- 上一篇：变速器噪音英文文献和中文翻译
- 下一篇：JSP应用程序框架英文文献和中文翻译
The switch lead needs to be connected to ground to cycle tothe next color (it actually goes color-off-color-off…). A relay switch is used to turnground on and off that third pin. One problem that arose was that these LEDs did notfunction correctly when hooked up the board. They would only cycle through threecolors, but they had worked fine in testing. I finally decided that ground must have beenacting funny on the board, so I hooked the LEDs to their own battery. This worked andthey now cycle through all seven colors, but the voltage of this new battery is higher, andthe lights may not be shining at he same frequencies as before, so their reading may notbe calibrated correctly now.Figure 8BehaviorsThe Robot was supposed to wall follow along the rails of the pool table. Thisproved difficult because since the robot is so long. I first the robot speeding up onewheel when there was a discrepancy between the side IR, but by the time the other sidewould catch up, it would be going to fast and Rack Attack would over correct. I fixedthis by only having one wheel ever change speeds instead of both. So one wheel either speeds up or slows down depending on what is needed. I also put a delay in betweenchanges, and I said that no changes should occur if the wheels the robot is makingforward progress towards correcting itself. This seemed to stop the over correctingproblem, but the robot does tend to lean to the left still.Below is a drawing of the projected path of the robot when navigating the table.Basically the robot drives, and till it gets close to the next wall and then turns before thewall and back up to the pocket where it stops so that the back left corner of the robot ispositioned at the pocket.Figure 9The side IR are supposed to also tell the robot when it has reached a side pocket.First when the front side IR reaches the pocket its value would drop suddenly, and therobot will slow down. Then once the robot waits for the back side IR to drop suddenly invalue, and then as soon as it starts raising its value (just passed the halfway point of thepocket) the robot would stop, and the arm would be lined up perfectly with the hole.Problems arise in the form of the robot not being aligned right with hole, from the robotbeing not straight, or not stopping when it should. ConclusionRack attack can move its arm to any spot defined by a radial coordinate system.The arm can go to the pocket (assuming he robot is parallel to the wall), pick up the poolball, drop it in the color sensor box, and move it to the rack. The color sensor box canreliably find the color of any ball, but it does have some trouble with stripe solididentification. It sometimes mistakes solids for stripes. The robot can wall follow well,but it does not accurately stop itself at the pocket. The IR sensors should be enough tonavigate the table, but much more time is needed to work on the programming.
Thingsjust don’t happen as you think they will when you are programming.Also the bump sensor on the end of the arm does not work well, and needs to beredesigned. In retrospect, it might have been better to attach some sort of camera to theend of the arm to find the balls in the pocket, but this would have added to the weight. SoI am still not sure what would work best. An IR on the end could tell you how far awaythe balls are, and that might be enough, but the vacuum cup might get in the way so thiscould be a problem. I still think the customized bump sensor would work best, but itneeds to have springs incorporated somehow, because gravity is not good enough to holdthe rods down when they are not in contact.I want to keep working on the robot though, even after the class is over, and oneday it will do everything we hoped it would
目 录
1.摘要…1
2.概要 ..1
3.介绍.….....1
4.集成系统..2
5.移动平台 2
6.驱动. ..3
7.传感器4
7.1.红外...4
7.2.撞...…5
7.3.光. ...5