-
00 : 01
© Distribution of this video is restricted by its owner
Transcript
×
Auto highlight
Font-size
| 00:06 | Welcome. I'm Professor Albert Cheng. in this five minutes I will overview |
|
|
|
| 00:14 | some key research in my research group also introduce my team to you. |
|
|
|
| 00:23 | you for your attention. So we going to look at how to build |
|
|
|
| 00:27 | next generation real time and cyber physical . So this is our research team |
|
|
|
| 00:39 | COVID-19 in front of the library, little bit of the real time systems |
|
|
|
| 00:48 | that you need to understand. And they are motivated by important applications such |
|
|
|
| 00:56 | this one which is the master over there was a priority inversion problem which |
|
|
|
| 01:05 | motivates our work in functional reactive The more complex a system becomes, |
|
|
|
| 01:19 | more difficult it is to maintain, analyze and to build. So there |
|
|
|
| 01:25 | different types of scheduling techniques, including and non preemptive. So it's okay |
|
|
|
| 01:34 | you're not familiar with the scheduling But the key point is we have |
|
|
|
| 01:39 | meet timing constraints such as deadlines and or delay requirements to ensure that a |
|
|
|
| 01:51 | time embedded and cyber physical systems can their logical as well as timing correctness |
|
|
|
| 02:02 | . So over the years, computer scientists and engineers in real time |
|
|
|
| 02:09 | have used a variety of techniques and a number of techniques to basically simplify |
|
|
|
| 02:16 | design and to ensure that these systems behave meeting timing constraints as well as |
|
|
|
| 02:25 | correctness constraints. So let's look at examples. So on the right, |
|
|
|
| 02:33 | see a spacecraft that has been developed here at the johnson Space Center. |
|
|
|
| 02:41 | on the left, it's a modern car with the control area networks, |
|
|
|
| 02:49 | computers, actuators and sensors. So have an embedded real time systems where |
|
|
|
| 02:56 | computer system is doing a lot of , but it's not the main part |
|
|
|
| 03:02 | the entire system. So obviously for car is to get you from A |
|
|
|
| 03:06 | B in the shortest time possible and safely as possible. There are lots |
|
|
|
| 03:14 | challenges such as complexity, reliability, and privacy. In addition to timing |
|
|
|
| 03:25 | in the interest of time, I not be able to go over the |
|
|
|
| 03:29 | . So I'm going to go over following slides more quickly. So one |
|
|
|
| 03:34 | the techniques that is being developed in lab is functional reactive programming or functional |
|
|
|
| 03:40 | systems. So utilizing techniques that are non procedural. So we're not using |
|
|
|
| 03:48 | CCPS class of python to program, rather we use a functional reactive programming |
|
|
|
| 03:55 | . So reactive systems can range from simple light bulb, the control switch |
|
|
|
| 04:01 | a pacemaker and a ps antilock brake in your car. Cyprus system, |
|
|
|
| 04:09 | , hyper physical systems. Combined, communication and actuacion elements includes the electric |
|
|
|
| 04:28 | . Mhm space uh and aircraft And is to ensure of course there's |
|
|
|
| 04:38 | collision and no accident. Even if humans make errors, functional reactive |
|
|
|
| 04:46 | As I said, it's a style is different from conventional procedural programs, |
|
|
|
| 04:51 | the challenge is how to determine their times accurately. Another big research area |
|
|
|
| 04:59 | hierarchical, real time scheduling. The of virtualization, but here also it |
|
|
|
| 05:05 | to meet timing constraints, not just effort requirements or performance requirements. Very |
|
|
|
| 05:13 | results. So we're pushing the boundary virtualization to increase the utilization while maintaining |
|
|
|
| 05:22 | ability and response time constraints. Other topics that are being done in my |
|
|
|
| 05:29 | includes intrusion detection in real time and physical systems and here are some of |
|
|
|
| 05:35 | recent publications in the area Real time learning. How do you trade off |
|
|
|
| 05:43 | with the quality of the deep learning or machine learning techniques and some of |
|
|
|
| 05:51 | papers in this area. The concluding include that we want to make |
|
|
|
| 05:58 | positive impact in society, not just papers and getting grants. So we |
|
|
|
| 06:05 | to make sure that the techniques developed our team is applicable applicable to society |
|
|
|
| 06:12 | make it better to make life more and safer. So, accomplishments from |
|
|
|
| 06:20 | team includes PhD students, graduates who outstanding awards as best student that's PhD |
|
|
|
| 06:31 | also best dissertation award and that includes students getting the best junior PhD student |
|
|
|
| 06:39 | and of course the professor Carlos, is an excellent um instructional assistant professor |
|
|
|
| 06:48 | also graduated from our group and of we have lots of outstanding undergrads and |
|
|
|
| 06:56 | are highlighted here. We have contributed and here are some of the publication |
|
|
|
| 07:03 | for authored by undergraduates. Thank you much for your |
|
|
|
A simple tooltip for the link