conventional environmental and physiological indicators?
8. Do selected volatile patterns precede independently confirmed adverse outcomes with useful lead time and acceptable false-alarm rates?

AIS proposes an experimental framework combining biological phenotyping, non-invasive VOC sampling, analytical chemistry, electronic sensing and computational modelling.

Selected mango trees will be characterised by:

• Cultivar, tree identity and developmental history.
• Flowering, fruit development and vegetative phenology.
• Visible plant condition and pest incidence.
• Temperature, relative humidity and soil moisture.
• Environmental exposure and orchard-management history.
• Physiological indicators such as leaf temperature, water status and chlorophyll-related measurements, where feasible.

Repeated observations will establish biological and environmental context for chemical sampling. Visual categories will require appropriate independent evidence before they are treated as confirmed physiological or pathological diagnoses.

Volatile emissions from intact plant tissues will be investigated using suitable headspace methods. These may include dynamic headspace sampling, controlled airflow, sorbent trapping and subsequent thermal desorption, according to the experimental design and available analytical facilities.

Sampling protocols will standardise or document:

• Plant tissue and developmental stage.
• Sampling duration and airflow.
• Chamber characteristics.
• Temperature and humidity.
• Time of day and sampling sequence.
• Biological and technical replication.
• Ambient-air references and procedural blanks.

Chamber-induced microclimatic changes, contamination and handling artefacts will be evaluated. Sampling is intended to avoid destructive tissue collection, while recognising that enclosure and handling can themselves influence emissions.

The objective is reproducible measurement of plant-associated volatiles under documented conditions.

GC-MS will serve as the principal analytical reference for characterising volatile composition and investigating relative abundance and temporal variation.

The proposed workflow includes:

• Chromatographic separation and mass-spectral acquisition.
• Spectral-library comparison and retention-index support.
• Authentic reference standards where available.
• Calibration and analytical quality control.
• Assessment of blanks, contamination and instrumental variability.
• Chemometric comparison of volatile profiles.

Tentative compound annotations will be distinguished from identifications confirmed with appropriate standards. Quantitative claims will require suitable calibration and validation.

AIS proposes developing a modular electronic-nose system for plant-associated volatile mixtures.

Its initial experimental architecture may include:

1. Headspace interface: A sampling chamber with controlled airflow, ambient-air references and filtration appropriate to the sampling design.
2. Chemical sensor array: Complementary sensors, initially considering metal-oxide semiconductor technologies, selected and refined following preliminary chemical characterisation.
3. Environmental measurements: Temperature, humidity and airflow monitoring.
4. Data acquisition: Time-stamped sensor recording, baseline correction and response normalisation.
5. Computational interface: Integration of sensor signals with biological and environmental observations.
6. Analytical validation: Comparison with GC-MS measurements, controlled reference mixtures and independently characterised plant conditions.

An electronic nose generally measures collective sensor responses rather than identifying individual molecules. Its value will depend on response stability, discrimination capability and calibration.

Evaluation will address sensor drift, humidity sensitivity, repeatability, environmental interference and biologically relevant discrimination. Development will progress from controlled reference mixtures to plant-headspace experiments and then orchard trials.

The proposed research dataset will combine:

• GC-MS chemical profiles.
• Electronic-nose responses, as prototype development progresses.
• Physiological and phenological observations.
• Meteorological and microclimatic measurements.
• Soil and environmental indicators.
• Tree identity, cultivar and spatial information.
• Sampling time, analytical batch and methodological metadata.
• Subsequent plant-health observations and outcomes.

Consistent metadata and quality-control records will support interpretation, reproducibility and independent evaluation.

Candidate methods include:

• Principal component analysis and multivariate exploratory statistics.
• Clustering and pattern discovery.
• Supervised classification and regression.
• Feature selection and interpretable modelling.
• Time-series analysis and change-point detection.
• Anomaly detection and uncertainty estimation.
• Sensor-drift correction and model calibration.

Method selection will depend on dataset size, quality and the research question. Statistical associations and model predictions will require separate biological evidence before causal interpretations are made.

Particular attention will be given to small-sample limitations, correlated observations, biological replication and data leakage. Model evaluation will separate independent trees, sampling periods, seasons or orchards wherever feasible, rather than randomly splitting closely related repeated measurements.

Performance will be compared with simpler statistical models and conventional environmental or physiological indicators.

Define sampling units, biological replication, phenological categories, independently assessed plant conditions and environmental measurements. Establish sampling protocols and quality-control procedures.

Expected outcome: A structured research protocol and initial phenotyping dataset.

Collect standardised plant-headspace samples and analyse them using GC-MS. Investigate temporal variability, phenological differences and exploratory associations with biological and environmental observations.

Expected outcome: The first structured instrumental VOC dataset and preliminary candidate chemical patterns.

Select sensors, construct sampling and acquisition interfaces, establish calibration procedures and compare responses with reference chemical measurements.

Expected outcome: A laboratory-evaluated prototype with documented performance and analytical limitations.

Integrate available chemical, sensor, biological and environmental data. Develop exploratory and predictive models and assess performance against conventional monitoring indicators.

Expected outcome: Preliminary models with documented uncertainty, performance and limitations.

Evaluate promising patterns and models using additional trees, sampling periods and environmental conditions and, where feasible, independent orchards and cultivars.

Expected outcome: Evidence supporting or rejecting specific monitoring and predictive applications.

The programme will prioritise:

• Independent biological reference measurements.
• Appropriate experimental and analytical controls.
• Reproducible sampling and documented quality assurance.
• Prespecified evaluation criteria wherever feasible.
• Assessment of uncertainty, false positives and false negatives.
• Separation of exploratory and confirmatory findings.
• Independent testing and comparison with conventional monitoring.
• Transparent reporting of negative, inconsistent and inconclusive results.

