Thank you Cyrill for these uploads! I am a Masters student doing a dissertation only project and your videos are saving me so much time and confusion! Really great.
Thanks for this uploads. They are a big help for me to get into the basics of SLAM. I have read most of your papers but with this videos, explain step by step, its much easier to get it. Thank you very much.
Thank you so much for the uploads. Feels so good to be getting my attention span back while watching your lectures! Hope I keep coming across and learning from amazing minds like you.
@@CyrillStachniss thank you so much for making these lectures available to everyone :) you are doing a great service to the scientific and engineering community :)
you are a life saver :) perfectly explained, thank you so much for recording this video, i was thinking there will be no lecture on SLAM but fortunately i found it. Thankyou again.
Feeling the same ! :) I am interested to know where you applied it , i am going to use it on making Autonomous QuadCopter but in my area there is no one who can hlp with it ! Any suggestions regarding how should i start ! Thanx
Excellent video. Now, when I was studying SLAM using Probabilistic Robotics I was a bit taken aback about the fact that odometry was actually perceived as control. But when one looks close at it, odometry is only reporting what the robot thinks it has done, which is basically what you told it to do (control). It basically just translates your commands into some calculation supposing that internal parameter (such as wheel diameter) are correct.
this is actually amazing, i have a question though, if i were to use model predictive control to minimize navigation errors, how could i approach that ? and how about fault tolerant control, do you have any example of applying it for navigation ? my system is a differential drive, thank you in advance
Perfectly Explained. So i was currently working with dynamic SLAM problem. Could you suggest which sensors can be more effective in dense dynamic environment?
Hi Mr. Stachniss! Thanks for this video-lectures! I am new in this area (SLAM). I am trying to implement SLAM in a car robot with a four ultrasonic and a stereo camera. I having difficulties choosing and adapting one of the slam algorithms to this particular case. Could you give some advice about it?
I am in 2nd Year of B.Tech course ! Can you suggest prior Mathematical knowledge for completing this course or list of prerequisites . I need to learn SLAM for an upcoming project on Autonomous Quadcopter !
Sir i have a doubt? in this tutorial you explained us how slam works and mapping but you given everything as LA derivations i have doubt here that how to convert these derivations in to real-time coding in python or in C++ to make practically possible,because i'm still learning and exploring...
That mint is scary. what if it sends out mappings of your house to dubious people. now they know your house inside and out. it would be a serious IoT issue. nice course btw. Thank you!
ok, im fine with that. Just thought that since m represent the map we obtain from the measurements, it change or let me say its getting bigger for every timestep we do. So that z(t) depends on a other map then then z(t-5) and so on. But i guess the map we obtained at (t-1) looks different then at (t) but its still the same map/environment we working with. Thanks for the Help.
At 18:00, shouldn't the grey star be below the gold star(the star below the word same). Because in its wrong path the distance of the star from the robot measured by the robot is less than the actual distance so, in estimation of the star's position from the robot's path(the real one through which it should go) should be less. so, it should be nearer.
Okay, Sir Cyril what if you don't want wire linings as demarcation. Can the SLAM method still be durable. Or what other methods of mapping and localization helps a lawn mower possibly mower an area effectively.
Dear professor Cyrill Stachniss, I am an undergraduate student of China. I am writing to ask would you please offer me some information about where can I buy the book you use in this course? I am really interested in SLAM. Thank you.
The Probabilistic Robotics Course should be a rather good starting point on the basics. From there, I would continue with tutorials and by implementing papers.
Thank you, sir, for these lectures. I'm doing a project on SLAM, and I have some questions about slam. can you please help me out to solve those problems? (Lots of love from INDIA)
In the old days, people were going to the landmarks (towers) and take measurements there towards other towers. This created these networks. The problem on the robot is a different one.
Do we allways just have one map? Should not on the Graphical Model for every timestep a map. Just like the observations z(t-1), z(t)...z(T) there is a map m(t-1), m(t)...m(T)?