Murray Wiki - User contributions [en]

Daniel Fremont, Sep 2018

2018-09-21T21:00:11Z

Jge: /* Agenda */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office (337 Annenberg)
* 11 am: seminar in 314 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:15 pm: VeHiCal meeting
* 2:00 pm: Mani Chandy, 212 Annenberg
* 2:30 pm: Tung, Steele Laboratory library
* 3:15 pm: Sumanth, Steele Laboratory library
* 4:00 pm: Open
* 4:35 pm: Open, if needed (replace with your name and location)
* 5:00 pm: Yisong Yue, 303 Annenberg

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
314 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Daniel Fremont, Sep 2018

2018-09-21T17:59:06Z

Jge: /* Seminar */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office (337 Annenberg)
* 11 am: seminar in 314 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:15 pm: VeHiCal meeting
* 2:00 pm: Mani Chandy, 212 Annenberg
* 2:30 pm: Tung, Steele Laboratory library
* 3:15 pm: Sumanth, Steele Laboratory library
* 4:00 pm: Jin, 337 Annenberg
* 4:35 pm: Open, if needed (replace with your name and location)
* 5:00 pm: Yisong Yue, 303 Annenberg

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
314 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Daniel Fremont, Sep 2018

2018-09-21T17:56:30Z

Jge: /* Agenda */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office (337 Annenberg)
* 11 am: seminar in 314 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:15 pm: VeHiCal meeting
* 2:00 pm: Mani Chandy, 212 Annenberg
* 2:30 pm: Tung, Steele Laboratory library
* 3:15 pm: Sumanth, Steele Laboratory library
* 4:00 pm: Jin, 337 Annenberg
* 4:35 pm: Open, if needed (replace with your name and location)
* 5:00 pm: Yisong Yue, 303 Annenberg

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
121 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Daniel Fremont, Sep 2018

2018-09-21T15:55:54Z

Jge: /* Agenda */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office (337 Annenberg)
* 11 am: seminar in 121 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:15 pm: VeHiCal meeting
* 2:00 pm: Mani Chandy, 212 Annenberg
* 2:30 pm: Tung, Steele Laboratory library
* 3:15 pm: Sumanth, Steele Laboratory library
* 4:00 pm: Jin, 337 Annenberg
* 4:35 pm: Open, if needed (replace with your name and location)
* 5:00 pm: Yisong Yue, 303 Annenberg

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
121 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Daniel Fremont, Sep 2018

2018-09-21T15:55:27Z

Jge: /* Agenda */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office (337 Annenberg)
* 11 am: seminar in 121 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:15 pm: Open (replace with your name and location)
* 2:00 pm: Mani Chandy, 212 Annenberg
* 2:30 pm: Tung, Steele Laboratory library
* 3:15 pm: Sumanth, Steele Laboratory library
* 4:00 pm: Jin, 337 Annenberg
* 4:35 pm: Open, if needed (replace with your name and location)
* 5:00 pm: Yisong Yue, 303 Annenberg

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
121 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Daniel Fremont, Sep 2018

2018-09-17T14:40:09Z

Jge: /* Agenda */

Daniel Fremont, a PhD student at UC Berkeley working with Sanjit Seshia, will visit Caltech on 21 Sep. Sign of for a time to meet with him below.

=== Agenda ===
* ~10:30-10:45 am: arrive at Caltech; meet in Jin Ge's office ()
* 11 am: seminar in 121 ANB
* 12 pm: Lunch with Jin, X, Y, Z (sign up if you want to join)
* 1:30 pm: Jin, Ann 337
* 2:15 pm: Open (replace with your name and location)
* 3:00 pm: Open (replace with your name and location)
* 3:45 pm: Open (replace with your name and location)
* 4:30 pm: Open, if needed (replace with your name and location)
* 5:15 pm: done for the day

