Progress 01/01/18 to 12/31/23
Outputs Target Audience: The targeted audiences for our efforts have been agricultural industry, agricultural producers, USDA, NIFA, NSF, and several agricultural commissions. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Paid for trips to several professional conferences such as ASABE and How have the results been disseminated to communities of interest?Multiple publications, patent, and news stories What do you plan to do during the next reporting period to accomplish the goals?Final report
Impacts What was accomplished under these goals?
Aim 1:Create vehicle mobility, design, safety, and route-optimization models for highly sloped land from data on traction, power usage, slip, maneuverability, and soil compaction. Vehicle behavior studies were performed first on a soil test bed that could be tilted to the desired angle then in a geographical location that was digitally mapped out. The results of these studies as well as digital maps of ageographical locationwere used to developmachine learning methods to route a team of robots along optimal trajectories. The articles: Badgujar (2023, Deep neural networks to predict autonomous ground vehicle behavior on sloping terrain field), Badgujar (2022, Artificial neural network to predict traction performance of autonomous ground vehicle on a sloped soil bin and uncertainty analysis), and Badgujar (2022, Experimental investigation on traction, mobility, and energy usage of a tracked autonomous ground vehicle on a sloped soil bin) were published on the findings and a review paper Figueroa (2023, . Application of computational intelligence methods in agricultural soil-machine interaction: A review). Further studies were done with the route planning and optimization with articles: Das (2019, Recurrent neural network based multi-robot route planning for steep-land harvesting systems) and Figueroa (2022, A distributed approach for robotic coverage path planning under steep slope terrain conditions) with also a Patent: Das(11856882:2022, Autonomous robot system for steep terrain farming operations). Aim 2:Refine a planter usable by small autonomous vehicles. Extensive effort was expended on the no till drill design for a small robotic vehicle. Due to lack of down force the system could not accomodate a conventional no till drill system. Due to tration needs (high sloped hills) shank systems were tested and deamed too draw bar needy. A final powered trencher system was designed then tested and found to be adequate but not efficient. An Auger seeder system was also designed and tested in differing vibration schemes and angles. The seeding system worked well for these smaller vehicle systems and proved robust for high sloped hill angles. These results were detailed the article Badgujar (2022,
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Dania Martinez Figueroa Chetan Badgujar, Sanjoy Das and Daniel Flippo. Application of computational intelligence methods in agricultural soil-machine interaction: A review. agri- culture, 13(2):357, January 2023
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Chetan Badgujar, Daniel Flippo, Sylvester Badua, and Carolyn Baldwin. Development and evaluation of pasture tree cutting robot. Croatian journal of forest engineering, 44(1), 2023
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Chetan Badgujar, Sanjoy Das, Dania M. Figueroa, Daniel Flippo, and Stephen Welch. Deep neural networks to predict autonomous ground vehicle behavior on sloping terrain field. Jour- nal of Field Robotics, 40(4):919933, 2023
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
H. Mansur, S. Welch, L. Dempsey, and D. Flippo. Importance of photo-realistic and dedicated simulator in agricultural robotics. Engineering, 15:318327, May 2023
|
Progress 01/01/22 to 12/31/22
Outputs Target Audience: The targeted audiences for our efforts have been agricultural industry, agricultural producers, USDA, NIFA, NSF, and several agricultural commissions. In addition to these audiences we speak to robotics communities and societies as well as STEM organizations. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?4 publications and 2 thesis's disseminated the results. What do you plan to do during the next reporting period to accomplish the goals?finish up robot testing and get field testing done to show the life cycle expectaion of the newly designed planter.
Impacts What was accomplished under these goals?
