Course given 2006/2007 in study period 4
To give an incentive to start working on the project early on, two labs are defined, lab0 and lab1. By completing lab0 you get 3 extra credits on the exam, and the second one is worth 5% of the grade (5 out of the max 35% in total for the project work).
3 new motors were bought before the course this year. Two groups had problems with their motors and we had to replace the encoders, using spare parts from old motors not working for other reasons. More care has to be taken next year to make sure that the motors are correctly connected to the boards at all times.
Patric Jensfelt replaced Henrik Christensen as lecturer for the course. The lecture notes were modified. Some example implementations were prepared to illustrate how robot localization can be accomplished in an environment similar to the soccer field using Kalman filters and particle filters. Robot hardware was brought to the lectures to show what kind of data common sensors can provide.
A guest lecturer was invited for the last lecture. Per Ljunggren, who lead the work at Elektrolux to develop the autonomous vacuum cleaner Trilobite, talked about the development process.
To increase the incentive for the students to meet the course requirements a week before the tournament, 5 out of the 35 point from the project work were given to those that completed the requirements the week before tournament deadline. This seems to have been a very good idea, every single student filling in the evaluation thought this was good and 32 out of 40 students met the early deadline.
Comments on lecture notes:
it realy sux not to get them in digital form.
---
It's sad that they are not avaliable on the cource home page as pdf's.
---
There were some type errors that need to be corrected
---
The lec
---
Very hard to understand in retrospect, especially since some parts that were on the original slides were missing on the lecture notes.
---
They only contain the headers, not too much information.
---
De var mer som en sammanfattning av boken, på gott och ont.
---
Not enough maths.. No use to buy a Kompedium, if we get slides.. the point of having something written is to get something we can refer back to, in case we need it later on. Not only has a list of ideas for the exam.
---
använde dem aldrig. fy på nada som inte vill ha ut dem på nätet. känns väldigt gammaldags. har man problem med att folk skriver ut för mkt får man väl lösa det problemet istället för att försöka göra en massa kringdribblande... förövrigt så kändes de inte så jättenödvändiga då i stort sett allt fanns i boken.
---
Concise and good
---
Det borde finnas mer utförlig text till dem. Mer än bara punktlistor. Det var ibland svårt att komma ihåg vad en sida handlade om. (Köpte inte boken)
---
Vissa rubriker var inte så tydliga, t.ex. "samples" samples för vad?
---
Som alltid med "powerpoint" presentationer så döljer det sig en massa bakom stödorden som inte kommer fram
Comments on course book:
pretty nice, the math parts was mostly confusing though.
---
Shallow but broad. A good starter!
---
Nice, light reading that also has substance. A good introductory book.
---
Its okay
---
Interesting albeit ultimately uneccesary for the project. Probably only useful when implementing the more fancy techniques like Kalman filtering.
---
Intressant
---
Didnt had time to read it and didnt need it to pass the course.
---
I didn't buy it.
---
läste hela, vissa delar flera ggr, det jag kände att jag behövde för att få koll till tentan. men som alltid så pluggas det för att maximera det som man kommer ihåg inför tentan, vilket ju inte är det bästa för långtidskunnandet.
---
Sometimes to deepgoing in some fields
---
Var bra komplement till föreläsningsanteckningarna.
---
Intressant läsning men kanske inte hjälper direkt för att bygga sin robot
Comments on lectures:
A little more about the eyebot and examples would have been nice.
---
I liked the real-life robot demo, heads up!
---
Unfortunatly the schedule was changed in a way I was unable to attend the lectures (the wednesday passes).
Something very unfortunate as I really liked the lectures.
---
They followed the lecture notes to closely. And the lecture notes followed the book much too closely. I think that lectures are a poor solution for this type of course. Use seminars where members from differens groups get togheter and examine case studies. The students should be required to read some course material detailing certain aspects of robotics before these seminars.
---
The lectures were good, interesting.
The double-sessions were maybe a bit long.
---
du är precis som jag inte den bästa pedagogen. men du försöker, och du gör saker som uppskattas, som att bjuda på godis när det var många timmar, fixa gästföreläsare och ta dit robotarna för att visa praktiska exempel. det som dock är lite trist är att det mesta andra är boken mer eller mindre rakt av. ur det hänseendet så tillför inte föreläsningarna så mycket utöver boken. det skulle kanske kunna vara värt ett försök att få föreläsningarna mer interaktiva, dvs, med mer interaktion med studenterna. visserligen finns det en hel del material som kanske inte behöver så mycket mer än en presentation bara, precis som det nu var.
---
Just adequate
---
* Jättekul andra föreläsning med många exempel på robotar. Himlans kul faktiskt.
* Skulle vilja ha fler real-life exempel för de andra föreläsningarna också.
---
När det kom kamanfilter och andra formler som jag aldrig sett förut så hade jag svårt att följa med
Comments on guest lecture:
maybe more interesting then relevant.
---
I missed that lecture.
