INTEGRATED SYSTEMS RESEARCH AND DEVELOPMENT IN AUTOMATION AND SENSORS FOR SUSTAINABILITY OF SPECIALTY CROPS

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: HATCH
• Reporting Frequency: Annual
• Accession No.: 1001069
• Multistate No.: W-2009
• Project Start Date: Oct 26, 2013
• Project End Date: Sep 30, 2018
• Program Code: [(N/A)]- (N/A)

Recipient Organization
UNIVERSITY OF CALIFORNIA, DAVIS
410 MRAK HALL
DAVIS,CA 95616-8671

Performing Department
Biological and Agricultural Engineering

Non Technical Summary
The Need: The continuing trend of declining available labor, especially the skilled labor, combined with an increasing consumer desire for a safe and high quality food supply, the pressure of global competition, and the need to minimize the environmental footprint, represents challenges for specialty crop sustainability in the US. Producers and processors are urgently seeking new devices and systems, which will aid them during production, harvesting, sorting, storing, processing, packaging, marketing, and transportation while also minimizing input costs. Currently, there is a lack of effective and efficient sensors and automation systems for specialty crops (fruits, vegetables, tree nuts, dried fruits and nursery). This is because many of the underlying biological processes related to quality and condition of fruits and vegetables are difficult to translate into engineering concepts. Biological variability, coupled with the variable environmental factors, makes it difficult to develop sensors and automation systems for effective implementation at various stages of the production, harvest and postharvest handling chain. Additionally, obtaining measurement of biological factors internal to the commodity is difficult using external, nondestructive sensors, as such devices or processes used must adapt to a wide variation in shape, size, and maturity of the commodity being processed. Much of the traditional methods for growing specialty crops, such as orchards, are based on conventional, labor-intensive production practices. Although slowly changing, crop architectures, systems, and growing practices must adapt to be able to accommodate mechanized production, rather than the other way around. It is a challenge for any single specialty crop sector to afford the cost of research, development, and commercialization of this complex level of automation due to its relative small scale in contrast to grain crop production. It is thus important for public agency entities to assist this economically vital agricultural sector with sensor and automation research and development using an integrated approach. Importance of the Work: The steady increase in global competition and the recent decrease in available labor, especially skilled labor, have increased the need for new technologies. The specialty crop industry in the United States faces significant challenges to remain competitive, and production efficiency is critical to keeping the specialty crop industry thriving. A system-wide approach to developing automation for the specialty crop industry is critically needed to address economic and environmental sustainability challenges. The proposed five-year project will address research and outreach needs for the specialty crop industry in these areas, working with researchers, manufacturers, and producers. Approval of this project will allow researchers and industrial partners to share their results and plan future efforts in a coordinated effort, which will help reduce redundancies and help promote better use of expertise. Objectives Study interactions between machinery and crop to provide basis for creating optimal mechanical and/or automated solutions for specialty crop production Design and evaluate automation systems which incorporate varying degrees of mechanization and sensors to assist specialty crop industries with labor, management decisions, and reduction of production costs. Methods Multiple Crop and Cross Platform Integration: A key and continuing theme in this project is a two dimensional integration of research activities; integration of biological concepts and quantifiable parameters that can be sensed; and integration within each objective among different specialty crops. Although the objectives seem sequential in execution, we anticipate a swirl of concurrent activities centered about and driven by the needs of specialty crop growers for automation of growing, harvesting and postharvest operations. A major task for the project leadership is to facilitate communication and collaboration among the members of this project, and between project members and other stakeholders in specialty crop agriculture.

