Show description

5) where beaconr is the actual radius of the beacon in the environment. By similar calculations, we can determine the distance and bearing (and the corresponding variances) of the various objects in the robot’s field of view. 6 VISUAL OPPONENT MODELING Another important task accomplished using the image data is that of opponent modeling. 4, each robot provides a maximum of four best estimates of the opponent blobs based on the current image frame. To arrive at an efficient estimate of the opponents (location of the opponents relative to the robot and hence with respect to the global frame), each robot needs to merge its own estimates with those communicated by its teammates.

Dashed lines show the resulting vectors. ) With the vectors shown, the robot will be turning towards its right as it moves diagonally forward and right. That walk is called SPLINE WALK, while the one being described here is called PARAM WALK. 3 discusses when each of the walks was used. More recent walking engines by some other teams take a similar approach to the ones described in this section, but are somewhat more flexible with regards to providing smooth omnidirectional motion and a set of different locus representations including rectangular, elliptic, and Hermite curve shaped trajectories [2].

To measure forward velocity, we use a stopwatch to record the time the robot takes to walk from one goal line to the other with its forward walking parameters. The time taken is divided into the length of the field, 4200 mm, to yield the forward velocity. The same process is used to measure sideways velocity. To measure angular velocity, we execute the walk with turning parameters. Then we measure how much time it takes to make a certain number of complete revolutions. This yields a velocity in degrees per second.