Aim 2: A two row till planter that was designed in 2020 and 2021 wastested in 2022. The prototype was tested for its functionality in laboratory vibration and slope tests. Results showed that the settled on auger design worked well with the vibration frequency generated by the robot and at slopes that are expected in this project. A publication was written and submitted and will be published in the final year. Aim 1:Create vehicle mobility, design, safety, and route-optimization models for highly sloped land from data on traction, power usage, slip, maneuverability, and soil compaction. Using experimental data collected from laboratory experiments, Dr. Flippo's team of investigators developed a novel deep learning model to predict robot's performance under varying land conditions. The model consistently outperforms all existing AI prediction models. Two journal articles from this research effort are currently under revision for journal resubmission after an initial round of review [1, 2]. The team has made further enhancement to their approach. In future, the extended model and more simulation results are planned to be submitted for journal publication. Dr. Flippo's team of investigators have developed new distributed algorithms for optimal multi-robotic planning in steep-land harvesting systems. Initial research with a single robot has been accepted for publication [3]. Further simulations by the team have shown effectiveness of this approach in optimally dividing the task into multiple sub-tasks to be carried out by an equal number of robots. The figure (see right) is an illustrative example that shows the outcome of the approach for task optimal coverage path planning and task division between three robots, on an uneven tract of land. The proposed approach also outperforms other recent algorithms on similar tasks. A manuscript intended for journal submission based on these results is under preparation [4]. Further research on imparting online learning capabilities for active coordination between several robots are being explored by Dr. Flippo's team, with the aim of future publications. Separately, the team is investigating hardware realizable multi-robot planning algorithms [cf. 5] for a possible patent as well as other future publications. [1] Chetan Badgujar, Dania Martinez Figueroa, Sanjoy Das, Daniel Flippo, "Deep neural networks to predict autonomous ground vehicle behavior on sloping terrain field,"Journal of Field Robotics, under revision. [2] Chetan Badgujar, Dania Martinez Figueroa, Sanjoy Das, Daniel Flippo, "Application of computational intelligence methods in agricultural soil-machine interaction : A review,"Agriculture,Design and Application of Agricultural Equipment in Tillage System(Special Issue), under revision. [3] Dania Martinez Figueroa, Sanjoy Das, Chetan Badgujar, Daniel Flippo, Stephen M. Welch, "A Distributed Approach for Robotic Coverage Path Planning Under Steep Slope Terrain Conditions,"Proc.IEEE Symposium Series on Computational Intelligence, 2022, in press. [4] Dania Martinez Figueroa, Sanjoy Das, Chetan Badgujar, Daniel Flippo, "Coverage Path Planning in Multi-robots for Steep-Land Harvesting Systems,"under preparation. [Planned for submission to an IEEE journal]. [5] Sanjoy Das, Stephen M. Welch, Daniel Flippo, "Recurrent Neural Network Based Multi-Robot Route Planning for Steep-Land Harvesting Systems,"Proc.International Congress on Advanced Applied Informatics, pp. 542-545, 2019. doi.org/10.1109/IIAI-AAI.2019.00116
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2022
Citation:
Chetan Badgujar, Daniel Flippo, and Stephen Welch. Artificial neural network to predict traction performance of autonomous ground vehicle on a sloped soil bin and uncertainty analysis. Computers and Electronics in Agriculture, 196:106867, 2022.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Chetan M. Badgujar, Daniel Flippo, Edwin Brokesh, and Stephen Welch. Experimental investigation on traction, mobility, and energy usage of a tracked autonomous ground vehicle on a sloped soil bin. volume 65(4), pages 835847. ASABE Annual International Meeting., 2022.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Dania Martinez Figueroa, Sanjoy Das, Chetan Badgujar, Daniel Flippo, Stephen M. Welch, "A Distributed Approach for Robotic Coverage Path Planning Under Steep Slope Terrain Conditions," Proc. IEEE Symposium Series on Computational Intelligence, 2022, in press.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Sanjoy Das, Stephen M. Welch, Daniel Flippo, "Recurrent Neural Network Based Multi-Robot Route Planning for Steep-Land Harvesting Systems," Proc. International Congress on Advanced Applied Informatics, pp. 542-545, 2019. doi.org/10.1109/IIAI-AAI.2019.00116
- Type:
Theses/Dissertations
Status:
Published
Year Published:
2022
Citation:
Brownstein, N. : Assessing the predictive ability of multispectral and geomorphometric data for soil rock fraction and bulk density . MS Thesis, MS in Geography, Kansas State University.