Animal Health Component

70%

Research Effort Categories

Basic

(N/A)

Applied

70%

Developmental

30%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
402	1119	2020	50%
404	1119	2020	50%

Knowledge Area
404 - Instrumentation and Control Systems; 402 - Engineering Systems and Equipment;

Subject Of Investigation
1119 - Deciduous tree fruits, general/other;

Field Of Science
2020 - Engineering;

Keywords

specialty crops

sensors

control

automation

Goals / Objectives
Study interactions between machinery and crop to provide basis for creating optimal mechanical and/or automated solutions for specialty crop production Design and evaluate automation systems which incorporate varying degrees of mechanization and sensors to assist specialty crop industries with labor, management decisions, and reduction of production costs

Project Methods
Objective: Study interactions between machinery and crops. Many of the interactions between the crop and machinery (such as pruning, thinning, harvesting, sorting, storage, and transportation) will involve physical contact. Most of these interactions must be nondestructive to the plant. Knowing the maximum forces, which may be applied without damage, will require sensing the mechanical properties of the plant. Research and design is required to develop sensors, which can nondestructively, rapidly and accurately determine the inherent strength of a plant tissue. Mechanical or automated harvesting of fruit crops is a critical issue in specialty crop production. Study of how mechanical impact and vibration energies are transmitted through canopies in different types of crop architectures will provide valuable information for developing efficient and effective interfaces for existing mechanical harvesters and to develop new technologies for mechanical and automated harvesting. Study of accessibility of machines to fruits, flowers and branches in different types fruit trees and bushes will be beneficial for improving or developing new technologies for not only harvesting, but also for pruning and thinning. At the same time, study of different types of crop architectures is essential to improve these systems for better accessibility to machines so that practically usable automation and mechanization solutions can be developed. Study of how chemicals sprayed by fixed and mobile spraying systems move in various types of crop canopies are essential to improve application technologies for better coverage and reduced drift and off-site movement. Multi-disciplinary work to simultaneously improve machines and crop architectures will be a key for the success in these areas. Objective: Design and evaluate automation systems A great deal of knowledge and equipment from the industrial and military use of automated equipment can be applied to specialty crops, but the wide variation and semi-chaotic nature of field operations provide a significant number of research challenges. A key challenge is adaptability to multiple crops and cropping systems. The capital investment of automated equipment is high, and the potential market is small. It is critical to spread research and development costs by developing automated systems, which can easily adapt to different crops and different cropping systems. An automated system for specialty crops is much more than an autonomous robot moving along the row of trees or plants. A cost effective system will require information from many sources both on-site and off-site, and autonomous robots or vehicles will require direction and coordination. Hence the efficient gathering of data and control of devices will be an integral part of an automated system. New autonomous orchard vehicles have been extensively tested and proven in the field. Further work can investigate integration of tasks with these vehicles and platforms, as well as fully automating certain operations, which have not yet been successfully addressed, such as autonomous fresh apple harvesting. It is not feasible to replace human labor with automation in all operations. In some operations the cost of replacing human dexterity and complex decision-making capability with equipment is not justified. In these operations, a semi-automated device assisting human labor may be a more optimum solution. Developing a man-machine system requires careful attention to the ergonomic (safety, productivity, comfort and intellectual engagement) needs of the human in the system. Very little research has been done in ergonomic design of specialty crop equipment, and additional research is required to design an optimum man-machine system for specialty crops. Decisions regarding the overall plan and control of an automated system will remain with a human operator. The quantity of operator decision-making information needed for autonomous or semi-autonomous operation of different equipment may potentially be unprecedented and unmanageable. There is a need to develop computer-based data systems which collect, verify, and organize raw data to present information to the operator such as the maturity of the crop, crop stress, and spatial variation. There is also a need to organize predictive models for crop needs such as pesticide application, pruning, thinning and harvesting. Finally, there is also a need to help the operator visualize the most effective use of the automated equipment.

