Obstacle Detector (Initial) Design

Assumptions

Naming conventions in algorithm documentation below do not attempt to precisely model expected naming conventio/cygdrive/c/cygwin64/home/svigi/Proj/BusLogic ns in code. They are meant to be a guide as to what class, attribute and method names may be.

maxDiff = 10 degrees ‘largest angle between features all on the same obstacle.

minFeaturesInObj = 3 This is the smallest number of clumped features to be

Considered an obstacle

featureCount = 1 Minimum assumed feature count

Initializations

currentFeatureID = 0 // Assume valid data input

lastObstacle = false // Used by iterator to signal that no further obstacles are to be found.

Top Level Algorithm

IDFirstObstacle(Features)

If CurrentFeatureID == -1, quit

lastObstacle = false

While (isFalseLastLastObstacle())

IDNextObstacle(Features)

obstacleHelper.Cout()

For (i=0; i>numObstacles; i++)

CoutObstacle()

CLASSES:

Class Features

Attributes:

Member functions:

Class Obstacles

IDFirstObstacle() This might be same as IDNextObstacle, but not sure yet.

// currentFeatureID = 0 Assume initialized and valid data input

// featureCount = 1 Assumed for first ID

leftEdgeID = currentFeatureID

nextFeatureID = currentFeatureID + 1 // Assume valid

While (nextFeatureInRange) {

If nextFeatureInRange

RightEdgeID = CurrentFeatureID

FeatureCount++

currentFeatureID++

If featureCount > minFeaturesInObj, set thisIsAnObj = true

ComputeMidpoint()

// Done with this obstacle. Next, we move on to evaluate for next feature and to

// see if possibly it is left edge of the next obstacle

IDNextObstacle()

// General: Cycle through the features in order to identify and locate the next obstacle.

// currentFeatureID // Available from Features array. Assume valid data.

// featureCount Currently valid in features obstacle

leftEdgeID = currentFeatureID // Candidate left edge

// nextFeatureID = currentFeatureID + 1 // Assume set and valid by nextFeatureInRange()

While (nextFeatureInRange) {

If nextFeatureInRange

RightEdgeID = CurrentFeatureID

FeatureCount++

currentFeatureID++

If featureCount > minFeaturesInObj, set thisIsAnObj = true

// Done with this obstacle. Next, we move on to evaluate for next feature and to

// see if possibly it is left edge of the next obstacle

// Todo: Handle condition where we run out of Features. Then, processing ends.

// Todo: Handle condition where an obstacle spans the perimeter at the boundary.

FirstObstacle()

CurrentObstacle = 0

NextObstacle()

If CurrentObstacle+1 invalid, set lastObstacle = true

Else CurrentObstacle++

LastObstacle()

Return lastObstacle boolean, true if last obstacle has been detected.

nextFeatureInRange()

return “quit” if nextFeatureValid is false

thetaDiff = data[currentFeatureID+1] - data[currentFeatureID]

result = (thetaDiff <= maxDiff)

return result

LastObstacle()

Return true if nextFeature fails

else return false

ComputeMidpoint()

ObstacleMidpoint = feature[currentFeatureID].right edge - left edge

- Todo, compute as average of theta values for all points from left edge to right edge.