Concluding remarks with project summary | |
Hi all, First and foremost I want to wish that all of you stay safe and manage to cope with these crazy times. If you need any assistance, my virtual door is always open. Second, thank you for the great course. I enjoyed the lectures and interactions with you very much. The projects were, in general, much more than I expected and I enjoyed reading them. The final course average is very high (roughly 95), but I thought that this represents the work that you did and your high understanding. The guidelines for the project grades were as follows: 70 - complete but sloppy work 80 - ok work with some missing things that should have appeared 90 - good work but with no deep insight 100 - great insights or more than expected implementation I provided some individual feedback via email but since many projects addressed the same topic and you all studied IRIS, I thought it is worth having group feedback: (1) As a general guideline - if you have the option to present results in a table or a plot, always go for a plot unless the actual numbers are important. (2) Always mention the setting - what params did you use? Why and how dis you choose them? At least give this some thought. (3) We are not in the business of story telling - try to formalize your work (e.g., description of heuristics that you use). (4) Use a spellcheck (seriously?) (5) When analyzing results - don't look at things from a black-box perspective but see what is under the hood. E.g., if one heuristic is better than the other, great! Why? is the interesting question - is it because it is faster to compute but just as informative (how would you measure this?) or does it have some additional operations that are worth while? Best, Oren |
פורסם ב-18/3/2020, 11:43:09 Created on 18/3/2020, 11:43:09 Создано18/3/2020, 11:43:09 تم النشر ب-18/3/2020, 11:43:09 |
HW3 is graded | |
HW 3 is graded and grades are published (1) Many of you observed that for w2 between 1.4 and 1.5 MHA* switches from using the anchor heuristic to using the guidance. If you state the phenomenon but give no reason why then four points were deduced. The reason is pretty intuitive - the anchor heuristic tries (roughly) to follow the diagonal path while the inadmissible heuristic wants to take the "Manhattan path". The ratio is roughly \sqrt{2}~1.4 which is where we observe the transition. (2) A couple of you submitted a word doc and not a pdf even though you were explicitly instructed to submit a pdf. One point was deduced in this case. If I did not return your HW then you either got 100, 96 or 95 correponding to no errors, error (1) or error (1) and (2), respectively. If you got another grade then I returned the HW with comments. |
פורסם ב-26/1/2020, 11:52:59 Created on 26/1/2020, 11:52:59 Создано26/1/2020, 11:52:59 تم النشر ب-26/1/2020, 11:52:59 |
Robotics courses @ Technion - spring semester 2020 | |
Here is the list of courses I mentioned yesterday in class: Introduction to robotics (046212)---Mondays 15:30-18:30, Aviv Tamar Cognitive robots (097244)---Mondays 13:30-16:30, Erez Karpas Autonomous navigation and perception(086762)---Wednesdays 14:30-17:30, Vadim Indelman Technion Robotics Seminar---Mondays 12:30-14:00 [times will change, co-orgamized] CS Robotics Seminar (236824)---Wednesdays 16:30-18:30, me |
פורסם ב-15/1/2020, 08:03:38 Created on 15/1/2020, 08:03:38 Создано15/1/2020, 08:03:38 تم النشر ب-15/1/2020, 08:03:38 |
משאל המרצה | |
שלום רב, כפי שהזכרתי בכיתה, אנו שמים דגש רב על איכות ההוראה. אחד הכלים העומדים לרשותנו הוא משאל המרצה. אנא מלאו אותו ברצינות ואם יש לכם ביקורת - כתבו אותה בצורה בונה כדי שנוכל ללמוד ממנה ולשפר את הקורס בהמשך. תודה, אורן |
פורסם ב-8/1/2020, 08:55:16 Created on 8/1/2020, 08:55:16 Создано8/1/2020, 08:55:16 تم النشر ب-8/1/2020, 08:55:16 |