For any early-warning claim, evaluation must establish useful lead time, acceptable false-alarm rates and incremental value under representative orchard conditions.

Subject to experimental demonstration and independent validation, AIS could support further research and technology development in:

• Plant stress monitoring: Investigating chemical information associated with defined water, heat, nutritional or biotic stress conditions.
• Phenological and reproductive monitoring: Examining volatile changes during flowering, fruit development and other developmental transitions.
• Early-warning research: Testing whether defined patterns precede independently confirmed adverse outcomes.
• Precision horticulture: Assessing whether volatile measurements improve monitoring and help prioritise field inspection.
• Portable machine olfaction: Evaluating calibrated electronic-nose devices outside analytical laboratories.
• Biological and environmental research: Investigating relationships among plant physiology, surrounding vegetation and ecological interactions.

These are prospective research opportunities. They are not established AIS products or demonstrated field capabilities.

Practical deployment will require evidence of incremental value, operational reliability, acceptable cost and performance across representative orchard conditions.

AIS has established its qualitative observational foundation, central hypothesis, research objectives and proposed technology architecture.

The initial scientific focus is mango orchard physiology and plant volatile emissions. Extension to other crops and ecological systems will depend on the evidence developed through the initial programme.

The initiative is preparing to move from qualitative observation and conceptual design toward structured biological measurements, instrumental VOC characterisation and prototype development.

AIS has not yet demonstrated a validated diagnostic signature, predictive AI model, operational early-warning system or field-deployable electronic nose.

Establish the first structured AIS dataset linking standardised plant VOC measurements with independently assessed physiological and environmental observations.

This dataset will provide the foundation for an initial validation prototype, subsequent sensor development and computational modelling.

The intended progression is:

1. Qualitative ecological observation.
2. Scientific hypothesis and experimental design.
3. Baseline phenotyping and instrumental VOC profiling.
4. Electronic-nose development and analytical comparison.
5. Data integration and computational modelling.
6. Independent biological and field validation.
7. Evaluation of practical field guidance.

AIS seeks interdisciplinary scientific engagement to develop reproducible experimental evidence from its observational and conceptual foundation.

EcoPlanetFarm welcomes collaboration with universities, national research institutions, agricultural scientists, analytical laboratories, sensor engineers and computational research groups.

• Plant sciences: Plant physiology, horticulture, phenology, plant pathology and chemical ecology.
• Analytical chemistry: VOC sampling, GC-MS characterisation, calibration and analytical validation.
• Sensor engineering: Electronic-nose architectures, chemical sensor arrays, instrumentation and signal processing.
• Computational research: Chemometrics, AI/ML, multivariate statistics, uncertainty modelling and sensor-data interpretation.
• Precision agriculture: Field instrumentation, environmental monitoring and practical evaluation.
• Ecological research: Soil–plant–microbe interactions, landscape ecology, biodiversity and atmospheric chemistry.
• Olfaction science: Botanical volatiles, natural aroma systems and sensory–chemical relationships.

Participation may include experimental design, instrumentation access, field sampling, GC-MS analysis, prototype engineering, computational modelling and independent validation.

The immediate collaborative objective is to establish a scientifically defensible research protocol and the first structured instrumental AIS dataset.

Landscape Aroma Systems (LAS) is a broader ecological research direction associated with AIS.

AIS initially focuses on plant-associated volatile signatures and physiological interpretation. LAS considers volatile chemistry at landscape and ecosystem scales.

Atmospheric VOC mixtures may originate from vegetation, soil microorganisms, decomposing organic matter and other biological or environmental sources. Their composition and measured concentrations vary with vegetation structure, phenology, microbial processes, temperature, humidity, atmospheric transport and chemical transformation.

LAS proposes investigating whether repeated spatial and temporal measurements reveal reproducible patterns associated with ecological dynamics.

Potential research directions include:

• Spatial and seasonal variation in landscape volatile profiles.
• Vegetation-associated chemical signatures.
• Soil–plant–microorganism interactions.
• Chemical responses to environmental disturbance.
• VOC emissions in regeneratively managed landscapes.
• Relationships among vegetation dynamics, ecological processes and atmospheric chemistry.

The principal challenge is to distinguish source-specific biological emissions from mixed atmospheric signals and environmental confounding.

LAS remains a future research direction. No validated ecosystem-health or biodiversity indicator is presently claimed.

Its longer-term purpose is to examine whether volatile chemistry can contribute an additional dimension to ecological observation and environmental research.

Originator and Research Initiator — Aroma Intelligence Systems (AIS)

Bimal Das is an independent researcher, food technologist, olfaction and sensory-science practitioner, natural-products researcher and sustainability practitioner. His professional experience spans more than three decades of botanical materials, sensory applications, product development and ecological practice.

He holds an MSc in Food Technology from CSIR–Central Food Technological Research Institute (CFTRI), Mysore, following undergraduate studies in chemistry.

His interdisciplinary interests include plant-derived volatile compounds, natural aroma systems, olfactory perception, botanical product science, regenerative agriculture, biodiversity and plant–environment interactions.

Since 2012, EcoPlanetFarm in Raigad, Maharashtra, has provided a living field setting for qualitative observations of orchard phenology, seasonal aroma variation, soil regeneration and ecological interactions.

These observations formed the conceptual foundation for AIS. The initiative seeks to translate field-derived questions into testable hypotheses through collaboration across biological sciences, analytical chemistry, instrumentation and computational research.

 12. AIS Research Materials

Watch the introductory video and explore the scientific presentation below for an overview of AIS—its origins in qualitative ecological observations, scientific hypotheses, proposed research methods and validation pathway.

These materials are intended for researchers, academic institutions and potential technology collaborators. The presentation is also available to download as a PDF.