Method for building a map of probability of one of absence and presence of obstacles for an autonomous robot

An autonomous robot, comprising:

• a plurality of directional distance sensors;
• a Robot kinematic joint model;
• an on-board memory to store a plurality of initial maps defined around said robot, said plurality of initial maps having predefined limits and being paved by pixels of a predefined dimension, where a value of a probability of at least one of presence or absence of an obstacle is stored;
• a module for initializing said plurality of initial maps by setting the value of each pixel to a predefined value;
• a module for acquiring, from at least one of said directional distance sensors, data representative of at least one of absence and presence of an obstacle in the environment of the robot;
• a module for concurrently applying, for the plurality of said directional distance sensors, to an initial map in the plurality of initial maps the followings steps:
• modifying values of probabilities of absence or presence of obstacle to a value closer to a predefined value;
• and
• updating values of probabilities of at least one of absence and presence of an obstacle in at least one pixel of the initial map from said data;
• and
• a module for, after updating, merging the plurality of initial maps by a threshold Tsobs for delimiting obstacle and unknown pixels and a threshold Tsfree for delimiting obstacle free and unknown pixels into a fused map by:
• setting the value of each pixel whose probability of absence of obstacle is below the threshold Tsobs in at least one of the initial maps to a value representative of an obstacle in the fused map;
• and
• setting the value of each pixel which is not an obstacle pixel in the fused map and whose probability of absence of obstacle is above the threshold Tsfree in at least one of the initial maps to a value representative of an obstacle free pixel in the fused map.

A non-transitory computer readable medium, comprising code that when executed by a computer cause the computer to automatically implement a method of determining a pixel map of the probability of at least one of absence and presence of an obstacle in the environment of the robot, said method comprising:

• initializing, in a memory, a plurality of initial maps defined around an autonomous robot, said plurality of initial maps having predefined limits and being paved by pixels of a predefined dimension, where a value of a probability of at least one of absence and presence of an obstacle in each pixel is set to a predefined value;
• acquiring, from at least a sensing procedure of a first type, data representative of at least one of absence and presence of an obstacle in the environment of the robot;
• concurrently applying, to an initial map in the plurality of initial maps the followings steps:
• modifying values of probabilities of absence or presence of obstacle to a value closer to a predefined value;
• and
• updating values of probabilities of at least one of absence and presence of an obstacle in at least one pixel of the initial map from said data;
• and
• after updating, merging the plurality of initial maps by a threshold Tsobs for delimiting obstacle and unknown pixels and a threshold Tsfree for delimiting obstacle free and unknown pixels into a fused map by:
• setting the value of each pixel whose probability of absence of obstacle is below the threshold Tsobs in at least one of the initial maps to a value representative of an obstacle in the fused map;
• and
• setting the value of each pixel which is not an obstacle pixel in the fused map and whose probability of absence of obstacle is above the threshold Tsfree in at least one of the initial maps to a value representative of an obstacle free pixel in the fused map.