Announcement

Collapse
No announcement yet.

robot goes backward instead of forward?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #76
    Re: robot goes backward instead of forward?

    Originally posted by timdavid View Post
    As Tom stated, it gives us something to do while we wait for "Senior Solutions".
    For some reason, I keep having visions of brick sorters being re-purposed to dispense medications.
    FIRST Tech Challenge Judge: 2010, Referee: 2017
    FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
    FIRST Robotics Competition judge (Chairman's Award): 2014
    Dean says I'm an "Oompa Loompa of Science"

    Comment


    • #77
      Re: robot goes backward instead of forward?

      Originally posted by Tom Mosher View Post
      Close-spacing is good for tracking both sides of a black line, for example. Widely separated sensors are good for getting the robot perpendicular to a line. It all depends on how you want to use them.
      Thank you for taking the time to explain all this. I feel I really understand very well now. We are trying to make use of the sensors. Someone said you should keep 2 - 3 cm in between, while another person said 1 cm is good enough. We are trying to use them, and we noticed that the robot confuses the "white" with the blue water area on its way going to the sink. We tried to calibrate sensors again and we looked at the value of the colors in there, but it never reached 60. It was maximum 54 in the blue area very close to the sink. It is weird that the robot stops there when it is supposed to stop on the white line instead of the blue water area. That's confusing!!

      Another issue is when each robot gives a different value when is calibrated. We have two robots that have EXACTLY the same design, but one gives a value of 41 on a black line, while the other one gives a value of 38 on a black line. We are doing the claibration in school and on the same table. Isn't it weird??

      The third thing is the tall white wheels are too long when attached to the robot, and when we do a number of missions they open wider. I am not sure how to explain this but the problem is we have to press on the wheels and fix them so they do not open wide after a couple of missions because they are tall.

      Comment


      • #78
        Re: robot goes backward instead of forward?

        Originally posted by Tom Mosher View Post
        Try an experiment, and see what works best for your robot and what you're trying to accomplish.
        Two minutes and a half is a very short time to do all the missions. Compared to videos and missions done by other teams on youtube, I felt we cannot do all the amazing mechanics and programming they're doing. Some of them are even changing wheels as they change missions.

        I really liked the video you attched to one of the replies for the red robot. That was a very simple design, with very simple mechanics and pieces used for the missions. The kids saw it and loved it. But the programming is amazing. I wish we could learn how to program same way they are programming. I wish I could see how they program their robot. It's a great one. The kids are trying and trying, but doing two missions together is challenging for us, and this is why we cannot finish in 2.5 minutes. I feel bad.

        Comment


        • #79
          Re: robot goes backward instead of forward?

          Originally posted by timdavid View Post
          You should get 12 points for that scenario, not 7. The number of points should only be reduced to 7 if one of the bacteria from any of the dispensers ends up touching the mat outside of base at the end of the match..


          I'm glad you are having fun with your team. I think it's great that kids from around the world participate in FLL. Having your season extend a little later is fine also. As Tom stated, it gives us something to do while we wait for "Senior Solutions".
          Yeah... Thanks. We'll try our best, and I hope my kids won't get disappointed if we don't win any award.

          Comment


          • #80
            Re: robot goes backward instead of forward?

            Originally posted by guest2012 View Post
            Two minutes and a half is a very short time to do all the missions. Compared to videos and missions done by other teams on youtube, I felt we cannot do all the amazing mechanics and programming they're doing.
            I do hope you realize that the videos on youtube represent the best efforts of some of the most accomplished teams in the world. I hope the videos inspire rookie teams to try more things and learn new techniques, but rookie teams should not feel discouraged if they cannot come close to duplicating the feats they see on youtube. Moreover, even with all the knowledge and techniques these teams had, this year none of the teams have managed to complete all of the missions in two and a half minutes.

            I think once you get to a tournament, you will have a more accurate understanding of what most teams accomplish on the mission field.

            Comment


            • #81
              Re: robot goes backward instead of forward?

              Originally posted by guest2012 View Post
              Two minutes and a half is a very short time to do all the missions. Compared to videos and missions done by other teams on youtube, I felt we cannot do all the amazing mechanics and programming they're doing. Some of them are even changing wheels as they change missions.

