... | ... | @@ -91,8 +91,6 @@ The suppressed functionality has also been implemented so that the behavior can |
|
|
|
|
|
In order to win a sumo battle an offense and a defense strategy needs to be determined. The initial idea for the offense strategy is to place a shovel in front of the LEGO car which goal is to lift the opposing LEGO car and thereby remove its control abilities. By combining the shovel with a sonar sensor the car should be able localize the opponent and drive towards it. The defense strategy involves keeping the LEGO car inside the ring. For this a light sensor is used that can distinguish between the black battle area and the white border. Whenever the border is detected the car will back up and make a turn.
|
|
|
|
|
|
[INDÆST BILLEDE: original setup]
|
|
|
|
|
|
### Software setup
|
|
|
|
|
|
The software for the sumo wrestler LEGO car is build upon the dynamic behavioral control paradigm. Three behaviors are used in this setup as illustrated in the following figure.
|
... | ... | @@ -205,7 +203,7 @@ With the additional light sensor the LEGO is able to stop, back up and turn no m |
|
|
|
|
|
## Conclusion
|
|
|
|
|
|
|
|
|
In this lesson we have developed a sumo wrestler LEGO car with both an offense and defense strategy. By using the behavioral control paradigm we are able to implement multiple behaviors which are triggered by the various sensors attached to the LEGO car. We found out that weight and weight distribution was an important factor in this setting. An equal weight distribution and a LEGO car which weight reached the maximum limit was the optimal solution. Localizing the opponent fast is a crucial task but with only a sonar senor in front the LEGO car, it needs to inspect the entire area around itself. Randomness has been introduced so that the LEGO car performs the inspection in a non-deterministic manner which seemed to be superior according to a deterministic inspection.
|
|
|
|
|
|
## References
|
|
|
|
... | ... | |