Navigation method and robot thereof

A robot, comprising:

• a plurality of distance sensors, configured to measure a plurality of to-be-sensed distances in different directions;
• a movement decision-making circuit, configured to repeatedly execute a decision-making process, wherein the decision-making process is to input the sensed distances, a first position coordinate, and first movement information into a neural network model to obtain second movement information output by the neural network model;
• a mobile device;
• and
• a control circuit, configured to drive, based on the second movement information, the mobile device to move from the first position coordinate to a second position coordinate, wherein
• when the movement decision-making circuit initially executes the decision-making process, the first position coordinate is set as an initial coordinate, and the first movement information is set as initial movement information;
• after the decision-making process is executed, the movement decision-making circuit sets the second movement information as the first movement information for a next round of the decision-making process;
• and
• after the control circuit drives, based on the second movement information, the mobile device to move from the first position coordinate to the second position coordinate, the movement decision-making circuit sets the second position coordinate as the first position coordinate of a next round of the decision-making process;
• wherein the first movement information comprises a first two-dimensional linear velocity and a first angular velocity, the movement decision-making circuit normalizes the first two-dimensional linear velocity to fall within an interval (0, 1), and the movement decision-making circuit normalizes the first angular velocity W1 to fall within an interval (−1, 1).