              I really liked the video you attched to one of the replies for the red robot. That was a very simple design, with very simple mechanics and pieces used for the missions. The kids saw it and loved it. But the programming is amazing. I wish we could learn how to program same way they are programming. I wish I could see how they program their robot. It's a great one. The kids are trying and trying, but doing two missions together is challenging for us, and this is why we cannot finish in 2.5 minutes. I feel bad.
              You shouldn't feel badly, it is impossible to do all the missions in 2.5 minutes. None of the videos you have seen accomplish that. Perhaps set a goal for the team that can be accomplished with their skills. When they succeed, then have a celebration.

              There is a lot that can be accomplished this year with just simple straight-line drives. Fancy programming and attachments aren't really required to have fun and enjoy the activity. If you're a new team, that's a reasonable goal.

              Later on, you can learn how to do more sophisticated programs. It takes time to gather the experience needed. You can't rush it, so don't try. Do what your team is capable of, and have fun! Awards aren't really necessary.
              FIRST Tech Challenge Judge: 2010, Referee: 2017
              FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
              FIRST Robotics Competition judge (Chairman's Award): 2014
              Dean says I'm an "Oompa Loompa of Science"

              Comment


              • #82
                Re: robot goes backward instead of forward?

                Originally posted by guest2012 View Post
                Thank you for taking the time to explain all this. I feel I really understand very well now. We are trying to make use of the sensors. Someone said you should keep 2 - 3 cm in between, while another person said 1 cm is good enough. We are trying to use them, and we noticed that the robot confuses the "white" with the blue water area on its way going to the sink. We tried to calibrate sensors again and we looked at the value of the colors in there, but it never reached 60. It was maximum 54 in the blue area very close to the sink. It is weird that the robot stops there when it is supposed to stop on the white line instead of the blue water area. That's confusing!!

                Another issue is when each robot gives a different value when is calibrated. We have two robots that have EXACTLY the same design, but one gives a value of 41 on a black line, while the other one gives a value of 38 on a black line. We are doing the claibration in school and on the same table. Isn't it weird??

                The third thing is the tall white wheels are too long when attached to the robot, and when we do a number of missions they open wider. I am not sure how to explain this but the problem is we have to press on the wheels and fix them so they do not open wide after a couple of missions because they are tall.
                A few comments on your specific questions:

                Using Light Sensors
                The light sensor "sees" a wide area below the sensor. The higher the sensor is, the wider the area it's seeing. If the sensor is a few mm above the mat, it will see an area smaller than a small fingernail. If the sensor is higher, maybe 2 cm, it will see an area about the size of a thumbnail.

                What are your two sensors being used for? That determines how far apart they should be.

                If you're using two sensors to track a line, start experimenting with the sensors as far apart as the width of the line, and low enough that each sensor sees half of the line. You can make adjustments and see how the behavior changes. When a sensor is over the edge of the line, a value around 50 is a good reference point.

                Tracking a line can also be done with just one sensor, as others have mentioned. The robot drives to keep the sensor over the edge of the line, not the middle.

                If you're trying to detect a line using two sensors (so the robot is perpendicular to it), then having them far apart is good.

                You can experiment with the light sensor using a small test program. Inside a loop, read the light sensor, convert it to text, and output it to the NXT display. That's a very handy technique. Make it a MyBlock, and you can drop it inside your mission program to help with testing.

                With your test program, hold the sensor in your hand, and you can move it around to different locations and heights, and see how the readings vary.

                If you're not getting 0 when the sensor is over Black, and 100 over White (plus or minus a few counts), you have calibration problems.

                Light sensor calibration:
                As the sensor passes over the border between a black line and the white background, the reading should change from 100 (when it's seeing only the white area) to 0 (when it's seeing only the black area). If that's not what you're observing, you need to run the Calibrate program.

                To calibrate, start the Calibrate program, put the sensor over Black, push the orange button, and wait a few seconds, then put it over White and push the orange button again. Wait a few seconds, and it's done.

                The light sensor calibration data is stored in the NXT brick's memory, so once it's calibrated, keep that sensor with that robot.

                If you have two sensors, you can only calibrate using one of them. The two sensors will give slightly different readings.

                Sensor calibration isn't very accurate. Having readings of 38 and 41 is very little difference, that's expected and not a problem. When you're over a black line, you should consistently get a very low value.

                Colors
                The Light sensor isn't very good at telling the difference between different colors, it just measures brightness. Blue looks a lot like Black to the Light sensor. If you want to measure the different colors of the mat (brown and blue), you will need to use the Color sensor (the one with three sensor elements in it).