Final Project | |
Final project published under HW folder. |
פורסם ב-7/1/2020, 14:52:20 Created on 7/1/2020, 14:52:20 Создано7/1/2020, 14:52:20 تم النشر ب-7/1/2020, 14:52:20 |
Program change - Jan 7 | |
In class I mentioned that we will have a guest lecture next week by Elon Rimon. In order to finish the material on inspection planning, next week will be a regular lecture and the guest lecture will be at Jan 21st. |
פורסם ב-2/1/2020, 14:54:16 Created on 2/1/2020, 14:54:16 Создано2/1/2020, 14:54:16 تم النشر ب-2/1/2020, 14:54:16 |
HW2 Graded | |
HW2 is graded. While not asked for explicitly, I was hoping that you will also report the std for RRT and RRT* runs. Average may be misleading and as many of you pointed out in the writeup, these algorithms have a very high variance in their running time. Feedback is returned only for groups where points were taken off. |
פורסם ב-30/12/2019, 07:35:32 Created on 30/12/2019, 07:35:32 Создано30/12/2019, 07:35:32 تم النشر ب-30/12/2019, 07:35:32 |
Talk at TASP seminar | |
For those of you that did not register to the TASO seminar mailing list, today Prof. Hugo Guterman from BGU will give a talk at 12:30 in Taub 337. Title: "Safety Performance Evaluation Methodology For Autonomous Platforms" Abstract: One of the most urgent challenges in the field of autonomous platforms is the lack of an efficient Safety Performance Verification technique. Existing tools for hardware and software reliability and safety engineering do not provide a comprehensive solution regarding algorithms that are based on Artificial Intelligent (AI) and Machine Learning. In order to deal with this gap a novel methodology that is based on statistical testing in simulated environment, is presented. The methodology was developed within the framework of the RobIL program and is beginning to be tested into a variety of autonomous platforms. The RobIL program is a collaborative effort of several leading Israeli academic institutions and industries in the field of robotics. The current objective of the program is to develop an Autonomous Off-Road Unmanned Ground Vehicle (UGV). In this talk we will present the RobIL project with special emphasis on the performance and safety verification methodology, some new results of implementation of the methodology will be presented. |
פורסם ב-23/12/2019, 06:46:02 Created on 23/12/2019, 06:46:02 Создано23/12/2019, 06:46:02 تم النشر ب-23/12/2019, 06:46:02 |
HW extension | |
Due to multiple requests. HW2 is extended by three days to this Wednesday. Happy Hannuka, Oren |
פורסם ב-22/12/2019, 08:22:02 Created on 22/12/2019, 08:22:02 Создано22/12/2019, 08:22:02 تم النشر ب-22/12/2019, 08:22:02 |
Mailing list for Technion robotics seminar | |
For those of you who do not know, we have a Technion robotics seminar. For those interested in keeping up with talks and announcements, join the mailing list (simply add yourself after clicking on the following link: https://robotics-seminars.net.technion.ac.il/ ) Oren |
פורסם ב-11/12/2019, 09:46:57 Created on 11/12/2019, 09:46:57 Создано11/12/2019, 09:46:57 تم النشر ب-11/12/2019, 09:46:57 |
HW2 Fixes | |
We will use the following google doc for updates regarding the HW2 code: https://docs.google.com/document/d/1mw_AHbk4jekXLcsvDYjenggQMd6NSW3tvztZzCaN1MM/edit?usp=sharing Please check it out before you contact me or Peer. It contains clarifications and minor fixes. |
פורסם ב-9/12/2019, 20:12:32 Created on 9/12/2019, 20:12:32 Создано9/12/2019, 20:12:32 تم النشر ب-9/12/2019, 20:12:32 |
HW1 Returned | |