---
Very interesting presentation.. Giving some ideas about real size projects..
---
det var en bra föreläsning. mer sådant! =)
---
It is good to realize the difficulties with building robots, and the guest lecture did it very well.
---
Superbra! Fler!
---
Mycket bra föreläsning där man fick lite insikt i svårigheterna att utveckla en ny produkt
Comments on the overlap:
Good overlap. If I was wondering anything about the content of the cource I could often find the answer in the book.
---
Not much overlapping.. maybe a bit with the AI course (same teacher, same examples :p)
---
Eftersom jag bara hade projektkurser så blev det aldrig krockar
Comments on the project:
id say more labs, or maybe demand more from them. lab0 was pretty much a joke that could be completed in a couple of hours.
---
Goals is very good to drive the project in a forward direction
---
Given how much time we spent on the project I think it should be a bigger part of the examination.
Perhaps more bonus points could be awarded for doing more technically advanced things. I.e. a lab2 and lab3 for doing X and Y.
---
lab0 and lab1 was great milestones for the project.
There should really be a lab0.5 where the robot must be able to distinguish between different object using the camera. And maybe a lab2 where the robot must be able to score a goal from a completely arbitrary position. Milestones are good. They do not take time from the project, instead they facilitate the project.
Paint the walls of the field in a cold white color. That is a hue that has a tad of blue in it to clearly distinguish it from the yellow goal.
Get matte balls.
Create truly diffuse lightning using a white sheet as a roof over the field that is illuminated by floodlights. It is then also easy to tune the color spectrum by adding color LEDS in different numbers to the overhead lighting armature.
---
It's annoying that it's so dependant on what people you get grouped with. If you don't know anybody or can't get a group with people you know it's like playing russian roulette.
---
Jag önskar att vi hade haft mer tid (mindre andra kurser) så man kunde lagt ner mer tid på projektet.
---
Our group was a bit lost.. no testing, last-minute coding.. Very disappointing when reaching the competition, feeling you could have done something much better..
---
nu har jag ju "fuskat" och läst det du skickat ut redan innan. lab0 var bra, så att man kom igång och gjorde något. som dock nämnts kan det vara bra att kanske ändra den så att det krävs något mer än att man klickar på en knapp, för att få igång mer programmerande. samtidigt finns det en poäng med att inte göra den för mycket. alla måste ha en chans att göra det också. lab1 var mycket bra. som alltid så finns det en tendens att man inte gör något förens man måste. vilket ju leder till att man kommer igång för sent. med lab1 en vecka innan tävlingen inser man om inte förens då, att det är ganska mycket att göra, och har åt minstonde en vecka på sig att fixa det ;)
---
Man kanske t.o.m. skulle ha fler steg innan tävlingen.
Kanske ett steg där alla får tävla mot "standardrobot" och se vad man kan förbättra innan tävlingen? t.ex. saker som man inte har tänkt på.
---
En stor del för att få en lyckad robot är att få en bra bildbehandling, kunna hämta ut bra och pålitlig information. Jag skulle behöva haft mer hjälp / teori om hur man löser detta. Detta pga av bildbehandlningen består av många steg och är komplex så när det går fel är det svårt att felsöka.
Comments on the project help:
Anders was very helpfull but hard to get a hold on sometimes
---
Anders was great. But most of the time you didn't need a physical person. Written explanations of the peculiarities of the EyeBot platform and guides would have been of great help.
---
The competition was, in my mind, far too early.. and the rapport expected far too late..
---
som nog ni också upptäckte så fick anders massa frågor när han väl var där, men inte alls så mycket frågor via mail el dyl. inte sant? visserligen var han där en hel del de sista dagarna innan tävlingen, vilket var väldigt bra. kanske kan vara en idé att schemalägga in en labass någon timme här o där andra dagar dock. ofta tänker man inte på att man har något problem, eller så tänker man att det går fortare att lösa det själv än att fråga om hjälp (vilket det kanske inte alls gör). att ha någon att fråga och bolla idéer mot som inte är i gruppen kan ju även vara bra (även om alla grupper egentligen väldigt mycket hjälper varandra så =)
---
Ja, tyvärr förstod man för sent hur snäll och hjälpsam han var. Säg det NOGA i början till nästa års studenter om han kommer att fortsätta :-)
Comments:
I found the exam to be accurate with the contents of the cource
---
The exam and the project seems to be widely separated. I think a seminar approach would be benificial here as well.
---
Kursen blev väldigt mycket inriktad på projektet, därav blev det som tentan handlade om ganska bortglömt. Det var dock bra att det inte var några matematiska formler på tentan eftersom jag har svårt att se nyttan i att kunna dem utantill.
---
The material taught and nothing else..
---
egentligen känns dock teoridelen och praktikdelen väldigt åtskiljda åt.
Comments:
I took the course in parallel with my exjobb.
---
The MVK (Mjukvarukonstruktion) cource have a project which took very much of the time
---
Jag hade Mjukvarukonstruktion och Operativssystem samtidigt, båda var väldigt tidskrävande kurser.