=== Seminar ===

Algorithmic Improvisation 
Daniel Fremont, UC Berkeley 

21 Sep 2018, 11 am 
121 Annenberg

Algorithmic Improvisation is a framework for automatically synthesizing systems with random but controllable behavior. It can be used in a wide variety of applications where randomness can provide variety, robustness, or unpredictability but safety guarantees or other properties must be ensured. These include software fuzz testing, robotic surveillance, machine music improvisation, randomized control of systems mimicking human behavior, and generation of synthetic data sets to train and test machine learning algorithms. In this talk, I will discuss both the theory of algorithmic improvisation and its practical applications. I will define the underlying formal language-theoretic problem, “control improvisation”, analyze its complexity and give efficient algorithms to solve it. I will describe in detail two applications: planning randomized patrol routes for surveillance robots, and generating random scenes of traffic to improve the reliability of neural networks used for autonomous driving. The latter application involves the design of a domain-specific probabilistic programming language to specify traffic and other scenarios.

Daniel Fremont is a PhD student in the Group in Logic and the Methodology of Science at UC Berkeley, working with Sanjit Seshia. He received a B.S. degree in Mathematics and Physics from MIT in 2013. His research is generally in the area of formal methods, focusing on the problems of counting and uniform generation of solutions to Boolean formulas. This includes developing practical algorithms to solve these problems, as well as finding new applications to the construction, verification, and testing of software, hardware, and cyber-physical systems.

Siva Seetharaman, 10 May 2018

2018-05-08T15:52:08Z

Jge: /* Schedule */

[[http:simons.berkeley.edu/people/siva-seetharaman|Siva Seetharaman]] is a doctoral candidate in the Department of Electrical Engineering at the University of Notre Dame. She obtained her undergraduate and Master’s degrees in Electrical Engineering from the PES Institute of Technology and the Indian Institute of Science, in 2011 and 2013, respectively. Sivaranjani’s research interests are in the area of distributed control for large-scale infrastructure networks, with emphasis on transportation networks and power grids.

=== Schedule ===

* 10:30 am: Richard Murray, 109 Steele Lab
* 11 am: informal seminar + Q&A
* Noon: lunch (Gabor Orosz to host)
* 1:15 pm: Jin
* 2:00 pm: open
* 2:45 pm: open
* 3:30 pm: open
* 4:15 pm: open (tentative; depending on flight times)
* 5:00 pm (done for the day)

=== Seminar ===

Congestion in Large-Scale Transportation Networks: Analysis and Control Perspectives 
Siva Seetharaman, Notre Dame 
10 May 2018, 11 am, 114 Steele Lab (library)

Fluid-like models their discretizations like the Cell Transmission Model (CTM), have proven successful in modeling traffic networks. However, given the complexity of the dynamics, it is not surprising that the stability properties of these models, especially in congested regimes, are not yet well characterized. The first half of this talk will propose a new modeling paradigm, where an analogy between discetized fluid-like traffic flow models and a class of chemical reaction networks is constructed by suitable relaxations of key conservation laws. This framework allows us to draw upon powerful structural results and entropy-like Lyapunov functions from chemical reaction network theory to study the existence and stability of congested steady states in networks with arbitrary toplogies. The second half of this talk will motivate compositional design approaches to mitigate congestion in large-scale transportation networks by describing a scalable distributed design that uses only local information to limit the propagation of congestion in the network.

Ben Richards, 20 Apr 2018

2018-04-17T13:56:19Z

Jge:

Ben Richards will visit Caltech on 20 Apr (Fri). Ben is an engineer for General Motors' North Hollywood Advanced Design Center who is interested in learning more about Caltech and our work on the VeHiCaL project.

Schedule:

* 2:00: Richard, 109 Steele Lab
* 2:45: Tung Minh Phan
* 3:30: Sumanth
* 4:15: Jin
* 5:00: Open

SURF discussions, Jan 2018

2018-01-22T17:30:58Z

Jge: /* 24 Jan (Wed) */

Slots for talking with applicants and co-mentors about SURF projects. Please sign up for one of the slots below. All times are PST. __NOTOC__

In preparation for our conversation, please do the following:
* SURF students should work with their co-mentors to find a time the meeting/Skype call. (For Skype calls, co-mentors should initiate.)
* Please make sure you have read the material in the description of your project, so that you are prepared to talk about what the project is about and we can narrow in on the key ideas that will be the basis of your proposal
* Please take a look at the [[SURF GOTChA chart]] page, which is the format that we will use for the first iteration of your project proposal.