HW1 grades are published. There are two general comments regarding the HW: (1) There was some confusion in the dry exercise if "line" referred to a line or a line segment. I accepted both interpretations as long as the explanation was correct. (2) Almost all groups did NOT account for the (potential) intersection of C-space obstacles creating new vertices. These should have been accounted for when computing the visibility graph. However, they are NEVER part of the reduced visibility graph (why?) and thus do not impact the computation of shortest path. I did not take off points for groups that did not address this point. For those who did, I added five extra points. This brought one group to a total of 105 points ;-) Feedback is given only to groups for whom some points were deducted. Finally, if you did not get a grade, please send me an email with all group members cc'ed or come see me after class. |
פורסם ב-8/12/2019, 20:36:42 Created on 8/12/2019, 20:36:42 Создано8/12/2019, 20:36:42 تم النشر ب-8/12/2019, 20:36:42 |
HW2 Published | |
HW2 is published. Submission date is in just over two weeks (Dec 22nd, end of day). Good luck! |
פורסם ב-5/12/2019, 11:54:16 Created on 5/12/2019, 11:54:16 Создано5/12/2019, 11:54:16 تم النشر ب-5/12/2019, 11:54:16 |
HW extension | |
As promised in class, the deadline for HW1 was extended by a couple of days to Friday, end of Day. |
פורסם ב-4/12/2019, 09:55:33 Created on 4/12/2019, 09:55:33 Создано4/12/2019, 09:55:33 تم النشر ب-4/12/2019, 09:55:33 |
Short Survey | |
Please take the time and fill in a quick (one question / ten second) survey by following the link https://forms.gle/n3MKB91Xt9BsAKNZ9 Thanks, Oren |
פורסם ב-2/12/2019, 11:00:47 Created on 2/12/2019, 11:00:47 Создано2/12/2019, 11:00:47 تم النشر ب-2/12/2019, 11:00:47 |
Updated files for HW1 | |
When uploading the file "obstacles" the system removed endlines. This caused an error for the input parser provided to you. To bypass this, all source files were removed and uploaded again as one zip file. Please download it to get the updated obstacles file (all other files remain unchanged). Best, Oren |
פורסם ב-26/11/2019, 12:06:36 Created on 26/11/2019, 12:06:36 Создано26/11/2019, 12:06:36 تم النشر ب-26/11/2019, 12:06:36 |
Typo in HW1 | |
There was a small typo in exercise 1 of HW1. A new version has been uploaded to the website. Best, Oren |
פורסם ב-21/11/2019, 12:15:28 Created on 21/11/2019, 12:15:28 Создано21/11/2019, 12:15:28 تم النشر ب-21/11/2019, 12:15:28 |
Technical support for homeworks | |
For all code-related questions in the homeworks, please contact our wonderful TA Pe'er Sagiv upeer.tutorialatgmail.com |
פורסם ב-21/11/2019, 10:13:35 Created on 21/11/2019, 10:13:35 Создано21/11/2019, 10:13:35 تم النشر ب-21/11/2019, 10:13:35 |
HW1 Published | |
HW1 is published. Submission date is in two weeks (Dec 4th, end of day). Good luck! |
פורסם ב-20/11/2019, 11:22:48 Created on 20/11/2019, 11:22:48 Создано20/11/2019, 11:22:48 تم النشر ب-20/11/2019, 11:22:48 |
שעות קבלה | |
מחר, 11.11 באופן חד פעמי לא יתקיימו שעות קבלה. אם משהו התכוון להגיע לשעת קבלה, נא לתאם פגישה אישית באימייל. עמכם הסליחה, אורן |
פורסם ב-10/11/2019, 21:03:46 Created on 10/11/2019, 21:03:46 Создано10/11/2019, 21:03:46 تم النشر ب-10/11/2019, 21:03:46 |
שינוי חדר לשיעור ראשון אוק 29 | |
השיעור יתקיים באופן חד פעמי בטאוב 4 |
פורסם ב-24/10/2019, 16:00:54 Created on 24/10/2019, 16:00:54 Создано24/10/2019, 16:00:54 تم النشر ب-24/10/2019, 16:00:54 |
ביטול שיעור שני וקיום שיעור השלמה | |
השיעור השני (5.11) לא יתקיים כמתוכנן. במקומו ינתן שיעור השלמה ביום ראשון 17.11 ב17:30 בטאוב 8 |
פורסם ב-7/10/2019, 07:07:41 Created on 7/10/2019, 07:07:41 Создано7/10/2019, 07:07:41 تم النشر ب-7/10/2019, 07:07:41 |
קיום שיעור ראשון בזמן ארועי יובל לפקולטה | |
השיעור הראשון (יום ג, 29.10) יתקיים במקביל לארועי היובל לפקולטה. הלימודים יתקיימו כסדרן. |
פורסם ב-6/10/2019, 11:12:18 Created on 6/10/2019, 11:12:18 Создано6/10/2019, 11:12:18 تم النشر ب-6/10/2019, 11:12:18 |