HW4 published |
| פורסם ב-20/1/2025, 10:00:02 Created on 20/1/2025, 10:00:02 Создано20/1/2025, 10:00:02 تم النشر ب-20/1/2025, 10:00:02 |
HW3 extension | |
|
Hi guys, Following multiple requests, HW3 is extended until the end of the weekend (25.1, end of day). Oren |
| פורסם ב-20/1/2025, 08:22:54 Created on 20/1/2025, 08:22:54 Создано20/1/2025, 08:22:54 تم النشر ب-20/1/2025, 08:22:54 |
HW2 Graded | |
|
Hi guys, I (finally) finished checking HW2 (sorry for the delay, I'm struggling with overcoming a flu as some of you may have noticed in class last Sunday). In general, the submissions were great, nice job. Below you can find all the common mistakes for which points were deducted. If you are not sure about your specific submission, shoot me an email and I will send the specific comment ids. Best, Oren ***************************** Question: 2 Comment id: (i) Points deducted: 3 Description: The analysis should discuss that smaller radii require less computations (good) but harms connectivity and thus potential path quality (bad). A claim such as ""a small change in the radius will harm the asymptotic optimality of the algorithm."" is wrong: the proof shows that any radius above a given threshold will ensure AO. However, fine tuning of what this radius is has a *dramatic* effect on the algorithm: reducing it even by a bit will reduce the number of samples used which on the one hand reduces the quality of the solution obtained for a given number of iterations BUT makes each iteration much faster." Question: General Comment id: (ii) Points deducted: 2 Description: Sloppy submissions (formatting, numerous typos, capitalization etc.) Question: 3.1 Comment id: (iii) Points deducted: 1 Description: Failing to mention that the path lies on the visibility graph Some of you described the visibility graph which is ok if done precisely but statements such as ""path vertices lie on […] and obs. Edges"" are imprecise and were not accepted. Question: 4.3 (P1) Comment id: (iv) Points deducted: 3 Description: Cost as a function of n should typically be monotonically decreasing for any k *if* you follow the hint and compute the PRM incrementally. If You do not have monotonically decreasing plots you should explain why this happens. Question: 4.3 (P2) Comment id: (v) Points deducted: 3 Description: Some of you plotted the runtime as a function of n and *not* the cost as a function of the runtime. Question: 5.2-a Comment id: (vi) Points deducted: 2 Description: Wrong values for sphere radii. Question: 5.2-5 Comment id: (vii) Points deducted: 4 Description: Some of you got wrong plots (not sure why). Intuitively, time should decrease as the radii factor increases while the false negatives should increase. If you got a different behavior, this should have raised some red flags… Question: 3.3 Comment id: (viii) Points deducted: 3 Description: A node should not be expanded if its cost (corresponding to the tether length) is above L. Some of you did not mention this termination condition Question: 3.4 Comment id: (ix) Points deducted: 4 Description: Insufficient details were provided (e.g., how to handle the fact that the goal's h-signature is not specified) Question: General Comment id: (x) Points deducted: 3 Description: If you did not explicitly mention how the h-signature is used (i.e., its representation and how it is updated when extending an edge). Question: 2 Comment id: (xi) Points deducted: Description: Plotting \eta_d in three different plots, each with a different scale does not allow for proper analysis and comparison Question: general Comment id: N/A Points deducted: 0 Description: When asked to discuss the results, simply describing the plots is insufficient. You are expected to discuss *why* we get these results |
| פורסם ב-16/1/2025, 10:36:57 Created on 16/1/2025, 10:36:57 Создано16/1/2025, 10:36:57 تم النشر ب-16/1/2025, 10:36:57 |
Lab meeting for HW3 | |
|
Hi guys, Recall that as part of the submission of HW3, you will have to execute a path you generted offline on the robots. Don't worry, we will help you out in this event which will occur on Sun 19.4 from 15:30-17:30. Make sure to clear your schedule and come to the lab. Oren |
| פורסם ב-9/1/2025, 14:50:37 Created on 9/1/2025, 14:50:37 Создано9/1/2025, 14:50:37 تم النشر ب-9/1/2025, 14:50:37 |
HW3 correction (1) | |
|
Hi guys, In Sec 3.1 we asked you to "Report your performance (cost of the path and execution time) for coverage of 0.5 and 0.75," this is a typo simply report the performance (cost of the path and execution time) simsilar to what was asked in Sec 2.2. Apologies for the confusion. |
| פורסם ב-7/1/2025, 15:36:33 Created on 7/1/2025, 15:36:33 Создано7/1/2025, 15:36:33 تم النشر ب-7/1/2025, 15:36:33 |
HW3 published | |
|
