Search

Your search keyword '"Dhami, Harnaik Singh"' showing total 4 results
Start Over Author "Dhami, Harnaik Singh"
4 results on '"Dhami, Harnaik Singh"'

1. Intermittent Deployment for Large-Scale Multi-Robot Forage Perception: Data Synthesis, Prediction, and Planning

Author

Liu, Jun, Rangwala, Murtaza, Ahluwalia, Kulbir Singh, Ghajar, Shayan, Dhami, Harnaik Singh, Tokekar, Pratap, Tracy, Benjamin F., and Williams, Ryan K.

Subjects
Computer Science - Robotics
Abstract

Monitoring the health and vigor of grasslands is vital for informing management decisions to optimize rotational grazing in agriculture applications. To take advantage of forage resources and improve land productivity, we require knowledge of pastureland growth patterns that is simply unavailable at state of the art. In this paper, we propose to deploy a team of robots to monitor the evolution of an unknown pastureland environment to fulfill the above goal. To monitor such an environment, which usually evolves slowly, we need to design a strategy for rapid assessment of the environment over large areas at a low cost. Thus, we propose an integrated pipeline comprising of data synthesis, deep neural network training and prediction along with a multi-robot deployment algorithm that monitors pasturelands intermittently. Specifically, using expert-informed agricultural data coupled with novel data synthesis in ROS Gazebo, we first propose a new neural network architecture to learn the spatiotemporal dynamics of the environment. Such predictions help us to understand pastureland growth patterns on large scales and make appropriate monitoring decisions for the future. Based on our predictions, we then design an intermittent multi-robot deployment policy for low-cost monitoring. Finally, we compare the proposed pipeline with other methods, from data synthesis to prediction and planning, to corroborate our pipeline's performance., Comment: 21 pages, 26 figures, submitted to IEEE Transactions on Automation Science and Engineering

Published
2021

3. DeepPaSTL: Spatio-Temporal Deep Learning Methods for Predicting Long-Term Pasture Terrains Using Synthetic Datasets

Author

Rangwala, Murtaza, Liu, Jun, Ahluwalia, Kulbir Singh, Ghajar, Shayan, Dhami, Harnaik Singh, Tracy, Benjamin F., Tokekar, Pratap, Williams, Ryan K., Rangwala, Murtaza, Liu, Jun, Ahluwalia, Kulbir Singh, Ghajar, Shayan, Dhami, Harnaik Singh, Tracy, Benjamin F., Tokekar, Pratap, and Williams, Ryan K.

Abstract

Effective management of dairy farms requires an accurate prediction of pasture biomass. Generally, estimation of pasture biomass requires site-specific data, or often perfect world assumptions to model prediction systems when field measurements or other sensory inputs are unavailable. However, for small enterprises, regular measurements of site-specific data are often inconceivable. In this study, we approach the estimation of pasture biomass by predicting sward heights across the field. A convolution based sequential architecture is proposed for pasture height predictions using deep learning. We develop a process to create synthetic datasets that simulate the evolution of pasture growth over a period of 30 years. The deep learning based pasture prediction model (DeepPaSTL) is trained on this dataset while learning the spatiotemporal characteristics of pasture growth. The architecture purely learns from the trends in pasture growth through available spatial measurements and is agnostic to any site-specific data, or climatic conditions, such as temperature, precipitation, or soil condition. Our model performs within a 12% error margin even during the periods with the largest pasture growth dynamics. The study demonstrates the potential scalability of the architecture to predict any pasture size through a quantization approach during prediction. Results suggest that the DeepPaSTL model represents a useful tool for predicting pasture growth both for short and long horizon predictions, even with missing or irregular historical measurements.

Published
2021

4. Using UAV Mounted LiDAR to Estimate Plant Height and Growth

Author

Dhami, Harnaik Singh, Electrical and Computer Engineering, Tokekar, Pratap, Williams, Ryan K., and Li, Song

Subjects
LiDAR, PrecisionAgriculture, Plot Detection, Plant Height Estimation, K-means
Abstract

In this thesis, we develop algorithms to estimate crop heights as well as to detect plots infarms. Plant height estimation is needed in precision agriculture to monitor plant health andgrowth cycles. We use a 3D LiDAR mounted on an Unmanned Aerial Vehicle (UAV) anduse the LiDAR data for height and plot estimation. We present a general methodology forextracting plant heights from 3D LiDAR with two specific variants for the two environments:row-crops and pasture. The main algorithm is based on ground plane estimation from 3DLiDAR scans, which is then used to determine the height of plants in the scans. For rowcrops, the plot detection uses a K-means clustering algorithm to find the bounding boxes ofthese clusters, and a voting scheme to determine the best-fit width, height, and orientationof the clusters/plots. This best-fit box is then used to create a grid over the LiDAR dataand the plots are extracted. For pasture, relative heights are estimated using data collectedweekly. Both algorithms we evaluated using data collected from actual farms and pasture.The accuracy in plot height estimation was +/- 5.36 % and that for growth estimates was+/- 7.91 %. Master of Science Plant height estimation and measurement is a vital task when it comes to farming. Knowing these characteristics help determine whether the plants are growing healthy and when to harvest them. On similar lines, accurate estimates of the plant heights can be used to prevent overgrazing and undergrazing of pastures. However, as farm and plot size increases, getting consistent and accurate measurements becomes a more time-consuming and manually intensive task. Using robots can help solve this problem because they can be used to estimate the height. With sensors that are already available, such as the 3D LiDAR that we use, we can use aerial robots to fly over the farm and collect plant data. This data can then be processed to estimate the plant height, eliminating the need to go out and manually measure every single plant. This thesis discusses a methodology of doing exactly this, as well as detecting plots within a farm. The algorithms are evaluated using data collected from actual farms and pasture.

Published
2019

Catalog

Books, media, physical & digital resources
See catalog results