Progress 10/26/13 to 09/30/18

Outputs
Target Audience:The targeted audiences during this project included the academic and scientific communities, growers, extension personnel and the general public. Details of such audiences are provided in the section about Dissemination. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?In the context of this project the following training and professional development took place: Three post-doctoral researchers were mentored. One is currently Assistant Professor at Penn State, the other is Assistant Professor in Brazil, and the third and most recent one is in the market for academic positions. One research engineer (Ph.D.) was mentored. He started his own start-up company on mushroom harvesting. Two PhD students were mentored. One graduated and the other will graduate by March 2019. Seven M.Sc. students were mentored. They all graduated and are working in various positions in industry. Six undergraduate students worked on internships. Five have rgaduated and one is still pursuing her degree. How have the results been disseminated to communities of interest?2014 February 12: Presentation at North Coast Pear Growers meeting, Lakeport, CA. 2014 February 4: Sacramento Western Pear Growers Meeting, Walnut Grove, CA. 2014 March 14: Ag Innovation Day, Growers and PCA community, Salinas, CA. 2015 March 17, Computation of performance metrics for mechanized harvesting using fruit tree models, Annual Meeting of UCCE Precision Ag Workgroup . ANR Building, Davis, CA. 2015 March 18, Automated Agriculture: Focus on Harvesting, CITRIS Spring Seminars. UC Berkeley and Webcast. 2/16/2016: 2016 Hood River Winter Horticulture Meeting, Oregon. 4/22/16: Seminar to undergraduate and graduate students at UC Merced. 4/28/16: Graduate seminar in Dept. of Mechanical and Aerospace Engineering, at UC Davis. 5/24/2106: Presentation to cling-peach growers hosted by Cling Peach Mechanization Research Fund 2017 February 6, Existing approaches, challenges, and possibilities for mechanical tree fruit and sweet cherry harvesting, San Joaquin County Robert J. Cabral Agricultural Center; THE 2017 DELTA PACKING GROWER DAY. 2017 March 8, Robotics in Ag Tech, UC Merced; 2017 CITRIS Agricultural Technology Fair. INVITED SPEAKER in Panel for Robotics in Ag Tech. 2017 April 4, Addressing Labor Shortages in Agriculture with Robotics & Automation, UC Davis Conference Center. Silicon valley Forum. 2017 April 27, Robotics in Agriculture, World Food Center Precision Ag Seminar Series. 2017 October, 15. From Agriculture to Activism: Digital media across Fields. UC Davis. 2017 October, 2. Agricultiral Robotics. UC Davis. Gowers delegation from Argentina. 2018 January, 8, 18. Mechanized harvsting. Modesto, CA. Cherry and clig peach growers. 2018 January, 11. Strawberry Automation Summit. California Strawberry Commission, Cal Poly at Saint Luis Obispo. 2018 March, 2. "Bio Automation Lab R&D Activities". 2018 April, 23. "Labor Saving Technologies". Innovator Summit, UC Davis. 2018 August, 22. "Bio Automation Lab R&D Activities". Trimble - UC Davis. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Toward Objective 3: Study interactions between machinery and crops A large-volume tree digitization system was designed, built and tested. Experimental data from ten Bartlett pear trees in a commercial orchard were acquired and the corresponding tree geometries and topologies were successfully encoded with RMS accuracy better than 0.2 cm. Data from digitized pear trees were used to assess by simulation the reachability of fruit by linear telescopic arms moving horizontally. It was estimated that approximately 70% of fruits were accessible this way, neglecting thin twigs and branches that might stand in the way between the arms and fruits. As another example of model-based analysis for harvesting equipment, it was assumed that after shaking trees with an orchard shaker, fruits dropped vertically down from their locations. The fruits were projected as cylinders to the ground and geometric collisions between each fruit and all branches were calculated. The collision information for four trees (fruit positions were also digitized) was calculated and approximately 58% of fruits would hit a branch before reaching the ground. In general, higher hitting rates for fruits at the upper zones would be expected since they travel longer distances until they reach the ground. However, in these particular trees a lot of the fruit were on branch segments that extended away from the lower branches. These results indicate that inserting catching layers inside the tree canopy could reduce the hitting rates. This approach will be pursued in the coming year. Simulation estimates of linear fruit reachability (LFR) for high-density, trellised pear and cling peach trees were computed. Individual fruit LFR is defined as a Boolean variable that is zero if the fruit's geometric projection along an approach direction vector results in collision with a branch; otherwise, it is one; linear only motion is considered for reaching fruits. The simulations used digitized geometric tree models and fruit locations. The calculated LFRs are averaged of individual fruit LFRs and constitute upper-bound estimates of the true LFRs, i.e., they are optimistic, because the tree models included only branches thicker than approximately 2.5 cm. Multiple 'harvesting passes' were simulated for each type of trees, in the sense that the kth-pass LFR of the fruits remaining on the trees were calculated after all reachable fruits were removed in the previous k-1 passes. Results showed that 92.2% of the pears and 97% of the cling peaches were linearly reachable after five "harvesting passes" at appropriate, optimal approach angles. The LFRs decreased monotonically - exponentially - as a function of the number of passes, thus indicating a diminishing return after more than three sets of approach angles were implemented. The first-pass LFR of cling peaches was significantly larger than that of pears, which can be attributed to simpler canopy and better fruit positioning. These preliminary results indicate that for some trees of SNAP-type architectures fruit reachability may not require complex and expensive arms with many degrees of freedom. Twenty cling peach and twenty pear trees were digitized (branches and fruits) and 3D models wee created. Simulation experiments were conducted using these 3D models to investigate the picking efficiency and throughput of multi-arm robotic harvesters. Simulation experiments were conducted using 3D models of cling peach and pear trees to investigate the picking efficiency and throughput of robotic harvesters with many cartesian arms, given gripper size and arm extension constraints (Arikapudi & Vougioukas, 2018a, 2018b). Toward Objective 5: Design and evaluate automation systems A path tracking controller was developed for a strawberry harvest-aid robot, which maintains the robot chassis in the center of the furrow and parallel to the furrow walls by minimizing the heading error between the robot's velocity vector and a look ahead point placed at a constant distance ahead of the robot in the center of the furrow. Experimental results indicated that the path tracking controller was able to maintain a lateral error of less than 2 centimeters from the center of a 50 centimeter wide furrow and heading error of less than 3 degrees. A prototype cart was built and instrumented with an Arduino microcontroller, and load cells to measure the weight of the harvested strawberries as the tray filled up with fruits in real-time. As the picker picked and placed the fruits on the cart, load cells measured the mass and an off-the-shelf GPS device installed on the cart recorded the location. The tilting angle of the cart was also monitored by a custom made AHRS fixed to the cart. All the collected data was stored on a micro SD card and then uploaded to a computer after the experiment in order to create a yield map for the plot of strawberry field harvested during the experiment. A relatively inexpensive real-time kinematic (RTK) GPS module was also used to provide ground truth location data. A field experiment was carried out and as a worker picked and placed the fruits in the cart, the weight data was measured and stored in a solid-state (SD) memory card, along with GPS position data. After harvest, a yield map was generated for an approximately 0.12 ha area plot of the field. Load cell measurements showed a mean absolute percentage error of 2.6% compared to the weights recorded by a digital scale. This error was reduced to 1.6% by employing a correction factor to the load cell data. A commercial picking bag was instrumented to measure harvested fruit weight. Two load cells were placed inside an enclosure, which was placed between the bag and its shoulder straps, without hindering picking motions. The load cells measure the forces exerted on the straps by the bag and fruits. All electronics were placed inside the enclosure, and included an Arduino, signal conditioning circuits, an Xbee shield, a GPS, and an SD card. The microcontroller transmits data in real time and saves time-stamped data on the SD card. Data are filtered using a median and a low-pass filter to reduce noise. Dynamic calibration was performed in the lab over the weight range of the bag's capacity (20 kg). Baseballs provided consistent weight and volume and were placed in the bag to provide a staircase ground truth weight signal. Two people carried the bag and moved in a manner analogous to pickers. Results showed a mean error of 0.39 kg, standard deviation 0.42kg, and 95th percentile 1.04kg. Major error sources included bag acceleration and body reaction force. A Linear Mixed Model was developed and trained to predict the picking time in manual strawberry harvesting. The dataset consisted of 161 picking times from 18 workers and was collected in strawberry fields in Salinas, CA. Prediction errors lower than 10% were achieved. A standard fruit-picking bag was retrofitted to measure the weight of manually harvested fruits in real time. Two load cells, electronics, and software running on a microcontroller were integrated in a custom-made enclosure. Calibration based on apple picking resulted in root mean square error that did not exceed 560 grams. A Linear Mixed Model was developed to predict the picking time in manual strawberry harvesting. Prediction errors lower than 10% were achieved (Khosro Anjom et al., 2018a, 2018b). A simulator was developed and calibrated that models crew harvest activities and robot fruit-transporting activities during strawberry harvesting. Robot scheduling algorithms were tested using the simulator (Jang, 2018; Peng et al., 2018; Seyyedhasani et al., 2018). A perception system was developed to detect trees and tree rows for autonomous orchard navigation (Durand-Petiteville et al., 2018).

