I have read many papers, and will likely read many more. Here is my attempt to keep them organized in some small way.
Languages
On Composing and Proving the Correctness of Reactive Behavior
- Harel, David and Kantor, Amir and Katz, Guy and Marron, Assaf and Mizrahi, Lior and Weiss, Gera
- (TODO: Summary) Examples of using "behavioral programming" using the Z3 SMT solver to make some claims. Only skimmed.
- Attach:Harel13Behavior.pdf
@inproceedings{harel2013composing,
title={On composing and proving the correctness of reactive behavior},
author={Harel, David and Kantor, Amir and Katz, Guy and Marron, Assaf and Mizrahi, Lior and Weiss, Gera},
booktitle={Embedded Software (EMSOFT), 2013 Proceedings of the International Conference on},
pages={1--10},
year={2013},
organization={IEEE}
}
Arrows, Robots, and Functional Reactive Programming
- Hudak, Paul and Courtney, Antony and Nilsson, Henrik and Peterson, John
- TODO: new summary. Get this: it's a description of a functional language, and I couldn't make heads or tails of it! It is supposed to describe how to program a robot with it, but I couldn't get through the languages description.
- Attach:Hudak03Yampa.pdf
@incollection{hudak2003arrows,
title={Arrows, robots, and functional reactive programming},
author={Hudak, Paul and Courtney, Antony and Nilsson, Henrik and Peterson, John},
booktitle={Advanced Functional Programming},
pages={159--187},
year={2003},
publisher={Springer}
}
The Yampa Arcade
- Antony Courtney and Henrik Nilsson and John Peterson
- This paper explains how to program a game of Space Invaders using Yampa. I still can't make heads or tails of it.
- Attach:Courtney03Arcade.pdf
@InProceedings{Courtney2003b,
author = "Antony Courtney and Henrik Nilsson and John Peterson",
title = "The {Y}ampa Arcade",
booktitle = "Proceedings of the 2003 {ACM SIGPLAN} {H}askell
Workshop ({H}askell'03)",
pages = "7--18",
year = 2003,
address = "Uppsala, Sweden",
publisher = "{ACM} Press",
month = aug
ALFA: A Language for Programming Reactive Robotic Control Systems
- Gat, Erann
- This paper describes the ALFA language, and gives two brief examples of robots using an earlier version. ALFA may be best understood in context of ATLANTIS, the system it was designed for. It provides the reactive layer, for which it is suitable: the module and channels abstraction allows for close correspondence to how systems are physically constructed and is limited enough to provide useful properties such as bounded response time.
- Attach:Gat91ALFA.pdf
@inproceedings{gat1991alfa,
title={Alfa: A language for programming reactive robotic control systems},
author={Gat, Erann},
booktitle={Robotics and Automation, 1991. Proceedings., 1991 IEEE International Conference on},
pages={1116--1121},
year={1991},
organization={IEEE}
}
Multi-Agent Planning
Introduction to Planning in Multiagent Systems
- de Weerdt, Mathijs and Clement, Brad
- This summary is seems far too short and lacking in detail. However, it serves its purpose as a good jumping off point to find relevant systems. Breaks down an approximate process for Multiagent planning, yet this seems to be very loosely followed (perhaps due to the field and no fault of the authors).
- Attach:deWeerdt09IntroMultiagentPlanning.pdf
@article{de2009introduction,
title={Introduction to planning in multiagent systems},
author={de Weerdt, Mathijs and Clement, Brad},
journal={Multiagent and Grid Systems},
volume={5},
number={4},
pages={345--355},
year={2009},
publisher={IOS Press}
}
Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations
- Shoham, Yoav and Leyton-Brown, Kevin
- This is a text book suitable for an advanced class on Multiagent planning. It covers a lot of ground, with the second chapter being of the most interest to me.
