ItemArticle

Neural network-assisted carbon nanotube electrochemical sensors for automated environmental risk assessment

Title
Neural network-assisted carbon nanotube electrochemical sensors for automated environmental risk assessment
Creator
Chakravarthy, V.J.; Thontadari, C.; Basavaraj G.N.; Sowndharya, R.; Aravindan, N.; Sangeetha, V.
Description
The current research proposes an intelligent network system that continuously tracks the quality of water flow, with particular attention to pollutants frequently occurring in runoff water from agricultural practices. It deploys high-performance electrochemical sensors based on carbon nanotubes (CNT) combined with a small neural network that functions directly on the built-in microcontroller. It is deployed on floating buoys powered by solar energy, where it can detect some critical contaminants in the rural water bodies, including nitrates, phosphates, atrazine, cadmium, and lead. The sensors work by transmitting their electrical signals through the sensors to the neural network, which provides precise identification of the level of pollutants as one of three risk levels: safe (below detection levels), manageable (within regulatory levels), and hazardous (exceeding regulations). Regarding power performance, results can be delivered over a relatively small-time delay (1.2 milliseconds per reading) and with low memory usage (1.8 MB), making it ideal for remote and low-powered sensors. It is more accurate (93.6 %) than typical machine learning models. Should pollutants exceed the above-prescribed limits, an automated warning will be generated, and the information will be immediately uploaded to a cloud-based dashboard. The dashboard will be closely monitored via remote control, and trend analysis will be conducted. By eliminating the need for manual water sampling, the system offers a scalable and energy-saving method for autonomous environmental testing, particularly in inaccessible locations. In the future, the study will focus on the use of federated learning, a technique that retains data locally to protect privacy, enabling more intelligent and collaborative conclusions across sensor networks. This prepares the ground for more intelligent and secure environmental surveillance systems in the future. 2025 Elsevier B.V.
Source
Microchemical Journal;Volume;218;Issue;;Article No.;115413;
Date
01-01-2025
Publisher
Elsevier Inc.
Subject
Carbon nanotube sensors; Electrochemical sensors; Environmental monitoring; Neural networks; Water quality monitoring
Coverage
Chakravarthy V.J., Faculty of Computer Applications, Dr. M. G. R. Educational and Research Institute, Maduravoyal, Chennai, 95, India; Thontadari C., Department of Computer Science, Christ University, Bangalore, India; Basavaraj G.N., Department of Information Science and Engineering, BMS Institute of Technology & Management, Karnataka, Bengaluru, India; Sowndharya R., Department of Computer Science and Engineering, Sona College of Technology, Salem, India; Aravindan N., Department of Fire Engineering, National Fire Service College, Nagpur, India; Sangeetha V., Faculty of Computer Applications, Dr. M. G. R. Educational and Research Institute, Chennai, 95, India
Rights
Restricted Access; Hardcopy may be available in the library
Relation
ISSN: 0026265X; CODEN: MICJA
Format
online
Language
English
Type
Article

Citation

Chakravarthy, V.J.; Thontadari, C.; Basavaraj G.N.; Sowndharya, R.; Aravindan, N.; Sangeetha, V., “Neural network-assisted carbon nanotube electrochemical sensors for automated environmental risk assessment,” CHRIST (Deemed To Be University) Institutional Repository, accessed June 18, 2026, https://archives.christuniversity.in/items/show/22393.

Output Formats