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Arikapudi R., Vougioukas, S.G., (2018). Estimating the Fruit Picking Throughput of a Telescopic Arm in High Density Trellised Pear Orchards. ASABE Annual International Meeting. Paper Number # 1801684 , Detroit, Michigan.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Arikapudi, R., Vougioukas, S. (2018). A Study on Pick-Cycle-Times of Robotic Multi-Arm Tree Fruit Harvesters. Intl. Conference on Agricultural Engineering (AgEng 2018).
• Type: Journal Articles Status: Published Year Published: 2018 Citation: Durand-Petiteville, A., Le Flecher, E., Cadenat, V., Sentenac, T., Vougioukas, S.G. (2018). Tree detection with low-cost 3D sensors for autonomous navigation in orchards. IEEE Robotics and Automation Letters. 3(4): 3876-3883.
• Type: Theses/Dissertations Status: Published Year Published: 2018 Citation: Jang, W.J., (2018). Investigation on the Harvest-aid Robot Scheduling Problem and the Implementation of Its Simulation Platform. M.Sc. Thesis. University of California, Davis.
• Type: Journal Articles Status: Published Year Published: 2018 Citation: Khosro Anjom, F., Vougioukas, S. G., Slaughter, D.C. (2018). Development of a Linear Mixed Model to Predict the Picking Time in Strawberry Harvesting Processes. Biosystems Engineering. (166): 76-89.
• Type: Theses/Dissertations Status: Published Year Published: 2018 Citation: Khosro Anjom, F. (2018). Predictive Modeling of the Temporal Distribution of Tray-Transport Requests for Robot-Aided Strawberry Harvesting. Ph.D. Dissertation. University of California, Davis.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Seyyedhasani, H., Peng, C., Vougioukas, S.G., (2018). Efficient Dispatching of a Team of Harvest-aid Robots to Reduce Waiting Time for Human Pickers. ASABE Annual International Meeting. Paper Number # 1801715, Detroit, Michigan.