- Attach:Shoham09MultiagentSystems.pdf
@book{shoham2009multiagent,
title={Multiagent systems: Algorithmic, game-theoretic, and logical foundations},
author={Shoham, Yoav and Leyton-Brown, Kevin},
year={2009},
publisher={Cambridge University Press}
}
ALLIANCE: An Architecture for Fault Tolerant Multi-Robot Cooperation
- Parker, Lynne E
- Robot behavior is decomposed into behavior sets (a common technique) which represent tasks such as moving a box to a loading area (the level at which these behaviors is constructed is an interesting problem, and dependent on the mission at hand an how it is decomposed). What ALLIANCE adds is a set of "motivations" such as impatience (how long a robot will wait for another robot to do a job) and acquiescence (how long a robot will try to complete a task). These allow a team of robots to accomplish a mission in a completely distributed method. Communication can be added make the team more efficient, but communication failures are automatically handled (as it is not a requirement). L-ALLIANCE is mentioned, which allows for automatic turning of the motivation parameters.
A simple waste cleanup experiment is used to demonstrate the robustness of the system. They introduced complete robot failures as well as task failures in the three robot team, which was able to still complete the mission. The motivations where shown to effectively allow the robots to select a task to perform.
- Attach:Parker98ALLIANCE.pdf
@article{parker1998alliance,
title={ALLIANCE: An architecture for fault tolerant multirobot cooperation},
author={Parker, Lynne E},
journal={Robotics and Automation, IEEE Transactions on},
volume={14},
number={2},
pages={220--240},
year={1998},
publisher={IEEE}
}
A Decentralized Architecture for Multi-Robot Systems Based on the Null-Space-Behavioral Control with Application to Multi-Robot Border Patrolling
- Marino, Alessandro and Parker, Lynne E and Antonelli, Gianluca and Caccavale, Fabrizio
- Describes applying the Null-Space-Behavioral approach to a decentralized linear border patrolling problem. A three layer architecture (not to be confused with ThreeLevel or 3T) is described, with a supervisor modeled as a finite state machine at the top level. The states are actions the robot should be performing, and the transitions are based upon sensor inputs. The actions are composed of behaviors; for example the Patrol Border action may be composed of the follow border, avoid obstacle, and detect intruder behaviors. NSB is the manner in which this composition happens: lower priority behaviors can contribute to control output, but only if the contribute does not conflict with that of high level behaviors. The mathematical description of this is better explained (though still beyond me) in Antonelli's paper.
Of interest to me is that they used Player Stage as their simulator, and present their results from said simulation. Experiments were also performed with actual robots.
- Attach:Marino12Decentralized.pdf
@article{marino2012decentralized,
title={A decentralized architecture for multi-robot systems based on the Null-Space-Behavioral control with application to multi-robot border patrolling},
author={Marino, Alessandro and Parker, Lynne E and Antonelli, Gianluca and Caccavale, Fabrizio},
journal={Journal of Intelligent \& Robotic Systems},
pages={1--22},
year={2012},
publisher={Springer}
}
Planning (navigation and path)
A Gradient Method for Realtime Robot Control
- Konolige, Kurt
- Combines a path planner (wavefront variant) with a local navigator (similar to potential fields) to create a planner which eliminates the problems of reconciling the two while navigating. Vectors are calculated for the entire path, and can be interpolated between if necessary. The "realtime" refers to after the path planner has run (only at the beginning hopefully). Can handle (some) new obstacles, but does not mention forcing re-planning.
- Attach:Konolige00Gradient.pdf
@inproceedings{konolige2000gradient,
title={A gradient method for realtime robot control},
author={Konolige, Kurt},
booktitle={Intelligent Robots and Systems, 2000.(IROS 2000). Proceedings. 2000 IEEE/RSJ International Conference on},
volume={1},
pages={639--646},
year={2000},
organization={IEEE}
}
The Null-Space-Based Behavioral Control for Autonomous Robotic Systems
@article{antonelli2008null,
title={The Null-Space-based Behavioral control for autonomous robotic systems},
author={Antonelli, Gianluca and Arrichiello, Filippo and Chiaverini, Stefano},
journal={Intelligent Service Robotics},
volume={1},
number={1},
pages={27--39},
year={2008},
publisher={Springer}
}
Kinematic Control of Platoons of Autonomous Vehicles
- Antonelli, Gianluca and Chiaverini, Stefano
- According to Papers#Marino12Decentralized this paper explains the Null-Space Behavioral approach. I don't know if it does that, as the math was beyond my understanding. A Closed-Loop Inverse Kinematic is used to calculate reference trajectories for a group of robots. Multiple goals are accounted for, and conflicts are handled via priority for each goal. Demonstrated with an entrapment / escort simulation in which nine robots encircle a moving target while still avoiding obstacles. The algorithm does not deal with local minima, which is expected as it is at the local planner level. It sends its reference trajectories (or interpolated points) to a local navigator, and depends on higher level planners to avoid local traps. It is, however, singularity-robust... whatever that means. See paper above for a more complete description of NSB.