                You can still use the Light sensor. Notice that the black lines are always surrounded by a white background. If you're trying to drive across the blue background and detect a black line, you can drive (slowly) across the blue until you find the white background, then drive slowly looking for the black line. It isn't complicated, it only takes four blocks (Move-unlimited, Wait (for White), Move-unlimited, Wait (for Black).

                Wheels
                It's difficult to tell without a picture. Perhaps like the connection between the wheel and motor is too weak for the weight of your robot. Perhaps the axles are too long, or the robot is too heavy.
                FIRST Tech Challenge Judge: 2010, Referee: 2017
                FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
                FIRST Robotics Competition judge (Chairman's Award): 2014
                Dean says I'm an "Oompa Loompa of Science"

                Comment


                • #83
                  Re: robot goes backward instead of forward?

                  Originally posted by Tom Mosher View Post
                  A few comments on your specific questions:

                  Wheels
                  It's difficult to tell without a picture. Perhaps like the connection between the wheel and motor is too weak for the weight of your robot. Perhaps the axles are too long, or the robot is too heavy.
                  Our light sensors are about 0.65 cm high from the mat. I dunno if you can tell from the picture, but these 3 different designs we did them at different times.

                  -----------------------

                  **This is my best student who built this first design and then fixed it:
                  Old Robot Design (1):

                  http://forums.usfirst.org/attachment...1&d=1334210102

                  -----------------------

                  **This is the design after fixing it and we are using it now:
                  New Robot Design (2):

                  http://forums.usfirst.org/attachment...1&d=1334209656

                  http://forums.usfirst.org/attachment...1&d=1334209836

                  -----------------------

                  **This is what two other determined students worked on recently and I really liked it:
                  Robot Design (3):

                  http://forums.usfirst.org/attachment...1&d=1334208936

                  http://forums.usfirst.org/attachment...1&d=1334209333

                  -----------------------

                  Let me know what you think, and if anything needs to be added or fixed Please.
                  Attached Files

                  Comment


                  • #84
                    Re: robot goes backward instead of forward?

                    Originally posted by guest2012 View Post
                    Our light sensors are about 0.65 cm high from the mat. I dunno if you can tell from the picture, but these 3 different designs we did them at different times.
                    ...
                    Let me know what you think, and if anything needs to be added or fixed Please.
                    Your team has built some interesting robots! You should be very proud of what your kids have accomplished this season. It looks like they have done an excellent job of building and evaluating different robot designs. If possible, have the kids write down the good (and possibly bad) characteristics of each design. Have them bring that information with pictures of the different robot designs to the robot design judging session at their tournament and show it to the judges. The judges are evaluating the team in part on its design process. The judges would love to see the process and thinking behind how the team came up with its robot design.

                    Also, be aware that a team can bring multiple different robots to the tournament. They can show the judges different robot designs that are still built (in addition to pictures). The team can also take 2 different robots to the robot game field for different matches. For example, a team can bring one robot to the field for match 1, a second different robot to the field for match 2, etc. This is not a strategy I necessarily recommend, but it is an option if you have a large team with multiple competing robot designs, and the team cannot decide on a single "best" design they want to use all the time.

                    I'm going to pass on making specific critiques of the robot designs for a couple of reasons. First, it can be hard to evaluate a robot design just from a static picture. It's easier to evaluate a design based on how it moves and accomplishes different mission goals. Second, it's more important for the kids to evaluate the different designs and come up with their own conclusions. The kids might learn that there is not always one single "best" design. There are usually tradeoffs between different design approaches. How the team evaluates these tradeoffs and makes decisions is a big part of FLL. Whether the team comes up with the "right" answer is not actually as important as whether the team can explain why they came up with their answer.

                    Having stated the above, I will give you a few general tips I have found useful in evaluating robots:
                    1. Pay close attention to the weight balance of the robot. Where is the center of gravity for the robot? How much weight of the robot is over the drive wheels? How much weight of the robot is over the non-drive wheel(s)? If too much of the weight of the robot is over the non-drive wheels, the robot may be very inconsistent in making turns.
                    2. If light sensors are too close to the mat, waves or ripples in the mat can cause problems. Usually, light sensors work relatively well at a height between 1 to 2 beams off the mat. I think a beam is around 0.8cm, so that would put the range at about 0.8 to 1.6 cms. It sounds like you may be a little closer than that, but not too far off.
                    3. It's been debated on this forum, but I recommend shielding the light sensors to reduce the impact of ambient light if possible. A light sensor shield is just some additional Lego pieces around or above the light sensors to block some of the ambient light and put the area under the light sensors more in the "shadows".