Progress 10/01/16 to 09/30/17

Outputs
Target Audience:The targeted audience included the academic and scientific communities, growers, and extension personnel. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?I mentored two undergraduate students four Ph.D. candidates and three MS.c. students. How have the results been disseminated to communities of interest?2017 February 6, Existing approaches, challenges, and possibilities for mechanical tree fruit and sweet cherry harvesting, San Joaquin County Robert J. Cabral Agricultural Center; THE 2017 DELTA PACKING GROWER DAY. 2017 March 8, Robotics in Ag Tech, UC Merced; 2017 CITRIS Agricultural Technology Fair. INVITED SPEAKER in Panel for Robotics in Ag Tech. 2017 April 4, Addressing Labor Shortages in Agriculture with Robotics & Automation, UC Davis Conference Center. Silicon valley Forum. 2017 April 27, Robotics in Agriculture, World Food Center Precision Ag Seminar Series. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 3: Study interactions between machinery and crops Twenty cling peach and twenty pear trees were digitized (branches and fruits) and 3D models were created. Simulation experiments were conducted using these 3D models to investigate the picking efficiency and throughput of multi-arm robotic harvesters. Important findings relating to the design of multi-arm robotic harvesters are being prepared for publication. Objective 5: Design and evaluate automation systems A Linear Mixed Model was developed and trained to predict the picking time in manual strawberry harvesting. The dataset consisted of 161 picking times from 18 workers and was collected in strawberry fields in Salinas, CA. Prediction errors lower than 10% were achieved. A manuscript was submitted and accepted for publication. A standard fruit-picking bag was retrofitted to measure the weight of manually harvested fruits in real time. Two load cells, electronics, and software running on a microcontroller were integrated in a custom-made enclosure. Calibration based on apple picking resulted in root mean square error that did not exceed 560 grams.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Wei-jiunn, J., Vougioukas, S.G. (2017). Evaluation of dynamic robot dispatching strategies using a strawberry-harvesting simulator. 1. ASABE Annual International Meeting. Paper Number 1701402. Spokane, Washington.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Arikapudi R., Vougioukas, S.G., (2017). Linear Reachability of Fruits in Orchard Trees with Actuator Constraints. ASABE Annual International Meeting. Paper Number 1701421, Spokane, Washington. http://doi:10.13031/aim.201701421
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Fei, Z., Shepard, J., Vougioukas, S. (2017). Instrumented picking bag for measuring fruit weight during harvesting. ASABE Annual International Meeting. Paper Number 1701385. Spokane, Washington. http://doi:10.13031/aim.201701385
• Type: Journal Articles Status: Accepted Year Published: 2018 Citation: Khosro Anjom, F., Vougioukas, S. G., Slaughter, D.C. (2018). Development of a Linear Mixed Model to Predict the Picking Time in Strawberry Harvesting Processes. Biosystems Engineering. (166): 76-89. https://doi.org/10.1016/j.biosystemseng.2017.10.006

