DPRG Hackathon Team Advances

DPRG’s own Hackathon team of Jain Shi, David Ackley, Omar Barlas, and Carl Ott have successfully advanced to the next tier of the AT&T IoT Hackathon. They are one of four teams in the nation to be invited to compete in the Mechatronics challenge at the AT&T Developers Conference in Las Vegas held in early January. With their selection, AT&T has provided help with airfare and travel expenses for two members of the team. The whole team plans to go to Las Vegas and they are sharing the cost. If they place in the top three teams, they would win a nice cash prize. Wish them luck!

The team will be discussing the Hackathon experience and their project at the December 10th, 2016 DPRG monthly meeting.

The featured picture is missing team member Carl Ott who is pictured below.

screenhunter_149-dec-07-15-59
                                   Carl Ott

 

DPRG Hackathon Team – Wins AT&T Mechatronics prize

It started with an email from an AT&T Dallas Hackathon organizer inviting DPRG to participate in their November event. A group of DPRG members quickly joined together to form a team. The Hackathon lasted an evening (Nov 18th) and a day (Nov 19th), with the option to work through the night. The team developed their project idea fairly quickly, but then wasted the first evening chasing false paths of hardware choices  and software packages. The next day with clearer minds, they were able to develop a prototype that they presented at the end of the event.

The attending crowd and judges received the presented prototype with interest and enthusiasm. A little later the team found that they had won the Mechatronics prize of $2500 and become semifinalists for the AT&T Invitational at their annual Developer Conference in January. The team is busily developing their idea and preparing a video for submission to this next tier of competition. Wish them luck!

dprg_hackathon_project-2

DPRG AT&T Hackathon Team

A good article about the AT&T Hackathon event including the DPRG team was written by David Moore of dallasinnovates.com .
The link is Dallas Hackathon Highlights Local Coders’ Creativity, Problem-Solving Skills .

RoboColumbus Plus 2016 Results

Avondale Park in Denton, TX provided an excellent venue for RoboColumbus Plus 2016 held on November 19th. The weather was a brisk clear Autumn day. Contestants and spectators all had a great time watching the robots make their attempts at the course. This year’s course was the most challenging course in RoboColumbus history.

robocolumbus_nov_2016_home_to_1st_cone_x
Image of home cone showing first target cone in the distance

RoboColumbus is an outdoor robot competition that was patterned after the RoboMagellan competition run by Seattle Robotics Society. RoboColumbus was always meant to be a training ground for DPRG members to improve their outdoor rover knowledge and be simpler than RoboMagellan. It started out as a course with one cone that was positioned about 50 feet from the starting line in a clear path. Participating robots drove autonomously from the starting line to the cone. The robots received points if they touched the cone and stopped.

In 2014, the distance between the starting cone and the target cone was increased to 100 yards. The path between the starting line and the target cone continued to have no obstacles. However, the course could have high grass, holes, and other terrain irregularities. The robots could score points for: coming close to the cone and not stopping (1 pt), being close to the cone and stopping (1 pt), touching the cone and stopping (1 pt).  A perfect score was 3 points.

In 2015 the name was changed to RoboColumbus Plus and  the course was expanded by an extra cone that had no clear path from the first target or home cones. The distance between each of the 3 cones (target 1- clear path, target 2 – obstacles in path, and home cone – obstacles in path) was kept at 100 yards. Scoring was similar to what was done in 2014. A prefect score was 9 points. Time of run determined winners when multiple robots achieved 9 points. This new contest and scoring method has worked well for both beginning and seasoned contestants. It allows beginners to see progress, yet challenges seasoned roboticists.

This year three roboticists participated. Two robots, jBot and B.U.R.P., scored perfect scores of 9 points. The remaining robot, Rover 3, came close to all three cones to score 3 points. B.U.R.P, which participated in the 2015 RoboColumbus Plus sported major improvements this year. Besides major software enhancements, it had a completely new obstacle avoidance sensor array with motorized pitch adjustment, new encoders, new IMU, and also added a camera and GPS. The winning robot, jBot, has achieved perfect scores for 2 years running.

robocolubus_nov_2016_winnersDoug Paradis (rover3 – 3rd),   David Anderson (jBot – 1st),   Scott Gibson (B.U.R.P. – 2nd)

 

The robot runs can be seen on youTube at RoboColumbus Plus 2016

 

 

Fall Indoor Competition – 6 Can contest Results

This year’s Fall indoor competition consisted of a single contest, the perennial favorite Six Can.  This contest was designed to be a training event for the Roborama Can-Can Soccer event held in May. It is similar to Can-Can Soccer without the opposing robot. A competing robot must collect 6 cans from the Can-Can Soccer arena and deposit them at the opposing goal.

This contest has seen a large variety of approaches to the problem over its lifetime. This year was not different. The winning robot, Can Man, used a spinning LiDAR (RoboPeak) and camera to determine location and target cans. The second place robot, VEX navigator, which used the VEX IQ modular system, determined it location with odometry. This robot used sonar and a Pixy (CMUCAM5) vision sensor to target cans. The third place robot, Rainman, used a completely targeting methodology and didn’t track robot location. It sported a LiDAR lite (v2), Pixy, and a goal trip switch to handle the tasks of capturing cans and depositing them at the goal.

fall_2016-6_can_winners

DPRG President Steve Edwards handing out prizes to the winners of Six Can.
Scott Gibson (1st),   Doug Paradis (2nd),  Ray Casler(3rd)

 

The robot runs can be seen on youTube at Fall Competition 2016 – Six Can .