{| border=1
|- valign=top
| width=30% |
==== 23 Jan (Tue) ====
* 12:00 pm PST: open
* 12:30 pm PST: open
* 1:00 pm PST: open
<hr>
* 4:00 pm PST: open
* 4:30 pm PST: open
<hr>
* 6:00 pm PST: open
* 6:30 pm PST: open
| width=30% |

==== 24 Jan (Wed) ====
* 7:30 am PST: open (hold for India/Europe)
* 8:00 am PST: open (hold for India/Europe)
<hr>
*12:15 pm PST: Filip
*12:45 pm PST: open (if needed)
<hr>
* 4:30 pm PST: open
* 5:00 pm PST: open
* 5:30 pm PST: open
|}

The agenda for the phone call is (roughly):

# Description of the basic idea behind the project (based on applicant's understanding)
# Discussion about approaches, things to read, variations to consider, etc
# Discussion of the format of the proposal
# Questions and discussion about the process

SURF 2018: Experimental verification of a semi-autonomous vehicle design based on human intention

2017-12-08T05:46:38Z

Jge:

While self-driving features such as adaptive cruise control and lane-keeping systems have been implemented in production vehicles, there are still many obstacles for the implementation of fully automated vehicles in real traffic. However, semi-autonomous driving can be achieved with few legal or policy obstacles. In particular, we are interested in a semi-autonomous vehicle design that can detect human intentions and interact safely with other road users. Such a vehicle design not only can enhance traffic safety, but it may also significantly benefit people who are physically challenged to drive.

In this study, we may first use data-driven methods to detect the intention of the human operator on-board and also the intention of other road users nearby. Then a higher-level advisory controller can be designed to execute, modify, or override the human intention, similar to the flight envelope protection system in aviation. Combining the intention detection, advisory controller and the lower-level driving/steering control, we have a semi-autonomous vehicle design.

The semi-autonomous vehicle design will be implemented and tested on F1/10 platform. The testing scenarios may include multi-lane driving, parking, or navigating intersections where other human-operated vehicles are around. For the controller design, general knowledge about planar rigid-body dynamics and PID control are desired. For the experimental implementation, programming experiences with Python and in ROS are needed.

References:

[1] S. A. Seshia, D. Sadigh and S. S. Sastry, "Formal methods for semi-autonomous driving," 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC), San Francisco, CA, 2015, pp. 1-5.

[2] B. Okumura et al., "Challenges in Perception and Decision Making for Intelligent Automotive Vehicles: A Case Study," in IEEE Transactions on Intelligent Vehicles, vol. 1, no. 1, pp. 20-32, March 2016.

[3] Duo Han, Yilin Mo and Richard M. Murray, “Synthesis of Distributed Longitudinal Control Protocols for a Platoon of Autonomous Vehicles”, Technical report done at Caltech, 2014.

[4] J. I. Ge, G. Orosz, and R. M. Murray. Connected cruise control design using probabilistic model checking. Proceedings of the American Control Conference, 4964-4970, IEEE, 2017.

[5] F110 autonomous race cars. <http://mlab-upenn.github.io/f110/index.html>

SURF 2018: Experimental verification of a semi-autonomous vehicle design based on human intention

2017-12-08T05:46:08Z

Jge: Created page with "While self-driving features such as adaptive cruise control and lane-keeping systems have been implemented in production vehicles, there are still many obstacles for the imple..."

While self-driving features such as adaptive cruise control and lane-keeping systems have been implemented in production vehicles, there are still many obstacles for the implementation of fully automated vehicles in real traffic. However, semi-autonomous driving can be achieved with few legal or policy obstacles. In particular, we are interested in a semi-autonomous vehicle design that can detect human intentions and interact safely with other road users. Such a vehicle design not only can enhance traffic safety, but it may also significantly benefit people who are physically challenged to drive.

