Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Data Optimisation of Machine Learning Models for Smart Irrigation in Urban Parks

Nasser Ghadiri, Bahman Javadi, Oliver Obst, Sebastian Pfautsch

TL;DR

This paper addresses the reliability and cost of data in smart irrigation for urban parks under climate stress by introducing two techniques: clustering-based estimation of missing sensor readings (using DTW-based K-means and K-shape clustering) and sequential robotic data collection to reduce fixed sensors. The DTW-based clustering consistently yields more accurate sensor replacements with lower variability, while the robotic sensing strategy reduces maintenance needs and maintains accurate soil-moisture predictions, achieving MAE improvements up to 17.2% on circular paths and 2.1% on linear paths. Overall, the methods enhance data quality, lower sensor-network costs, and improve resilience of the SIMPaCT irrigation system against sensor failures, supporting scalable deployment in similar urban parks. The work demonstrates practical gains in data reliability and operational efficiency, with future plans to generalize to other parks and weather conditions.

Abstract

Urban environments face significant challenges due to climate change, including extreme heat, drought, and water scarcity, which impact public health, community well-being, and local economies. Effective management of these issues is crucial, particularly in areas like Sydney Olympic Park, which relies on one of Australia's largest irrigation systems. The Smart Irrigation Management for Parks and Cool Towns (SIMPaCT) project, initiated in 2021, leverages advanced technologies and machine learning models to optimize irrigation and induce physical cooling. This paper introduces two novel methods to enhance the efficiency of the SIMPaCT system's extensive sensor network and applied machine learning models. The first method employs clustering of sensor time series data using K-shape and K-means algorithms to estimate readings from missing sensors, ensuring continuous and reliable data. This approach can detect anomalies, correct data sources, and identify and remove redundant sensors to reduce maintenance costs. The second method involves sequential data collection from different sensor locations using robotic systems, significantly reducing the need for high numbers of stationary sensors. Together, these methods aim to maintain accurate soil moisture predictions while optimizing sensor deployment and reducing maintenance costs, thereby enhancing the efficiency and effectiveness of the smart irrigation system. Our evaluations demonstrate significant improvements in the efficiency and cost-effectiveness of soil moisture monitoring networks. The cluster-based replacement of missing sensors provides up to 5.4% decrease in average error. The sequential sensor data collection as a robotic emulation shows 17.2% and 2.1% decrease in average error for circular and linear paths respectively.

Data Optimisation of Machine Learning Models for Smart Irrigation in Urban Parks

TL;DR

This paper addresses the reliability and cost of data in smart irrigation for urban parks under climate stress by introducing two techniques: clustering-based estimation of missing sensor readings (using DTW-based K-means and K-shape clustering) and sequential robotic data collection to reduce fixed sensors. The DTW-based clustering consistently yields more accurate sensor replacements with lower variability, while the robotic sensing strategy reduces maintenance needs and maintains accurate soil-moisture predictions, achieving MAE improvements up to 17.2% on circular paths and 2.1% on linear paths. Overall, the methods enhance data quality, lower sensor-network costs, and improve resilience of the SIMPaCT irrigation system against sensor failures, supporting scalable deployment in similar urban parks. The work demonstrates practical gains in data reliability and operational efficiency, with future plans to generalize to other parks and weather conditions.

Abstract

Urban environments face significant challenges due to climate change, including extreme heat, drought, and water scarcity, which impact public health, community well-being, and local economies. Effective management of these issues is crucial, particularly in areas like Sydney Olympic Park, which relies on one of Australia's largest irrigation systems. The Smart Irrigation Management for Parks and Cool Towns (SIMPaCT) project, initiated in 2021, leverages advanced technologies and machine learning models to optimize irrigation and induce physical cooling. This paper introduces two novel methods to enhance the efficiency of the SIMPaCT system's extensive sensor network and applied machine learning models. The first method employs clustering of sensor time series data using K-shape and K-means algorithms to estimate readings from missing sensors, ensuring continuous and reliable data. This approach can detect anomalies, correct data sources, and identify and remove redundant sensors to reduce maintenance costs. The second method involves sequential data collection from different sensor locations using robotic systems, significantly reducing the need for high numbers of stationary sensors. Together, these methods aim to maintain accurate soil moisture predictions while optimizing sensor deployment and reducing maintenance costs, thereby enhancing the efficiency and effectiveness of the smart irrigation system. Our evaluations demonstrate significant improvements in the efficiency and cost-effectiveness of soil moisture monitoring networks. The cluster-based replacement of missing sensors provides up to 5.4% decrease in average error. The sequential sensor data collection as a robotic emulation shows 17.2% and 2.1% decrease in average error for circular and linear paths respectively.
Paper Structure (20 sections, 5 equations, 5 figures, 1 table)

This paper contains 20 sections, 5 equations, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. The Proposed Data Optimisation Techniques
  4. Using Sensor Time Series as Input for Clustering
  5. Design Considerations and Methodology Selection
  6. DTW-based K-means Clustering Method
  7. k-Shape Clustering Method
  8. Silhouette Score for Evaluation of Clustering
  9. Replacing the Time Series of a Missing Sensor
  10. Sequential Data Collection Method for Robotic Sensing
  11. Experimental Evaluation
  12. Experimental Dataset
  13. Using Silhouette Score for Determining k in Clustering
  14. Clustering of Moisture Sensor Time Series
  15. Results Of Sensor Replacement
  16. ...and 5 more sections

Figures (5)

  • Figure 1: The SIMPaCT system architecture
  • Figure 2: Clustering for sensor SENS0098-SM-SOPA replacement: (a) A cluster obtained using the K-shape clustering method, that contains five sensors showing nearly identical trends across both months. (b) A cluster from the DTW K-means clustering method, which includes 12 sensors, shows similarities in the peaks and waveform patterns for both months
  • Figure 3: Clustering for sensor SENS0107-SM-SOPA replacement: (a) A cluster obtained using the K-shape, which contains 10 sensors that exhibit nearly identical trends across both months, with the exception of one sensor that reports significantly different values in September. (b) A cluster from the DTW K-means results, which includes six sensors. These sensors demonstrate similarities in the peaks and waveform patterns for both months.
  • Figure 4: MAE box plot for using two clustering methods to replace SENS0098-SM-SOPA and SENS0107-SM-SOPA.
  • Figure 5: Selected areas for robot emulation and the resulting MAE values for each area: median MAE for the circular path is slightly lower than the base values, with a slightly higher spread