                    Keep having fun! Your team is doing excellent work!
                    Last edited by timdavid; 04-12-2012, 02:53 PM.

                    Comment


                    • #85
                      Re: robot goes backward instead of forward?

                      Originally posted by guest2012 View Post
                      Our light sensors are about 0.65 cm high from the mat. I dunno if you can tell from the picture, but these 3 different designs we did them at different times.


                      Let me know what you think, and if anything needs to be added or fixed Please.
                      Thanks for sharing the photos, those are two very different and interesting robot designs.

                      As I mentioned before, I'm curious how the two light sensors are being used.

                      Experiment with the height and spacing of the sensors to get the best performance for your purpose.

                      As timdavid said, it's important that the team make their own evaluations.

                      For the robot with the large wheels, my observation is that there is a lot of weight resting on the main wheels, and also a lot of weight on the two skid-wheels in the front. It's possible that could make the robot difficult to turn, and the amount of torque required from the rear wheels might be causing the main wheels to slide sideways on the axles. That would let the axles bend, and make the tops of the wheels tilt inward. Is that what you see when you mentioned the problem that the wheels "open wider"?
                      FIRST Tech Challenge Judge: 2010, Referee: 2017
                      FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
                      FIRST Robotics Competition judge (Chairman's Award): 2014
                      Dean says I'm an "Oompa Loompa of Science"

                      Comment


                      • #86
                        Re: robot goes backward instead of forward?

                        Originally posted by Tom Mosher View Post
                        [B]

                        If you're trying to detect a line using two sensors (so the robot is perpendicular to it), then having them far apart is good.

                        You can experiment with the light sensor using a small test program. Inside a loop, read the light sensor, convert it to text, and output it to the NXT display. That's a very handy technique. Make it a MyBlock, and you can drop it inside your mission program to help with testing.

                        With your test program, hold the sensor in your hand, and you can move it around to different locations and heights, and see how the readings vary.

                        If you're not getting 0 when the sensor is over Black, and 100 over White (plus or minus a few counts), you have calibration problems.

                        Light sensor calibration:
                        As the sensor passes over the border between a black line and the white background, the reading should change from 100 (when it's seeing only the white area) to 0 (when it's seeing only the black area). If that's not what you're observing, you need to run the Calibrate program.

                        To calibrate, start the Calibrate program, put the sensor over Black, push the orange button, and wait a few seconds, then put it over White and push the orange button again. Wait a few seconds, and it's done.

                        The light sensor calibration data is stored in the NXT brick's memory, so once it's calibrated, keep that sensor with that robot.

                        If you have two sensors, you can only calibrate using one of them. The two sensors will give slightly different readings.
                        Thank you for all the explanation. I am using a test program now. I just realized that the calibration we were doing was inaccurate. The two light sensors we're using for the same robot have different calibration values. One gives a value of 9 while the other one gives a value of zero. They require much effort and time to understand how they work, and when to use them. I hope they won't cause many problems on the day of the competition.

                        Comment


                        • #87
                          Re: robot goes backward instead of forward?

                          Originally posted by timdavid View Post

                          Having stated the above, I will give you a few general tips I have found useful in evaluating robots:
                          1. Pay close attention to the weight balance of the robot. Where is the center of gravity for the robot? How much weight of the robot is over the drive wheels? How much weight of the robot is over the non-drive wheel(s)? If too much of the weight of the robot is over the non-drive wheels, the robot may be very inconsistent in making turns.
                          2. If light sensors are too close to the mat, waves or ripples in the mat can cause problems. Usually, light sensors work relatively well at a height between 1 to 2 beams off the mat. I think a beam is around 0.8cm, so that would put the range at about 0.8 to 1.6 cms. It sounds like you may be a little closer than that, but not too far off.
                          3. It's been debated on this forum, but I recommend shielding the light sensors to reduce the impact of ambient light if possible. A light sensor shield is just some additional Lego pieces around or above the light sensors to block some of the ambient light and put the area under the light sensors more in the "shadows".


                          Keep having fun! Your team is doing excellent work!
                          Thanks for complementing. I try to keep the kids busy and working, and some of them feel like they want to build their own robot. Even that we are one team, but they enjoy the challenge. They bring up some very nice ideas, and sometimes I really get surprised!

