Robot, system and method detecting and/or responding to transitions in height

A system comprising:

• a robot adapted to travel,
• a sensor adapted to sense information of surroundings of the robot, and
• at least one processor configured to:
• generate a three-dimensional data set corresponding to a heightmap of the surroundings of the robot based on the information of surroundings of the robot,
• detect a transition between different height levels in the three-dimensional data set, and
• categorize the transition between the different height levels using at least one characteristic,
• wherein the robot is adapted to perform at least one response action, which response action depends on a categorization of the transition,
• wherein the at least one processor is configured to detect a transition between different height levels in the three-dimensional data set by:
• (A) allocating data points in the three-dimensional data set to different bins, depending on the height of the data points;
• (B) identifying two bins with the most data points;
• and
• (C) selecting, on each of the two bins with the most data points, identified in (B), a respective first subset of data points from data points lying within a predetermined distance from the respective mean value of the data points within the respective bin;
• (D) generating a first estimation of said transition using a first classifier algorithm performed on the first subsets of data points;
• and
• (E) using the two bins identified in (B), refining the first estimation by:
• (E)(i) selecting, on each of the two bins with the most data points, identified in (B), a respective second subset of data points from data points lying within a predetermined distance from the first estimation;
• and
• (E)(ii) generating a second estimation of said transition using a second classifier algorithm performed on the second subsets of data points.

A data processing system adapted to detect a transition between different height levels in a three-dimensional data set, the data processing system comprising at least one processor, wherein the data processing system is configured to:

• (a) allocate data points in the three-dimensional data set to different bins, depending on the height of the data points;
• (b) identify two bins with the most data points;
• and
• (c) select, on each of the two bins with the most data points, identified in (b), a respective first subset of data points from data points lying within a predetermined distance from the respective mean value of the data points within the respective bin, and
• (d) generate a first estimation of said transition using a first classifier algorithm performed on the first subsets of data points, and
• (e) using the two bins identified in (b), refine the first estimation by:
• (e)(i) selecting, on each of the two bins with the most data points, identified in (b), a respective second subset of data points from data points lying within a predetermined distance from the first estimation;
• and
• (e)(ii) generating a second estimation of said transition using a second classifier algorithm performed on the second subsets of data points.