In this study, we may first use data-driven methods to detect the intention of the human operator on-board and also the intention of other road users nearby. Then a higher-level advisory controller can be designed to execute, modify, or override the human intention, similar to the flight envelope protection system in aviation. Combining the intention detection, advisory controller and the lower-level driving/steering control, we have a semi-autonomous vehicle design.

The semi-autonomous vehicle design will be implemented and tested on F1/10 platform. The testing scenarios may include multi-lane driving, parking, or navigating intersections where other human-operated vehicles are around. For the controller design, general knowledge about planar rigid-body dynamics and PID control are desired. For the experimental implementation, programming experiences with Python and in ROS are needed.

References:

[1] S. A. Seshia, D. Sadigh and S. S. Sastry, "Formal methods for semi-autonomous driving," 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC), San Francisco, CA, 2015, pp. 1-5.
[2] B. Okumura et al., "Challenges in Perception and Decision Making for Intelligent Automotive Vehicles: A Case Study," in IEEE Transactions on Intelligent Vehicles, vol. 1, no. 1, pp. 20-32, March 2016.
[3] Duo Han, Yilin Mo and Richard M. Murray, “Synthesis of Distributed Longitudinal Control Protocols for a Platoon of Autonomous Vehicles”, Technical report done at Caltech, 2014.
[4] J. I. Ge, G. Orosz, and R. M. Murray. Connected cruise control design using probabilistic model checking. Proceedings of the American Control Conference, 4964-4970, IEEE, 2017.
[5] F110 autonomous race cars. <http://mlab-upenn.github.io/f110/index.html>

Karan Kalsi, Oct 2017

2017-10-02T15:41:03Z

Jge: /* 4 Oct (Wed) */

__NOTOC__
[[Image:Karan_Kalsi.jpg||right]]
Dr. Karan Kalsi is currently a power systems research engineer at the Pacific Northwest National Lab, Richland, WA, USA. He is the lead investigator on Department of Energy funded research to develop efficient, reliable and secure control strategies for Smart Grid assets. He is the principle investigator on the topic of developing future power grid control paradigms as part of the laboratory’s Future Power Grid Initiative. He is the Co-principle investigator on EED SEED project related to modeling and controls for distributed energy resources in microgrids. Karanjit is also actively involved with IEEE including being the 2011 Chair of the IEEE Richland Section Graduates of the Last Decade (GOLD) Affinity Group and was recently elected as the 2012 Vice-chair of Technical Activities and Outreach of the IEEE Richland Section Power and Energy Society.

More: http://energyenvironment.pnnl.gov/staff/staff_info.asp?staff_num=2178

== Schedule for Karan Kalsi visit, 2-5 Oct 2017 ==

=== 2 Oct (Mon) ===
* Morning: HR check-in, office setup
* Lunch: Richard
* 1:30 pm: Open
* 2:30 pm: Sofie Haesaert (Annenberg 310)
* 3:30 pm: Colloqium reception (location?)
* 4 pm: CMS colloqium - George Papanicoulou (Stanford), 105 Annenberg

=== 3 Oct (Tue) ===
* 9:30 am: Open
* 10:15 am: Tony Fragoso (Annenberg 3rd floor treehouse lounge)
* 11 am: Aaron Ames (266 Gates-Thomas)
* IST lunch bunch
* Afternoon off

=== 4 Oct (Wed) ===
* Morning off
* Lunch: Jin (and anyone else who wants to join)
* 3 pm: CDS tea
* 4 pm: Open
* 4:45 pm: Open

=== 5 Oct (Thu) ===
* 10 am: Soon-Jo Chung (235 Guggenheim) CAST and aerospace robotics research
* 11 am: Ioannis Filippidis (330 Annenberg)
* Lunch: Open
* 2 pm: head for airport