Progress 10/01/15 to 09/30/16

Outputs
Target Audience:The targeted audience included the academic and scientific communities, growers, and extension personnel. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Mentored one undergraduate student, Joshua Munic. How have the results been disseminated to communities of interest?2/16/2016: 2016 Hood River Winter Horticulture Meeting, Oregon. 4/22/16: Seminar to undergraduate and graduate students at UC Merced. 4/28/16: Graduate seminar in Dept. of Mechanical and Aerospace Engineering, at UC Davis. 5/24/2106: Presentation to cling-peach growers hosted by Cling Peach Mechanization Research Fund What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 3. Study interactions between machinery and crop to provide basis for creating optimal mechanical and/or automated solutions for specialty crop production. Mechanical or automated harvesting of fruit crops is a critical issue in specialty crop production. Study of accessibility of machines to fruits, flowers and branches in different types fruit trees and bushes will be beneficial for improving or developing new technologies for not only harvesting, but also for pruning and thinning. At the same time, study of different types of crop architectures is essential to improve these systems for better accessibility to machines so that practically usable automation and mechanization solutions can be developed. Prototype robotic fruit harvesters typically utilize multiple degree-of-freedom (DOF) arms. The working hypothesis is that as tree branches constrain fruit reachability, redundancy is necessary to navigate through branches and reach fruits inside the canopy. However, modern commercial orchards increasingly adopt trees of SNAP architectures (Simple, Narrow, Accessible, and Productive). In this period, simulation estimates of linear fruit reachability (LFR) for high-density, trellised pear and cling peach trees were computed. Individual fruit LFR is defined as a Boolean variable that is zero if the fruit's geometric projection along an approach direction vector results in collision with a branch; otherwise, it is one; linear only motion is considered for reaching fruits. The simulations used digitized geometric tree models and fruit locations. The calculated LFRs are averaged of individual fruit LFRs and constitute upper-bound estimates of the true LFRs, i.e., they are optimistic, because the tree models included only branches thicker than approximately 2.5 cm. Multiple 'harvesting passes' were simulated for each type of trees, in the sense that the kth-pass LFR of the fruits remaining on the trees were calculated after all reachable fruits were removed in the previous k-1 passes. Results showed that 92.2% of the pears and 97% of the cling peaches were linearly reachable after five "harvesting passes" at appropriate, optimal approach angles. The LFRs decreased monotonically - exponentially - as a function of the number of passes, thus indicating a diminishing return after more than three sets of approach angles were implemented. The first-pass LFR of cling peaches was significantly larger than that of pears, which can be attributed to simpler canopy and better fruit positioning. These preliminary results indicate that for some trees of SNAP-type architectures fruit reachability may not require complex and expensive arms with many degrees of freedom. Objective 5: Design and evaluate automation systems It is not feasible to replace human labor with automation in all operations. In some operations the cost of replacing human dexterity and complex decision-making capability with equipment is not justified. In these operations, a semi-automated device assisting human labor may be a more optimum solution. A worker's picking rate during tree fruit harvesting can provide useful information for better workforce management; orchard platform crew management; yield maps (in combination with position). A commercial picking bag was instrumented to measure harvested fruit weight. Two load cells were placed inside an enclosure, which was placed between the bag and its shoulder straps, without hindering picking motions. The load cells measure the forces exerted on the straps by the bag and fruits. All electronics were placed inside the enclosure, and included an Arduino, signal conditioning circuits, an Xbee shield, a GPS, and an SD card. The microcontroller transmits data in real time and saves time-stamped data on the SD card. Data are filtered using a median and a low-pass filter to reduce noise. Dynamic calibration was performed in the lab over the weight range of the bag's capacity (20 kg). Baseballs provided consistent weight and volume and were placed in the bag to provide a staircase ground truth weight signal. Two people carried the bag and moved in a manner analogous to pickers. Results showed a mean error of 0.39 kg, standard deviation 0.42kg, and 95th percentile 1.04kg. Major error sources included bag acceleration and body reaction force.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Vougioukas, S.G., Arikapudi, R. and Munic, J., 2016. A Study of Fruit Reachability in Orchard Trees by Linear-Only Motion. IFAC-PapersOnLine, 49(16), pp.277-280.
• Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Vougioukas, S., Arikapudi R., Munic, J. (2016). Upper Bound Estimates of Fruit Reachability in Orchard Trees using Linear Motion. Proceedings of the Intl. Conference on Agricultural Engineering (AgEng  CIGR 2016), Paper 529, Aarhus, Denmark.
• Type: Journal Articles Status: Published Year Published: 2016 Citation: Arikapudi, R., Vougioukas, S.G., Jim�nez- Jim�nez, F., Farangis Khosro Anjom, F. (2016). Estimation of Fruit Locations in Orchard Tree Canopies Using Radio Signal Ranging and Trilateration. Computers and Electronics in Agriculture (125):160-172.
• Type: Journal Articles Status: Under Review Year Published: 2016 Citation: Vougioukas, S.G., He, L., Arikapudi, R. (2016). Orchard Worker Localisation Relative to a Vehicle Using Radio Ranging and Trilateration. Biosystems Engineering (147): 1-16.