- Type:
Theses/Dissertations
Status:
Published
Year Published:
2022
Citation:
Badgujar, Chetan. : Robotic farming on marginal, highly sloped lands. PhD Dissertation, PhD in Biological and Agricultural Engineering, Kansas State University, 2022
|
Progress 01/01/21 to 12/31/21
Outputs Target Audience: The targeted audiences for our efforts have been agricultural industry, agricultural producers, USDA, NIFA, NSF, and several agricultural commissions. In addition to these audiences we speak to robotics communities and societies as well as STEM organizations. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest? Badgujar, Chetan, Daniel Flippo, Ed Brokesh, and Stephen Welch. "Experimental Investigation on Traction, Mobility, and Energy Usage of the Tracked Autonomous Ground Vehicle on a Sloped Soil Bin."Transaction of ASABE, Under reivew. Badgujar, Chetan, Daniel Flippo, and Stephen Welch."Artificial Neural Network to Predict Traction Performance of Autonomous Ground Vehicle on a Sloped Soil Bin and Uncertainty Analysis." Computer and Electronics in Agriculture, Under review. What do you plan to do during the next reporting period to accomplish the goals? Still hoping to get a planter that will do no till without the downforce as well as incorporate the routing plan work that is being accomplished.
Impacts What was accomplished under these goals?
Aim 1: Soft computing based approach/models was developed for real time optimization of a vehicle traction performance on highly slope terrain. further tests have been done on real world slopes with transitions. This data has been used to further train the model and 2 papers are currently under review for the research. Aim 2: A two row till planter was designed, fabricated and tested in 2020. The prototype was tested for its functionality in laboratory soil bin. A field trail of two vehicles with planter mounted was conducted to seed distribution in field conditions. Initial results were encouraging, though a need of optimization and modification has been identified. Lab tests are ongoing testing the vibration as well as the slope of the planter proficiencywhile under these conditions. We are still only at 3/5 graduate student capacity due to COVID and visa regulations.
Publications
- Type:
Journal Articles
Status:
Under Review
Year Published:
2021
Citation:
1. Badgujar, Chetan, Daniel Flippo, Ed Brokesh, and Stephen Welch. Experimental Investigation on Traction, Mobility, and Energy Usage of the Tracked Autonomous Ground Vehicle on a Sloped Soil Bin.�Transaction of ASABE, Under reivew.?
- Type:
Journal Articles
Status:
Under Review
Year Published:
2021
Citation:
2. Badgujar, Chetan, Daniel Flippo, and Stephen Welch.��Artificial Neural Network to Predict Traction Performance of Autonomous Ground Vehicle on a Sloped Soil Bin and Uncertainty Analysis. Computer and Electronics in Agriculture, Under review.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Chetan Badgujar, Dan Flippo. (2021) Performance of Autonomous Ground Vehicle on Varying slope. Kansas State University in the 18th Capitol Graduate Research Summit, Topeka KS
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Research and the State GRADUATE STUDENT POSTER SESSION 2021, ARTIFICIAL NEURAL NETWORK TO PREDICT TRACTION PERFORMANCE OF AUTONOMOUS GROUND VEHICLE ON A SLOPED SOIL BIN AND UNCERTAINTY ANALYSIS. Chetan Badgujar and Daniel Flippo
|
Progress 01/01/20 to 12/31/20
Outputs Target Audience: The targeted audiences for our efforts have been agricultural industry, agricultural producers, USDA, NIFA, NSF, and several agricultural commissions. In addition to these audiences we speak to robotics communities and societies as well as STEM organizations. Changes/Problems:We are still understaffed due to visa regulations, COVID, and having a harder time getting graduate students to come to Kansas. We have just recently had 2 graduate students arrive. So we will need to due a no cost extension. What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?The presentation on "Tractive performance of tracked ground autonomous vehicle on varying slopes" was delivered, by Chetan Badgujar, atK-State Research and State forum furtherselected to present research at Kansas Capitol event (Capitol Graduate Research Summit 2021).? This poster was one of 10 selected from a field of 35 to represent KSU at the 18th Capitol Graduate Research Summit, which is attended by state legislators, the Kansas Board of Regents and the general public. What do you plan to do during the next reporting period to accomplish the goals?Seeder calibration to obtain desired seed rate (pounds/acre). A variety of field test to check seed distribution pattern and optimize the planter performance in terms of power and efficiency. Laboratory simulation studies including finite element or discrete element to check soil and tool interaction. The focus of this next year is navigation and guidance. Our new graduate student that is joining this month will develope this concept.
Impacts What was accomplished under these goals?
