• Developed machine vision/learning algorithms (Python) to detect features from RGB and hyperspectral images.• Developed algorithms (ROS & Gazebo) for customized control and path planning of wheeled robots, Cartesianrobots, and articulated robotic arms. Improved inverse kinematics and hardware interface of the existing methods.• Designed an entirely new mobile Cartesian robot for in-field plant scanning with high coverage (LabVIEW &FPGA). The system had customized controllers, circuits, and mounting solutions(Paper 1).• Developed a 6-joint articulated robot system for indoor plant scanning. (Paper pending)• Designed and manufactured customized 3D printed & machined parts for robots.• Utilized deep learning with a machine vision algorithm to detect soybean leaf poses in the field with a test accuracyof 89%. (Paper pending)• Designing and developing a heavy-payload drone robotic system for carrying the articulated robot for in-fieldautomation applications. (Paper pending)

• Composed lecture, lab, and exam contents.• Held lab sessions with 30 students to teach circuits & robotics related experiments.• Composed changes in the middle of the semester for switching to online teaching.

• Designed and implemented ergonomic improvements into manufacturing workstations.• Led streamlining and cost reduction meetings.

• Combined triangulation and image processing to build a control system on LabVIEW and Arduino for a robot to locate target objects and determine distance.• Calibrated long distance measurement using the RSSI signal from Xbee.• Designed and assembled camera sets and the body frame of the robot.

• Developed image processing logarithm to extract physiological data of plants to examine their health condition.• Implemented image processing logarithm into iOS and Android Apps using Swift and Java, optimized the Apps’ memory allocation by 60% and designed icons, logos and user interfaces.

• Revised midterm exam questions related to Matlab and data analysis.• Guided 25 students through Matlab projects on noise reduction using frequency domain transformation.