Progress 10/01/14 to 09/30/15

Outputs
Target Audience:The targeted audience included the academic and scientific communities, growers, and extension personnel. 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?2015 March 17, Computation of performance metrics for mechanized harvesting using fruit tree models, Annual Meeting of UCCE Precision Ag Workgroup . ANR Building, Davis, CA. 2015 March 18, Automated Agriculture: Focus on Harvesting, CITRIS Spring Seminars. UC Berkeley and Webcast. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 3. Study interactions between machinery and crop to provide basis for creating optimal mechanical and/or automated solutions for specialty crop production. Mechanical or automated harvesting of fruit crops is a critical issue in specialty crop production. Study of accessibility of machines to fruits, flowers and branches in different types fruit trees and bushes will be beneficial for improving or developing new technologies for not only harvesting, but also for pruning and thinning. At the same time, study of different types of crop architectures is essential to improve these systems for better accessibility to machines so that practically usable automation and mechanization solutions can be developed. Objective 5: Design and evaluate automation systems It is not feasible to replace human labor with automation in all operations. In some operations the cost of replacing human dexterity and complex decision-making capability with equipment is not justified. In these operations, a semi-automated device assisting human labor may be a more optimum solution. During this period a prototype cart was built and instrumented with an Arduino microcontroller, and load cells to measure the weight of the harvested strawberries as the tray filled up with fruits in real-time. As the picker picked and placed the fruits on the cart, load cells measured the mass and an off-the-shelf GPS device installed on the cart recorded the location. The tilting angle of the cart was also monitored by a custom made AHRS fixed to the cart. All the collected data was stored on a micro SD card and then uploaded to a computer after the experiment in order to create a yield map for the plot of strawberry field harvested during the experiment. A relatively inexpensive real-time kinematic (RTK) GPS module was also used to provide ground truth location data. A field experiment was carried out and as a worker picked and placed the fruits in the cart, the weight data was measured and stored in a solid-state (SD) memory card, along with GPS position data. After harvest, a yield map was generated for an approximately 0.12 ha area plot of the field. Load cell measurements showed a mean absolute percentage error of 2.6% compared to the weights recorded by a digital scale. This error was reduced to 1.6% by employing a correction factor to the load cell data.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Khosro Anjom, F., Rehal, R., Vougioukas, S. (2015). A Low-Cost, Efficient Strawberry Yield Monitoring System. ASABE Annual Intl. Meeting; Paper Number 152189408, New Orleans, USA.
• Type: Journal Articles Status: Under Review Year Published: 2015 Citation: Vougioukas, S.G., He, L., Arikapudi, R. (2015). Orchard Worker Localisation Relative to a Vehicle Using Radio Ranging and Trilateration. Biosystems Engineering.
• Type: Journal Articles Status: Under Review Year Published: 2015 Citation: Arikapudi, R., Vougioukas, S.G., Jim�nez- Jim�nez, F., Farangis Khosro Anjom, F. (2015). Estimation of Fruit Locations in Orchard Tree Canopies Using Radio Signal Ranging and Trilateration. Computers and Electronics in Agriculture.

