PhytoNodes for Environmental Monitoring: Stimulus Classification based on Natural Plant Signals in an Interactive Energy-efficient Bio-hybrid System

Buss, Eduard; Rabbel, Tim-Lucas; Horvat, Viktor; Križmančić, Marko; Bogdan, Stjepan; Wahby, Mostafa; Hamann, Heiko

doi:10.1145/3524458.3547266

Cited by 1 publication

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This challenges our goal of an energy-efficient, self-sufficient measurement unit, so this method is only applicable when sufficient energy is available. Given these results, and also comparing them to our previous work [14] where we achieved good results with deep learning methods for a qualitatively different dataset, we conclude that the decision regarding which classification technique to use, needs to be done on a case-to-case basis. Depending on the amount of available data and the feature selection, the simple statisitical methods can outperform the deep learning techniques.…”

Section: Discussionmentioning

confidence: 54%

“…In our experiments, we obtained approximately 1300 time series as a training dataset which is probably not sufficient for networks with larger numbers of trainable weights. In previous work [14], the same dataset was used. However, the time window used for classification was slightly larger (about 11.5 min compared to our 9.8 min), including about 1.5 min from the poststimulus phase.…”

Section: Discussionmentioning

confidence: 99%

“…In previous work, we designed an energy-efficient sensor node called PhytoNode for plant electrical potential (not used here). It can run classification models on board and interact, for example, with handheld devices of close-by citizens via Bluetooth Low Energy [14]. In this previous work, we exposed Zamioculcas Zamiifolia plants to four stimuli: wind, heat, blue light, and red light for data collection.…”

Section: Introductionmentioning

confidence: 99%

“…In our EU-funded project 'Watch-Plant' [16,17] (2021-2024), we use phytosensing to measure environmental features via living plants. We develop a self-powered, energy-efficient sensor network of electronic devices attached to plants [14,18]. This forms a so-called bio-hybrid system of strongly coupled technological devices and living organisms.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Stimulus classification with electrical potential and impedance of living plants: comparing discriminant analysis and deep-learning methods

et al. 2023

View full text Add to dashboard Cite

The physiology of living organisms, such as living plants, is complex and particularly difficult to understand on a macroscopic, organism-holistic level. Among the many options to study plant physiology, electrical potential and tissue impedance are arguably simple measurement techniques to gather plant-level information. Despite the many possible uses, our research is exclusively driven by the idea of phytosensing, that is, interpreting living plants' signals to learn information about surrounding environmental conditions. As ready-to-use plant-level physiological models are not available, we consider the plant as a blackbox and apply statistics and machine learning to automatically interpret measured signals. In simple plant experiments, we expose Zamioculcas zamiifolia and Solanum lycopersicum (tomato) to four different stimuli: wind, heat, red and blue light. We measure electrical potential and tissue impedance signals. Given these signals, we evaluate a large variety of methods from statistical discriminant analysis and from deep learning for the classification problem of determining the correct stimulus to which the plant was exposed. We identify a set of methods that successfully classify stimuli with good accuracy without a clear winner. The statistical approach is competitive, partially depending on data availability for the machine learning approach. Our extensive results show the feasibility of the blackbox approach and can be used in future research to select appropriate classifier techniques for a given use case. In our own future research, we will exploit these methods to drive a phytosensing approach for air pollution monitoring in urban areas.

show abstract

Section: Discussionmentioning

confidence: 54%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%