The Paul G Allen School of Computer Science & Engineering - Remembering Paul Allen

Paul G. Allen School of Computer Science & Engineering-A powerhouse in robotics & AI at Washington state

 

Paul Allen

The founders of Microsoft, Paul Allen and Bill Gates have deep history with the University of Washington. They used to use the computer lab at the University of Washington to do research but one day got banned for abusing it. However, as we can see today, this marvelous institution and engineering school is now named after Paul Allen for his contributions to many fields including aerospace, real estate, sports and arts.

Paul Allen along with aerospace engineer Burt Rutan is behind the creation of Stratolaunch Systems which built the worlds largest twin-fuselage Stratolaunch Jet.

 

Stratolaunch

Paul Allen also owned the Seattle Seahawks which under him, went on to win the Superbowl XLVIII in 2014. 

Paul Allen Seahawks

Back to this amazing school.

The Paul G. Allen School of Computer Science & Engineering at the University of Washington is one of the top-notch schools in the US producing important research and superstars in robotics.

Today we will talk about the few robotics & automation labs at this impressive institution and some of the work that have been produced there.

Human Centered-Robotics Lab

The Human-Centered Robotics lab activities are based on robotics development that would assist people in everyday tasks.

The laboratory applies human-centered approach for problems resolving, with an idea to bring users (with no technical knowledge) exceptional robotics functionality. Some notable representative lab accomplishments are: project EMAR - as an interdisciplinary effort that investigates effective measurement of adolescent stress levels by means of a social robot.

EMAR

EMAR

Furthermore, professor Maya Cakmak and postdoctoral researcher Maru Cabrera attended robotics games and conference RoboCup 2019 and even won the first even Vising Fellows Award at that tournament.

In addition, this lab was responsible for some workshops organization such as “Designing a Social Robot for Mental Health Support”.

Movement Control Lab

The Movement Control lab researches biological movements, their modeling and control. With support of MuJoCo physics engine as the key tool, the smart people in the lab focuses on effective methods development and their applicability to more complex problems.

Note: There is a very interesting video at the link above. The virtual objects in the video move when a virtual force is acted on them - a virtual engine powered by MujoCo.

Dr. Emo Todorov is currently developing this engine for Roboti LLC and MujoCo also used by a wide community of researchers and developers.

Some of the most important laboratory results are include modeling and simulation of multi-joint dynamics with contact – a project that combines

1. Recursive algorithms in generalized coordinates and velocity-stepping using contact impulse solvers.

2. Algorithms and software development for Model-predictive control with full dynamics in real time, contact-invariant optimization (CIO) as a new approach to optimal control

3. Simultaneous state estimation and system identification problems resolving - by jointly optimizing the movement trajectory and the model parameters.

The Movement Control lab has significant results in the design of robotic hands which have become faster and almost as compliant as human hands.

Also some notable promotions of the very impressive people that worked in the Movement Control Lab includes Vikash Kumar who now works at OpenAI and Yuval Tassa who is now a research scientist at DeepMind.

Speaking OpenAI, lets see a robot hand solve a Rubiks Cube.

 

Neural Systems Lab

The research at the Neural Systems Lab is concentrated on building intelligent robotic systems and brain-computer interfaces (BCIs).

A BCI basically uses input from the brain to produce certain types of responses. Imagine being able to think your essay and letting the computer type the essay for you. Imagine using your brain to make movements in a video game or control a robot arm.

The Neural Systems lab’s projects and publications are focused on probabilistic information processing and building biologically-inspired robots that can learn through experience and imitation. We have no doubt that research from this lab will be used in crucial technology that will be coming in the near future.

Personal Robotics Lab

Professor Siddhartha Srinivasa, who is Boeing endowned founded the Personal Robotics Lab in 2006.

The mission of the Personal Robotics Lab is to develop the fundamental building blocks of perception, manipulation, learning and human-robot interaction in order to provide capability for robots in performing complex physical manipulation tasks around people.

Some work from professor Srinivasa and the Personal Robotics Lab include..

HERB

HERB Robot

 

MuSHR

MuSHR

MuSHR is a low-cost open-source platform to research AI and robotics on a mobile/racing car environment. This platform can becruical to prepare the next generation of engineers by combining a toy element in robotics learning and research.

CHIMP CMU

Needs no introduction, this chimplike, morphing robot is also product of Professor Srinivasa and Carnegie Mellon University. CHIMP  (Mobile Intelligent Robot) loves to rescue people and complete complicated tasks in degraded and dangerous environments.

 

Robotics and State Estimation Lab

The Robotics and State Estimation Lab was established in 2001. It is focused on problems in robotics with object recognition and tracking, human robot interaction and human activity recognition.

See video below of regression networks with visual tracking of random objects. This advanced system can track a number of random objects at the same time by having the network learn how an objects appearance might change.

 Very interesting video and you can see how the researcher have fun while building sound technology that is on the leading edge of robotics and visual tracking. 

There is no doubt that the Paul G Allen School or Computer Science and Engineering is a scientific and robotics powerhouse that will continue to produce important research and ideas in robotics.

We leave you with a quote from Paul Allen:

You look at the things you enjoy in your life, but much more important is what you can do to make the world a better place.

We do not own the rights to images or videos used in this article.

Leave a comment