 | | NXT Program: How to avoid stop-and-go? —bigboss97
|
| | | I've just started with my very first vehicle according to the instruction. Then I added the distance sensor. So, my car will turn when it come too close to an obstacle. I use a loop with distance as exit condition, i.e. the loop repeats the forward (...) (17 years ago, 20-Dec-07, to lugnet.robotics)
|
| | | |
| | | | | Re: NXT Program: How to avoid stop-and-go? —Brian Davis
|
| | | | (...) Try a Move block set to "unlimited": that will turn the motors on and let them run. Then have a Wait block to wait for the US sensor to report an object, or something similar inside a Loop. When an object (like a wall) is detected, the next (...) (17 years ago, 21-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? —bigboss97
|
| | | | But if the Move block is unlimited will the next block (Wait) be processed? and when it goes to Wait block isn't that mean Move block is done and it will stay in the Wait block for ever, because the vehicle is not moving? (...) (17 years ago, 21-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? —Brian Davis
|
| | | | (...) Yes. Move (unlimited) Starts the motors and *leaves them running* while progressing on to the next block in a sequence. The motor state will then remain unchanged until some other block (like a later Move or Motor block) changes it. Try it an (...) (17 years ago, 21-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor in angle) —bigboss97
|
| | | | Brian, It works fantastic. Thanks. Now I've put 2 US Wait blocks. The 1st one to slow down for < 30cm and the 2nd one to turn for < 20cm. This works pretty well when the vehicle runs ~90 degree toward the wall. But when the angle is flat it only (...) (17 years ago, 22-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor in angle) —Brian Davis
|
| | | | (...) At a glancing angle (not approaching the wall "head on" at 90°, but at an angle) very little sound energy will be reflected back towards the sensor. If there isn't enough energy reflected back for the sensor to "hear" the echo, it won't detect (...) (17 years ago, 22-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor in angle) —bigboss97
|
| | | | Strangely, it slows down, i.e. it detects the further distance (less reflection?) in the same angle. I tried to put the US-sensor vertically. It still shows the same effect. I also notice the closer it comes the larger angle it needs to get a (...) (17 years ago, 23-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor in angle) —bigboss97
|
| | | | (...) (17 years ago, 24-Dec-07, to lugnet.robotics)
|
| | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor in angle) —Brian Davis
|
| | | | | (...) That is a very interesting site, with some interesting information... but you might like to do those experiments yourself. For instance, the odd "ledge" in their data at 48 cm is not a feature with my US sensors - they may have been working (...) (17 years ago, 25-Dec-07, to lugnet.robotics)
|
| | | | | |
| | | | | | Re: NXT Program: How to avoid stop-and-go? (US-Sensor Test Values) —bigboss97
|
| | | | I tried to use a program to log the distance while the US-sensor is moving forward. Unfortunately, it doesn't go straight all the time. So, I simply display the US value and put the vehicle in different position manually. Here are my test values for (...) (17 years ago, 27-Dec-07, to lugnet.robotics)
|
| | | | |
