Week 4

This week Brigette and Wargo handed over the schematics to Gardner to begin the PCB layouts. They have began testing the H-Bridge to figure out the extreme conditions required to break components. This week they ran through frequency tests on the PWM and direction signals to assure that all the current components will withstand rigorous switching. Next week Brigette and Wargo plan on continuing testing while developing the formal test document for the H-Bridge.

Although minimal tangible progress was made on Karl's part this past week, he was able to fully install the necessary programs (including python) and libraries to adequately begin developing programs for the beaglebone and minibox. He has been familiarizing himself with the the syntax of python, and has completed the first draft of the odometry code in python.


By the end of this week Tanner and Jason have accomplished installing the back caster wheels, fabricating and mounting main drive axle L-brackets, and mounting intermediate U-channels. They are still waiting on Patrick to finish machining the rest of the intermediate section. Next week, they hope to have battery brackets mounted and all of the drive system complete excluding the chain.


Harris and Gabe have enabled their opencv code in python to register video processing and analysis properties. They have developed a code that uses Hough Transformation lines to determine straight lines within registered images and are currently refining their code to filter the hough transform lines so that they may give us the lines of the side walk edges. Their edge detection code is currently being refined so that it can be used to localize the position of the camera with respect to the initial starting point of the robot. For this they will come up with a filtering algorithm to get rid of unnecessary lines of the hough output. This will enable them to localize the hough outputs. This will be started on Monday.

Week 3

This week, Brigette and Trent worked on schematics for the electronic system. The schematics were finalized and confirmed by Dr. Isenberg and Dr. Davis will be reviewing them next week. When all the schematics are approved, construction on the PCB designs will begin. The parts for the prototype H-Bridge have arrived allowing Brigette and Trent to replace existing components with new components. Unfortunately the new MOSFETs are not handling the voltage and current spikes that the H-Bridge is producing when connected to the large motor so, next week the pair of them will be looking for a solution to this problem and potentially ordering new parts.
Brigette updated the Gantt Chart with everyone's current progress and will continue to update the Gantt chart to attempt to keep the project on track.

This week the vision team (consisting of Gabe and Harris) has been able to successfully perform an edge detection on a single still image and is near to completing a Hough analysis of the edges. This will allow the team to identify the edges of a sidewalk from which to judge the robot’s position.

The team has begun programming on robot’s mini-box computer which will eliminate some integration time in the future. However, another computer, running Linux, has been set up to run Open CV so development can occur even more quickly, by working simultaneously on both machines.Next week the team will complete analysis of still images and will begin to work with live stream input from the webcam. Planning has also begun on the correct method of localization using the information that will be provided by the Hough Analysis of the robot camera stream.

This week Jaya updated and completed the basic odometry program. Next week he will be implementing a control law titled “back-stepping” into the program which will allow the integration of a kinematic controller and a torque controller using a dynamical extension. The major problem of using this method is that Jaya will have to teach himself the material, which can lead to misunderstandings and it may take some time to fully grasp the context. One of the major benefits of using this method is that the odometry algorithm will be able to travel a more accurate path to find its destination points.

According to the gantt chart of assigned tasks, Karl was scheduled to assist Jaya in developing the computer torque controller. The calculations and pseudocode behind the basic controller were already developed in the Robotics II class last semester. With this said, Karl and Jaya decided that (knowledge of working in Python aside), Jaya had the material he needed to begin development of the controller on his own. Karl then decided to begin the task of developing a simulated mobile robot to demonstrate the dynamics and functionality of the controller on the basic system before it is implemented in the real robot. Using a template designed by Dr. Isenberg, Karl completed a simulated animation in MATLAB, able to demonstrate a controller designed in MATLAB. In the upcoming week, Karl will now attempt to develop the same/similar model in Python to work with the Python-based controller.

By the end of this week Jason and Tanner accomplished the following: mounted the motors to the platform base, provided Patrick with all the parts to be fabricated (all but one of the ordered parts has been delivered), and cut the L-brackets. Now the L-brackets just need to be drilled before mounting them to the robot base. Currently the pair are waiting for Patrick to finish the parts they gave him, so they can finish the drive system on the robot. Next week, Tanner and Jason hope to have all parts back from Patrick, and finish fabrication of the L-brackets for the wheel axels finished. With these parts completed, they can finish installing the drive system on to the robot base except for welding the axel into place. In addition, a caster wheel will be chosen, bought, and installed to the base.

Week 2

As week two has come to a close, the Riddle Rascals have made significant progress on their project. To begin, Jason Chong and Tanner Reznicek made great progress in finalizing the CATIA model and ordering parts that are required to complete the fabrication of the platform base. Next week they hope that the parts that they ordered come in so that they can continue to fabricate the platform base for parts to be mounted to.

Next, Karl Pruhsmeier has greatly assisted in the team’s part ordering process by creating a standardized order form that is used to submit part orders to Jim Weber. Karl has also been spending time assisting Jaya Nwobu in developing and completing pseudo code for the odometry software. Karl then began on developing a simulated model of the robot in MATLAB, which will be used to help visualize the control programs being developed for the physical robot.

Trent Gardner has been working to write code and develop the parts needed for the motor decoders.

The team of Haris Tyeb and Gabe Bentz, is working on Open CV and has one computer running Open CV on Windows and has begun familiarization and testing with the Open CV software. Currently, the team is able to obtain a live stream of video from the webcam to be used on the project. Code is currently being researched and developed to perform basic edge detection on the stream. This will lead to the capability to distinguish the edges of sidewalks and gain the location of the robot. The team is also working on integrating Open CV onto a Linux computer. This will allow them to familiarize themselves with how the code and software will operate on the MiniBox computer that will be placed on the robot platform, which using a Linux operating system. Next week the Open CV team intends to have completed code for edge analysis of a video stream. From there they will begin using the information obtained to develop localization techniques in order to give the robot information on its location based on edge analysis of sidewalks. This progress will be dependent on the amount of open source code available and whether additional software will have to be installed and implemented to allow for the increased functionality of Open CV.

Brigette Cochran has completed work on the Gantt chart used to schedule events for the semester. This chart will be extremely useful for keeping the team on time for task completions. Brigette has also been working with Trent Wargo to develop schematics for the electronic system in order to determine proper connections with all electronic circuits. The schematics will also be used to develop the PCBs in the future. Finally, Brigette and Trent Have completed a working prototype of an H-Bridge. They have ordered the final parts needed to complete testing and should soon start on PCBs for both H-bridges that are needed.

Week 1

Having a clear schedule will keep the Riddle Rascals on track to complete the necessary stages of this capstone project. This week the Riddle Rascals completed the Gantt Chart. The goal was to schedule, in detail, all the tasks necessary to ingrate the IMU, GPS, and computer vision onto a visually appealing autonomous robot. The Gantt Chart has extra time allocated to keep the team on track if any particular stage falls behind. Also, the Gantt Chart delegates who will be working on what parts of the project to assure that each team member is allocated a fair amount of work over the course of the semester.

Welcome

Welcome to the official website of Tour PNADDER, the 2015 Robotics Detail project at the Prescott Campus of Embry-Riddle.

Check here to get updates on project progress.