Aim 1: Soft computing based approach/models was developed for real time optimization of a vehicle traction performance on highly slope terrain. Aim 2:A two row till planter was designed, fabricated and tested in 2020. The prototype was tested for its functionality in laboratory soil bin. A field trail of two vehicles with planter mounted was conducted to seed distribution in field conditions. Initial results were encouraging, though a need of optimization and modification has been identified. We are still only at 2/5 graduate student capacity due to COVID and visa regulations. We should have 2 more graduate students joinging us this month and the final one joinging this Spring.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2020
Citation:
TRACTIVE PERFORMANCE OF SMALL TRACKED GROUND AUTONOMOUS VEHICLE ON VARYING SLOPES, Chetan Badgujar, Daniel Flippo (2020) State GRAD Forum, March 25th
- Type:
Other
Status:
Published
Year Published:
2020
Citation:
Using-mechanically-engineered-robots-to-feed-the-world article in Futurum magazine. https://futurumcareers.com/using-mechanically-engineered-robots-to-feed-the-world
|
Progress 01/01/19 to 12/31/19
Outputs Target Audience:On September 21st I gave a seminar to 50 STEM high school, Junior High, and STEM teachers about engineering practices with robotics. I presented showing the AgDrone vehicle that was made possible through this NRI2017 grant. On June 14th 2019 the Kansas State University Biological and Agricultural Departementhosted an advisory meeting with industry advisors from John Deere, Case New Holland, and Teejet. For the morning session 20 STEM science teachers joined us to discuss the 2050 dilemma (feeding over 9 billion people in 2050) and howw robotics are working to mitigate that problem. It was a successful day with the teachers eyes being opened to both higher academic and industry views of: Failure: Fail early and fail often in design, this was somethign newto the STEM teachers and they heard it from the industry partners as well as saw cases of failure in the academic research sense. 2050: This was a brand new concept to them and one that needs to be taught early on to grow awareness. So that the future generation will vote and work toward the same goals that the USDA are working toward. Another artifact, from this adivsory meeting, was a collaboration between one of the companies who wished to lease the algorithm discussed in the meeting. This is currently ongoing and the company is looking into either an additional research project or leaseing the patent license. Magazine articles: An Article about the NRI-2017 agricultural automation on high sloped hills is inSEEK Magazine, which is the Kansas State University's research magazine.The audience for this magazine includes donors and potential donors; administrators and researchers at other universities; stakeholders in agriculture and ag research. Another article is in the K-Statermagazine, whichgoes out to our K-State Alumni Association membership of more than 41,000 alumni. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Software training in Unreal Engine 4 modeling and simulation. Chetan Badgujar has been learning UE4 for the purpose of robotic simulation. How have the results been disseminated to communities of interest?On September 21st a seminar was givento 50 STEM high school, Junior High, and STEM teachers about engineering practices with robotics. I presented showing the AgDrone vehicle that was made possible through this NRI2017 grant. On June 14th 2019 the Kansas State University Biological and Agricultural Departementhosted an advisory meeting with industry advisors from John Deere, Case New Holland, and Teejet. For the morning session 20 STEM science teachers joined us to discuss the 2050 dilemma (feeding over 9 billion people in 2050) and howw robotics are working to mitigate that problem. It was a successful day with the teachers eyes being opened to both higher academic and industry views of: Failure: Fail early and fail often in design, this was somethign newto the STEM teachers and they heard it from the industry partners as well as saw cases of failure in the academic research sense. 2050: This was a brand new concept to them and one that needs to be taught early on to grow awareness. So that the future generation will vote and work toward the same goals that the USDA are working toward. Magazine articles: An Article about the NRI-2017 agricultural automation on high sloped hills is inSEEK Magazine, which is the Kansas State University's research magazine.The audience for this magazine includes donors and potential donors; administrators and researchers at other universities; stakeholders in agriculture and ag research. Another article is in the K-Statermagazine, whichgoes out to our K-State Alumni Association membership of more than 41,000 alumni. What do you plan to do during the next reporting period to accomplish the goals?Aim 1: Geologically improve methods to easily estimate crucial ag-soil variables for robot-ag on hilly terrain: bulk density and stoniness Aim 2: Refine the planter design and measure efficiency and seed placement and compare to conventional systems. Aim 4: Refine stripper header design. Refine composer/conductor system
Impacts What was accomplished under these goals?
