Open Access Kent State (OAKS)

The objective of the proposed research is to investigate, design and demonstrate a robotic lawn mower that uses GNSS and electronic compass as sensors for autopilot. The autopilot will allow the mower to navigate along a predetermined trajectory between waypoints, replacing thus the uneconomical random motion of the prior-art robotic mowers. The autopilot / controller will be based on ArduPilot platform. The software upload, the configuration of the controller, the calibration of the sensors and the definition of the autopilot trajectory will be achieved using the freeware graphical user interface (GUI) Mission Planner. The trajectory of the mower will be defined in Mission Planner using waypoints on a satellite map of the lawn.

The novelty of the proposed technology is the open-source approach, which achieves low cost and accessibility. It is also possible, that this may be the first demonstration of a robotic lawn mower that uses GNSS autopilot with predetermined programed trajectory. The benefit for this technology is represented by its economical procedure of achieving its tasks.

In the NASA Robotic Mining Competition, the goal is to design a robotic system that is capable of completing challenging simulations of tasks an autonomous Martian system would encounter. One of the largest obstacles when designing any autonomous robotic systems is localization and orientation. While many systems for localization have been designed and implemented such as GPS and ultrasonic ranging, the conditions of the Martian surface exclude many such systems and demand a minimalistic system. By utilizing strobing sheets of light and precise timing, it is possible to calculate the position of the robot as a function of time intervals in order to achieve high definition localization and accurate positioning, while being viable for use on the Martian surface.

In this research, we are pursuing the robustness of a self-excited vibrational system with negative damping. In practice, this is manifested as conductor galloping of overhead power lines, which is an oscillation of the lines caused by strong winds. Improved transmission tower designs are needed which are capable of combating excessive stresses exerted on the tower by the galloping power lines. Our model of this self-excited system shows that the oscillations can be controlled by adding a boundary velocity feedback controller at the transmission tower. Using a decomposition method, we show there is a closed form analytical solution which predicts the system can be stabilized under certain conditions. Through this research, power transmission systems that are more reliable and resistant to galloping can be engineered.

A lightweight, flow-insensitive pointer analysis tool called srcPtr is presented. The research investigates various methods for pointer analysis to determine which are scalable to large scale software systems. Pointers are variable that indirectly reference other variables within a source code program. Pointer analysis is a non-trivial activity for determining what variables are being pointed by which pointer and how these indirect references can alter program behavior. Pointer analysis is critical for compiler optimization, impact analysis, fact extraction, debugging, and program comprehension. A prototype tool is being constructed as a test bed for determining what analysis approaches are scalable while simultaneously providing a good approximation (full pointer analysis is NP-hard). The implementation, called srcPtr, is built on top of the srcML. srcML is an infrastructure that supports the analysis and exploration of source code and was developed here at Kent State. The research involves the implementation of various published algorithmic aspects of pointer analysis and determine, via experimentation, which can be accomplished within the parameters of scalability. The approach will be evaluated against Codesurfer a commercial program analysis tool. It is expected that srcPtr will be less accurate than Codesurfer but have drastically faster performance. Once completed, srcPointer, will be one of the only stand alone open-source pointer-analysis tool available. It will facilitate additional research that requires efficient pointer analysis currently not available. Additionally, industry professionals can use srcPtr to both obtain valuable data about their code and decide whether a more accurate, heavy-weight solution is appropriate and cost effective.