Hi guys, HW2 is published and due on 20.1. For any support contact Eyal Weiss (ueweissatcampus.technion.ac.il) and cc me (uosalzmanatcs.technion.ac.il) Oren |
| פורסם ב-26/12/2024, 16:03:30 Created on 26/12/2024, 16:03:30 Создано26/12/2024, 16:03:30 تم النشر ب-26/12/2024, 16:03:30 |
Feedback for HW1 | |
|
Happy Hanukkah / Christmas! HW1 is graded. In general, everything was great, nice job! Common mistakes are below. If you did not get your a pdf back, this means that you got 100 or that your mistakes are covered below. Oren ****************************************** Q1b Some incomplete answers. You were expected to explicitly state that (iii) the Mink. sum of two general line segments is a parallelogram not a rectangle, a quadrilateral or a polygon (iv) the Mink. sum of two disks is a disk whose centers is the Mink. sum of the individual Disk's centers radius is the sum of the individual Disk's radii 1 point was deducted for each incorrect answer. Q2a Some of you explained what to do when he obs is not convex (typical answer was to triangulate it), others failed to give concrete examples in their answer. The question required you to discuss what happens when you do not do this and back your answer with an example. Between 2 and 5 points were deduced here. Q2b (*) Wrong complexity analysis (e.g., O(n^2) instead of O(n^3) for a naive implementation caused 5 points to be deduced. (*) Some of you split the parameters to number of vertices and num of edges - these are the same. 2 points to be deduced. Q4 (*) Some of you gave a wrong answer to 4.2.1c The correct one is that the final orientation: θ1 + θ2 + θ3 = 60◦ 2 points to be deduced. (*) Other minor errors caused 1 points to be deducted |
| פורסם ב-26/12/2024, 15:34:35 Created on 26/12/2024, 15:34:35 Создано26/12/2024, 15:34:35 تم النشر ب-26/12/2024, 15:34:35 |
Update & extension for HW2 | |
|
Hi guys, During the course we will learn about representing the homotopy class of planar paths and Q3 in HW2 studies this point. I was under the wrong assumption that we will already cover this material when I published HW2 so you will have this knoledge at your disposal. I am sorry for this mistake and apologise if this caused frustration. (1) Please srudy the topic of representing homotopy class by going over to (i) slides 31-32 of Lec 8 MHA* and (ii) Min 11-17 of the video on MHA*. Both can be found under the course material. (2) I added an extension to the HW, it is now due on the 2.1. Best, Oren |
| פורסם ב-24/12/2024, 07:52:03 Created on 24/12/2024, 07:52:03 Создано24/12/2024, 07:52:03 تم النشر ب-24/12/2024, 07:52:03 |
Small fix to HW2 code | |
|
Hi guys, We had a small error in the code for HW2 as pointed out by some students (thanks!). I uploaded an updated version of the code. Please download it and use that one. Sorry for the inconvenience, Oren |
| פורסם ב-19/12/2024, 04:30:00 Created on 19/12/2024, 04:30:00 Создано19/12/2024, 04:30:00 تم النشر ب-19/12/2024, 04:30:00 |
HW2 published | |
|
Hi guys, HW2 is published and due on 26.12. For any support contact Eyal Weiss (ueweissatcampus.technion.ac.il) and cc me (uosalzmanatcs.technion.ac.il) Oren |
| פורסם ב-5/12/2024, 17:39:48 Created on 5/12/2024, 17:39:48 Создано5/12/2024, 17:39:48 تم النشر ب-5/12/2024, 17:39:48 |
Class today is in person | |
| Just to avoid confusion - class today will be held in person |
| עדכון אחרון ב-1/12/2024, 12:32:31 Last updated on 1/12/2024, 12:32:31 Последняя модификация1/12/2024, 12:32:31 تمت الحتلنة الأخيرة ب-1/12/2024, 12:32:31 |
HW1 published | |
|
Hi guys, HW1 is published and due on 12.12. Not that some of the material has not been covered yet. For any support contact Eyal Weiss (ueweissatcampus.technion.ac.il) and cc me (uosalzmanatcs.technion.ac.il) Oren |
| פורסם ב-19/11/2024, 12:57:35 Created on 19/11/2024, 12:57:35 Создано19/11/2024, 12:57:35 تم النشر ب-19/11/2024, 12:57:35 |
Material uploaded | |
|
Hi guys, I uploaded the recorded lecture from Sunday (this is in a new folder so you will have access to old videos as well). Best, Oren |
| פורסם ב-12/11/2024, 07:31:57 Created on 12/11/2024, 07:31:57 Создано12/11/2024, 07:31:57 تم النشر ب-12/11/2024, 07:31:57 |
Welcome to Alg Motion Planning (236767) | |
|
Welcome to Algorithmic Robot Motion Planning Due to the situation, we will have most of the classes via zoom with some (after Hanukah) n the lab where we will experiment with the robots. I will explain everything on Sunday when we will meet using the following zoom link: https://technion.zoom.us/j/92577974134. The lecture + recitation gives us the windown of 14:30-17:30 but the meeting will likely end before 17:00. Take care, Oren |
| פורסם ב-7/11/2024, 09:27:28 Created on 7/11/2024, 09:27:28 Создано7/11/2024, 09:27:28 تم النشر ب-7/11/2024, 09:27:28 |