---
jag läste egentligen ingen annan kurs aktivt parallellt.. även om jag var antagen till några osv. men de var nedprioriterade redan sedan innan för andra kurser som jag eg inte läste nu.. osvosv ;)
---
some of it sleeping on the couch
---
Kursen tog mycket tid, men mycket pga att den var så kul.
the competition is realy fun.
the project in general is realy fun.
---
1. The laid back atmosphere.
2. The competition
3. The exam
4. Lab0 and lab1
5. Patric and Anders beeing helpful and nice
---
The project, the most other things were good too.
---
Project
Lab0 Lab1
Competition
---
Keep the project
Lab 0
Lab 1
---
1) The tournament.
2) The mechanical engineering part of the project. Great fun and a vast unexplored potential for interesting solutions.
3) The freedom in the software design.
---
the example program
the exam
---
The tournament must provaile!
---
* Bra verklighetsanknytning
* Intressanta vinklar till robotiken (de viktigaste bitarna togs upp)
* Projektet var väldigt lärorikt
* Kul kurs!
---
The fotballrobot is a very good way to understand how hard and fun robotics and autonomic systems realy are.
The competition which rounds up everything and makes you work as hard as possible.
The exam was realy good with nice relevant questions.
---
- The robot building !!
- The early labs.
- The sweets for the break between 2 courses (thanks by the way :p)
- The grading structure Robot/Exam that allows you to catch up on the robot... or the exam.
- The guest lecturer.
---
tävlingen
labbarna
---
1. Lab0 and Lab1 to incentivate an early start
2. Guest lecture
3.
---
lab0 & lab1
book
---
Nice boys (I'm bi) keep it up next year
I learned to play snooker
I stopped sleeping
I started to like tequila
I got it good
---
Tävlingen.
Anders Dovervik som labassistent.
Gästföreläsaren.
the camera is realy realy realy bad. it took up most of the time spent in the project to make it work. we could have done a much beather robot if we didnt have to spend al that time trying to write a good cameracode.
the "robotics lab" was way to hot and no oxygen at al. kinda hard to do smart stuff when you cant breath.
the project and lectures realy dont mix. we hardly used any knowledge learnt form lectures in the project. (mostly because we had to get that stupid camera to see stuff). it would have been realy fun to use localisation and stuff but there realy wasnt time.
---
1. Demos creating, transfering and running a program on the eyebot.
2. Lectures talk more about coding on the eyebot.
3. The workshop was a bit crowded during normal hours.
4. Adding to the rules during the cource.
5. The horrible lousy cams.
---
A lecture on probabilistic sensor representation would have been nice.
Perhaps a brief overview of image processing / classification techniques would have been helpful to some groups, since interpreting data from the camera is a big part of the project.
Perhaps a place where peculiarities of the used platform could be documented would be useful - e.g. a wiki or something?
---
Port the development kit for Eyebot to Windows/MacOSX and provide usb<->serial adapter for each group. So people will have better opportunity to work outside lab, as it's very cramped.
---
Exam credit should be less instead report and project credits should be increased
---
1) The lecture system is archaic. Case seminars where the students have read about specific (not necessarily overlapping!) areas beforehand would be much much better.
2) The ventilation in the workshops.
3) The restrictions in the rules regarding robot constructions combined with the provided material really narrows it down to a Differential-Drive Robot with a Roller. Not much room for explorative designs.
---
faulty hardware and lack of replacements
the eyecam
the lab and workshop were very crowded at times
---
Some more information about electronics and circuitry. The hardware was of varying quality between groups.
---
Shorter competition, this year it wasnt that many contestents but it was quite long anyway.
More spare parts just in case.
...
---
- Too many methods presented during the lectures, it looks like a catalog, we only get an overview of things.
- Buying the slides.. (the slides being the best thing to look at on your computer..)
- The fact the kompediums were made of slides and not text (no point in printing it then...)
- The size of the programming room.. 40 people in 20 m².. OUCH.
- The programming langage, but not much can be done about that..
---
kamerorna? :o
fixa så vi får öppna fönstren i labrummet.
---
1. Equal hardware for everybody. It is not fair that some groups need to solve hardwareproblems while other groups can spend their time in solving the requirements of Lab0, Lab1 and the competition.
2.
---
ta bort kvalet och gör lab1 till den avgörande för att klara projektdelen.
---
the camera
---
Need more nuts and screws
Addicted to X-ray energy drink
Nightmares about hue values
I stopped sleeping
I started to like tequila
---
Att alla på något sätt får spela mot varandra, då alla förmodligen har kämpat så hårt med både hård- och mjukvara och sovit på skolan.
Fler gästföreläsare.
En "standardrobot" som alla får mäta sig mot, så att det blir lika för alla.
Mer processorkraft så att man känner att det vore värt att försöka implementera något av det coola från föreläsningarna utan att riskera att vara för sölig på tävlingen.