Summer 2015 Research
In the summer of 2015 i was introduced to and learned about
motion planning and its many applications. My partner at the time
Brandon Martinez
and I were able to test our newly formed algorithm which proved to be up to six
times faster than previous methods. Our algorithm began planning from the origin
and then proceeded to plan in the space we had provided. All the testing was
done through simulations which eliminated the need to use physical robots. The
reason our algorithm worked great was because it had a constant input and
feedback loop that interacted with the user throughout the planning process.
The user was able to guide the robot towards the goal by inputting attract
regions that the robot biased its planning towards. This greatly benefitted the
planning process because we greatly reduced the time taken to solve the problem,
which saved the algorithm its computational resources. By combining a human's
problem solving skills and a computer's speed and precision, we hope that our
algorithm one day be using in unmanned aerial vehicles or unmanned ground
vehicles in disaster situations. There are so many applications to our algorithm,
e.g., computational biology, robotics, satelite docking, disaster relief,
virtual reality, etc. Below I included the concluding paper and poster I made
for the summer of 2015's research. We took the poster to a poster symposium held
at Texas A&M University. We competed against over 60 undergraduate students and
were able to walk out of there with second place! |
Summer 2016 Research
In the summer of 2016 I will be doing research in the Parasol lab once again but
instead of using simulations to test our method, we will be using physical robots.
My two mentors this year will be
Read Sandstrom and
Saurabh Mishra. Read is
well on his way to getting his Ph.D, while Saurabh is about to get his Masters. So far
they have been great mentors and i have learned so much in the few weeks we have
been here. So far, we developed a program for the robot to travel x meters
forward, x degree turn, and x meters backward. Currently my two partners and I,
Brandon Martinez and
Bryan Rodriguez have
been modeling the environment we are going to use in our experiments with markers
throughout the lab. In this summer's project, we will look at modeling the first
stage of a construction problem: moving resources from a supply/drop-off site to
a construction site. We still have a long way to go since we still need to get
the markers working, get the IR sensors working which will detect the virtual wall,
and create a robot swap behavior when robots are running low on battery. We have
been split into three teams. Team one, ArUco team, is in charge of getting the
ArUco recognition working. Team two, Environment modeling team, is responsible
for setting up the simulated environment. Team three, IR team, is responsible
for figuring out the IR sensors. My team is team two, and we are in the processs
of completing the setting up and measuring of the markers throughout the lab. |
Click to go back to my home page