- Attach:Antonelli06KinematicPlatoons.pdf
@article{antonelli2006kinematic,
title={Kinematic control of platoons of autonomous vehicles},
author={Antonelli, Gianluca and Chiaverini, Stefano},
journal={Robotics, IEEE Transactions on},
volume={22},
number={6},
pages={1285--1292},
year={2006},
publisher={IEEE}
}
Sensor Networks
A Light-Weight Event-Driven Protocol for Sensor Clustering in Wireless Camera Networks
- Medeiros, Henry and Park, Johnny and Kak, Avinash
- An algorithm for tracking a moving object with a network of cameras. That's exactly what the title says. TODO: Better Summary.
- Attach:Mereiros07ALight.pdf
@inproceedings{medeiros2007light,
title={A light-weight event-driven protocol for sensor clustering in wireless camera networks},
author={Medeiros, Henry and Park, Johnny and Kak, Avinash},
booktitle={Distributed Smart Cameras, 2007. ICDSC'07. First ACM/IEEE International Conference on},
pages={203--210},
year={2007},
organization={IEEE}
}
Real-Time
Feasibility Analysis of Fault-Tolerant Real-Time Task Sets
- Burns, Alan and Davis, Robert and Punnekkat, Sasikumar
- Very short paper. Gives a feasibility test for a RT task set that can tolerate faults. Looks into several recovery schemes: re-execution, recovery blocks, exception handling and checkpointing. "Checkpointing will be beneficial only for tasks with computation time greater than checkpoint overhead." Recovery blocks sounds a little similar to n-version. TODO: finish summary.
- Attach:Burns96FeasibilityFTRT.pdf
@inproceedings{burns1996feasibility,
title={Feasibility analysis of fault-tolerant real-time task sets},
author={Burns, Alan and Davis, Robert and Punnekkat, Sasikumar},
booktitle={euromicro-rts},
pages={0029},
year={1996},
organization={IEEE}
}
Scheduling Analysis under Fault Bursts
@inproceedings{many2011scheduling,
title={Scheduling analysis under fault bursts},
author={Many, Florian and Doose, David},
booktitle={Real-Time and Embedded Technology and Applications Symposium (RTAS), 2011 17th IEEE},
pages={113--122},
year={2011},
organization={IEEE}
}
Bits and Peices
Byzantine Fault-Tolerant Publish/Subscribe: A Cloud Computing Infrastructure
DO NOT CITE (Position Paper)
@inproceedings{chang2012byzantine,
title={Byzantine Fault-Tolerant Publish/Subscribe: A Cloud Computing Infrastructure.},
author={Chang, Tiancheng and Meling, Hein},
booktitle={SRDS},
pages={454--456},
year={2012}
}
Integral Channel Features
- Doll{\'a}r, Piotr and Tu, Zhuowen and Perona, Pietro and Belongie, Serge
- libccv implements this for pedestrian detection. I use it for an example computational load. TODO: Read and summarize.
- Attach:dollar2009integral.pdf
@inproceedings{dollar2009integral,
title={Integral Channel Features.},
author={Doll{\'a}r, Piotr and Tu, Zhuowen and Perona, Pietro and Belongie, Serge},
booktitle={BMVC},
volume={2},
number={3},
pages={5},
year={2009}
}
libccv
- I'm not sure how to cite libccv, and liuliu (the only attribution I found) hasn't made any recommendations. I saw one paper which cited it as:
CCV-A Modern Computer Vision Library, "libccv.org", [Online] 2015, http://libccv.org/
A Study in Malloc: a Case of Excessive Minor Faults
- Ezolt, Phillip
- Not really academic. More of a detective mystery involving quirks of large memory allocations.
- Attach:ezolt2001malloc.pdf
@inproceedings{ezolt2001study,
title={A study in Malloc: a case of excessive minor faults},
author={Ezolt, Phillip},
booktitle={Proceedings of the 5th Annual Linux Showcase and Conference},
year={2001}
}