                          I read your critiques carefully, and I found them very useful. I will ask the kids to write down some notes about the robot designs. It might be too late now to decide on one as we need the first one, and maybe we'll use the second one if time permits. Thank you for all the help, but I might get back for a little more questions :P

                          Comment


                          • #88
                            Re: robot goes backward instead of forward?

                            Originally posted by Tom Mosher View Post
                            Thanks for sharing the photos, those are two very different and interesting robot designs.

                            As I mentioned before, I'm curious how the two light sensors are being used.

                            Experiment with the height and spacing of the sensors to get the best performance for your purpose.

                            As timdavid said, it's important that the team make their own evaluations.

                            For the robot with the large wheels, my observation is that there is a lot of weight resting on the main wheels, and also a lot of weight on the two skid-wheels in the front. It's possible that could make the robot difficult to turn, and the amount of torque required from the rear wheels might be causing the main wheels to slide sideways on the axles. That would let the axles bend, and make the tops of the wheels tilt inward. Is that what you see when you mentioned the problem that the wheels "open wider"?
                            We are currently using the two light sensors so the robot can track the two white lines, and I think we should keep them far apart. Sometimes we use them to detect a black line. We are going to fix it and keep 2 cm in between. I think this is the best idea?

                            I think you're right about putting too much weight over the wheels. The robot can make good turns, but I hope it won't affect the motion of the robot when doing missions far away from the base. I really liked the second design because it is slim, but good things come last... Aghhh...

                            The light sensors are giving us a very hard time. I hope it will be a happy ending. My kids are excited and are eager to learn. I don't have all the answers for their questions, and I am learning as much as they are.

                            Comment


                            • #89
                              Re: robot goes backward instead of forward?

                              Originally posted by guest2012 View Post
                              We are currently using the two light sensors so the robot can track the two white lines, and I think we should keep them far apart. Sometimes we use them to detect a black line. We are going to fix it and keep 2 cm in between. I think this is the best idea?

                              I think you're right about putting too much weight over the wheels. The robot can make good turns, but I hope it won't affect the motion of the robot when doing missions far away from the base. I really liked the second design because it is slim, but good things come last... Aghhh...

                              The light sensors are giving us a very hard time. I hope it will be a happy ending. My kids are excited and are eager to learn. I don't have all the answers for their questions, and I am learning as much as they are.
                              The most common way to use the light sensor is to track the edge of a black line. That's why the black lines are surrounded by white, to give the robot a high-contrast pattern to detect. When the sensor is over the black edge, it should give a reading very close to 50. When you're using two sensors to track a line, space them the same distance apart as the width of the line. The both sensors see a value of 50 (or very close to it). I don't have a mission mat with me, so I can't measure it. I do have a Mindstorms test mat, and the black line is 2 cm wide.
                              FIRST Tech Challenge Judge: 2010, Referee: 2017
                              FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
                              FIRST Robotics Competition judge (Chairman's Award): 2014
                              Dean says I'm an "Oompa Loompa of Science"

                              Comment


                              • #90
                                Re: robot goes backward instead of forward?

                                Originally posted by guest2012 View Post
                                Thank you for all the explanation. I am using a test program now. I just realized that the calibration we were doing was inaccurate. The two light sensors we're using for the same robot have different calibration values. One gives a value of 9 while the other one gives a value of zero. They require much effort and time to understand how they work, and when to use them. I hope they won't cause many problems on the day of the competition.
                                When you use two sensors, one of them is always going to be calibrated differently. That's because there is only one set of calibration data stored in the NXT brick. But having values of 0 and 9 is pretty good, it shouldn't cause you much trouble if you're trying to detect the value of 50 when your over the edge of a black line.

                                When you get to the contest site, be sure to run the calibration program on the table you're going to be using. The lighting in the room will probably be different than where you practice, and the light sensor might detect the difference. You will have about a minute to set up your robot on the table, that's when you can run the Calibrate program.
                                FIRST Tech Challenge Judge: 2010, Referee: 2017
                                FIRST LEGO League Mentor, Instructor, and/or Referee/Head Referee since 2011
                                FIRST Robotics Competition judge (Chairman's Award): 2014
                                Dean says I'm an "Oompa Loompa of Science"

                                Comment

                                Working...
                                X