Progress 10/26/13 to 09/30/14

Outputs
Target Audience: he targeted audience included the academic and scientific communities, growers, and extension personnel. 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? 2014 February 12: Presentation at North Coast Pear Growers meeting, Lakeport, CA. 2014 February 4: Sacramento Western Pear Growers Meeting, Walnut Grove, CA. 2014 March 14: Ag Innovation Day, Growers and PCA community, Salinas, CA. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 3. Study interactions between machinery and crop to provide basis for creating optimal mechanical and/or automated solutions for specialty crop production. Mechanical or automated harvesting of fruit crops is a critical issue in specialty crop production. Study of accessibility of machines to fruits, flowers and branches in different types fruit trees and bushes will be beneficial for improving or developing new technologies for not only harvesting, but also for pruning and thinning. At the same time, study of different types of crop architectures is essential to improve these systems for better accessibility to machines so that practically usable automation and mechanization solutions can be developed. During this period, a large-volume tree digitization system was designed, built and tested. Experimental data from ten Bartlett pear trees in a commercial orchard were acquired and the corresponding tree geometries and topologies were successfully encoded with RMS accuracy better than 0.2 cm. Objective 5: Design and evaluate automation systems It is not feasible to replace human labor with automation in all operations. In some operations the cost of replacing human dexterity and complex decision-making capability with equipment is not justified. In these operations, a semi-automated device assisting human labor may be a more optimum solution. During this period a path tracking controller was developed for a strawberry harvest-aid robot, which maintains the robot chassis in the center of the furrow and parallel to the furrow walls by minimizing the heading error between the robot's velocity vector and a look ahead point placed at a constant distance ahead of the robot in the center of the furrow. Experimental results indicated that the path tracking controller was able to maintain a lateral error of less than 2 centimeters from the center of a 50 centimeter wide furrow and heading error of less than 3 degrees.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Arikapudi, R., Durand-Petiteville, A., Vougioukas, S. (2014). Model-based assessment of robotic fruit harvesting cycle times. ASABE Annual Intl. Meeting; Paper Number 1913999, Montreal, Quebec, Canada.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: arangis Khosro, A., Rehal, R., Fathallah, F., Wilken, K., Vougioukas, S. (2014). Sensor-based Stooped Work Monitoring in Robot-aided Strawberry Harvesting. ASABE Annual Intl. Meeting; Paper Number 1913911, Montreal, Quebec, Canada.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Wei-jiunn, J., Lewis, G., Hoachuck, J., Slaughter, D., Wilken, K., Vougioukas, S. (2014). Vibration-reducing Path Tracking Control for a Strawberry Transport Robot. ASABE Annual Intl. Meeting; Paper Number 1914011, Montreal, Quebec, Canada.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: He, L., Arikapudi, R., Khosro Anjom, F., Vougioukas, S. (2014). Worker Position Tracking for Safe Navigation of Autonomous Orchard Vehicles Using Active Ranging. ASABE Annual Intl. Meeting; Paper Number 141913710, Montreal, Quebec, Canada.