Aim1: 1) Tilt bed testing was finished and data is being analysed to be used in simulation. 2) Unreal engine 4 has been used to create a simulation of a trageted location in Kansas using Satelite topological data as well as a simplified 3D CAD model of the AgDrone robot. PAramteres of this simulation can be changed to match that of the real life tilt bed testing, such as coefficient of friction and powe usage, so that the simulation can be used to create optimized routes in the Composer algorithm 3) Robots now function under a waypoint path planning algorithm that is simple and robust. 4) A grad student was found and hired (8/16/2019)to locate possible places to experiment with that meet the requirements. Aim2: 1) several planter designs have been tested and evlauated. Simple shank, conventional disk, trench, powered tiller wheel have all be tested on different soils for simplicity, ease of use, power consumption, and rock and root capability. The powered till wheel has shown the best performance for our small ag robots. We have dug many furrows as well as planted wheat. We are refining the planter device by increasing the tiller motor and upgrading the seeding mechanism to maintain efficiency during sloped events. Aim4: 1) work has started on a small scale wheat strip header for the agdrone vehicle. Literature review has shown several studies done with larger headers as well as smaller 2 wheel tractor headers.
Publications
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2019
Citation:
1. Das, S., Welch, S., & Flippo, D. (2019). Recurrent neural network based multi-robot route planning for steep-land harvesting systems. In 8th international congress on advanced applied informatics.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2019
Citation:
1. Badgujar, C., Flippo, D.K. & Sharda, A. (2019). Tilt bed testing for autonomous vehicles to get BSFC data, ASABE Annual International Meeting
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2019
Citation:
2. Flippo, D.K., Badgujar, C. & Sharda, A. (2019). Quick release battery for autonomous agricultural vehicles, ASABE Annual International Meeting
|
Progress 01/01/18 to 12/31/18
Outputs Target Audience: The targeted audiences for our efforts have been agricultural industry, agricultural producers, USDA, NIFA, NSF, and several agricultural commissions. Changes/Problems:We lost one Co-PI to an industrial job and replaced him with another KSU faculty. This was approved by NIFA. Since start date was later PhD student was not hired for earliest timing. Student arrived in Fall of 2018. What opportunities for training and professional development has the project provided?Unreal engine four training for the simulation engine needed for the project. How have the results been disseminated to communities of interest?2 publications this year as well as a poster session What do you plan to do during the next reporting period to accomplish the goals?Aim 1: Continue with data collection on slope testing with the robot. Use this data to facilitate simulation models Aim 2:Plant wheat in test field this summer Aim 3:none Aim 4:Test initial Composer algorithm to optimize robot routing
Impacts What was accomplished under these goals?
Aim 1: 1) Tilt bed was designed and fabricated to measure data for robots on slopes 2) Traction tests were done at different speeds on solid concrete to getbaseline. 3) Initial traction tests were conducted in Tilt bed to validate mobility tests. 4) Optimization algorithms were begun to govern route planning. 5) GPS RTK system tested on parking lot. System communication system incorporates the RTK error corrections to facilitate just one 900mghz radio signal. 6) Prototype battery and receptacle were designed and fabricated. 7) Batteries were load tested and proved to operate under load for more than 3 hours each which will give the robots 6 hrs of battery life. 8) New battery updates allow smarter batteries that are self diagnostic and protective. Batteries are now easily charged from a simple power supply which allows for easy, simple, fast, and safe charging. 9) Receptical tested and passed vibration test. Aim 2: 1) Prototype 001 Planter was designed and fabricated and tested for best flute design. 2) Several planter methodologies were tested and evaluated. Shank, Disk, powered trench, and powered oscillation. From the tests it looks like shank drill will work the best and be the simplest to accomodate for small autonomous vehicles. PhD student was hired and arrived in Fall of 2018.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2018
Citation:
Schwindt, J., Flippo, D., & Sharda, A. (2018, May). Developing a wireless communication system for the KSU biological and agricultural engineering ag drone robot. International Journal of Engineering Inventions e-ISSN: 2278-7461, 7(5), 18-24.
- Type:
Journal Articles
Status:
Published
Year Published:
2018
Citation:
Scheer, J., Flippo, D., & Sharda, A. (2018). Wheat drill for a small autonomous vehicle. American Journal of Mechanical Engineering and Automation, 5(1), 9-14.
|
|