In the third quarter of this project, we summarized the results of our previous experiments that evaluated the influence of several limiting factors on the performance of a ground penetrating radar (GPR) to accurately detect HLB-affected citrus roots and determine their main structural characteristics. We prepared a manuscript that includes most of our experiments/results, and submitted it for peer-review to the Agronomy Journal, Special Issue of Precision Agriculture. The manuscript was accepted and published:Zhang X., Derival M., Albrecht U., and Ampatzidis Y., 2019. Evaluation of a Ground Penetrating Radar to Map Root Architecture of HLB-infected Citrus Trees. Agronomy (Special Issue: Precision Agr.), 9(7), 354. https://doi.org/10.3390/agronomy9070354. Received: 3 May 2019 / Revised: 23 June 2019 / Accepted: 1 July 2019 / Published: 3 July 2019.We have acknowledged CRDF:Funding: This research was funded by the University of Florida Citrus Initiative and the Citrus Research and Development Foundation. As explained in the previous reports (and based on our preliminary experiments), in order to collect accurate data from the GPR, a layout of the scans has to be prepared. This layout involves three concentric circles with a distance of 1 foot from each other with the trunk as the center. These concentric circles are drawn manually on the ground after measuring the distance from the trunk using a measuring tape and making marks. Once the circles are drawn around the tree, the scanner is moved along the circles manually by an operator moving under the trees. This process can be sometimes challenging as there may not be enough space to perform the scanning. As this process is done manually, the circles may not be perfect and there is a chance for errors to show up in the results due to irregular marking and drawing. In order to eliminate these problems and irregularities and also as an attempt to ease the scanning process we plan to automate the whole marking and scanning process.The idea is to put the scanner in an enclosed unit that is capable of being controlled remotely using the user. This enclosure is connected to the tree using an adjustable length bar to allow radial movement of the scanner. The adjustable length bar can have lengths varying from 1 to 3 feet. To achieve this, we built an aluminum chassis with wheels in a way that the scanner is touching the ground at all the time. However, with this design, we noticed some issues, especially with the wheels when rotating was stuck in the soil. This could be solved by increasing the size of the wheels but, since the scanner has to be in contact with the ground all the time, we did not proceed with this approach. Therefore, our next idea was to build an agile tracked chassis which can have a much better movement in soil and rough terrains. The track is connected to two drive wheels on either side of the chassis, which are in turn connected to motors controlled by a microcontroller. The motors are connected to a Cytron SmartDriveDuo Smart Dual Channel 10A Motor Driver, which allow the user to control the motors remotely using a remote controller. A RadioLink T8FB 2.4GHz 8CH Transmitter w/ R8EH 8CH Receiver was used for this purpose. Both the motors are connected to Motor Driver circuit and each of them can be controlled independently by the remote controller, which allows the chassis to make turns and perform radial movements. With this chassis, the scanner moves well on the ground even when there are irregularities present. In the next quarter, we will attach the adjustable length bars to the aluminum chassis and recognize the best way to attach them to the tree. The challenging aspect here is the varying size of the trunks, so the tree holder has to have a capability to adjust itself to hold on to varying trunk sizes and also at the same time allow the chassis to rotate around itself. We are currently working on this issue and have found certain methods to solve this problem to develop an automated system.