Below are the current list of algorithms offered on the SWARM Simulation Platform.
Description
Given a point cloud, detect the predicted slope of the ground and use that slope to avoid hitting the ground.
How It Works
A goal point is provided to the algorithm. At each iteration, the point cloud is provieded as a N x M channel matrix, where N is the number of channels in the LiDAR and M is the number of points in each channel.
The algorithm uses the top row of the matrix and chooses the point in the middle of the row. Then, the algorithm chooses the point in the middle of the bottom row. The slope of the ground is then calculated.
If the slope that is calculated falls within the set of parameters, the vehicle sets the next point to be the goal point provided but modifies the altitude of the agent to follow the gradient of the slope.
For downward slopes, the vehicle uses a trick where when we were detecting a slope but now don't, then we force the agent to move slightly up before being allowed to go down.