My interest in Artificial Intelligence dates back to my undergraduate, when I learned about Neural Networks. It was the first time I came to know that machines can be intelligent enough to do our work and they can learn to do things on their own. I personally love to cook and I envision building a robot that knows how to cook and at the same time, can adapt and learn in itself to try new variations and combinations to make new and tasty dishes. I think that would be a great innovation!
I joined Washington State University in Fall 2015 as a Masters Student in Computer Science, before that I completed my B.S. from JSS Academy of Technical Education, India.
I am currently advised by Dr. Matthew Taylor who gave me an opportunity to gain some relevant experience and do some cool research in the field of Machine Learning and Robotics.
Cherry, grape, honeycrisp apple, and blueberry growers lose $80 million annually to bird damage in the state of Washington (WA) alone. Over 50% of sweet cherry growers in Washington consider bird damage as one of the significant factors affecting profit. Netting, auditory scare devices, visual scare devices, chemical applications, and active methods such as trapping, falconry, and lethal shooting are the most common ways bird damage is minimized currently. Netting and falconry are the only methods viewed by most growers as effective, but are costly, and netting is potentially lethal to a host of non-target wildlife. Growers of a wide range of crops have a critical need for a safe, cost effective method for persistent bird deterrence, resulting in reduced costs and increased sustainability.
Unmanned Aerial Systems (UAS) have recently grown in popularity in military, civilian, and agriculture domains due to their decreasing cost, maneuverability, and ability to tackle multiple types of missions. WA is expected to become a leader in the development and manufacture of UAS, with anticipated $1.3 billion in economic activity and 6,746 jobs by 2017. It is expected that agriculture will see more than 80% of the total growth in UAS-related economic activities.
Therefore, UAS and its application in agriculture is a critically important emerging area for WSU to lead the effort in the nation. Lack of such research activities at WSU is a critical drawback and is limiting our ability to bring WSU to the forefront of this emerging issue. We strongly believe that the seed funding in this area will help us be highly competitive in securing extramural funding and develop highly beneficial agricultural applications of UAS for WA growers.
The long-term objective of the proposed research is to develop an autonomous UAS for bird deterrence, as well as fully test its effectiveness and financial implications. In addition to bird deterrence in high-value fruit crops, such a system would also be able to deter birds in a variety of other situations, such as at landfills, dairy farms, and airports.
The goal of this research is to improve the modern process of counting individual trees in a nursery, which at the moment is inefficient, costly, and inaccurate. Multiple times a year, employees must walk down every row in a tree nursery and count each tree in order to have an estimate of the nurseries stock. This tedious and time consuming job could be replaced by an autonomous rover, creating the opportunity to save time and money, as well as to both get a more accurate estimate of the number of trees and to record the diameters of the different trees.
Our research project designs, builds, programs, and tests an autonomous rover that can be placed in a tree nursery and successfully navigate through each row. Although this has been done before by other universities and companies, such systems run $50,000 or more. We currently are operating under a total budget of a $1,000.
In order to stay under budget we decided to create a low-cost LIDAR system to accurately estimate the position of the trees in relation to the robot. This custom LIDAR system maps the surrounding unknown environment and plots each data point. We then use various algorithms to cluster the data into 4 points, 2 points on each row. These points are used to create lines that symbolize the left and right rows of trees. The rover then determines its position in the row and calculates the movement needed to drive down the middle of the row.
Video and Media