Method for controlling mobile robot based on Bayesian network learning

1. A method for controlling the navigation of the mobile robot, in which a control unit applies a drive control signal to a robot driving unit to control the navigation of the mobile robot using sensing information of a sensing unit and preset data that is stored in a storage unit and includes grid information for a navigation environment, the method comprising:

a path planning step of calculating a navigation path of the mobile robot from a departure point to a destination point based on congestion map information calculated using grid map information for a navigation environment included in the preset data and the sensing information;

a navigation step of acquiring navigation environmental information including obstacle map information using the sensing information sensed by the sensing unit in a process in which the mobile robot moves along the navigation path calculated in the path planning step, and allowing the mobile robot to navigate from the departure point to the destination point by performing a position estimation operation using the acquired navigation environmental information; and

a congestion map building and calculating step of building a congestion map for the navigation environment using a navigation environment database having stored therein navigation cost information including information regarding a probability of existence or absence of an obstacle in a relevant grid in the navigation environmental information acquired in the navigation step after the completion of the navigation step,

wherein the navigation cost includes information regarding a static obstacle that is in a fixed position in the navigation environment and information regarding a dynamic obstacle that is moveable in the navigation environment, and

wherein building the congestion map includes calculating the probability of the presence or absence of the dynamic obstacle.

2. The method according to claim 1 , wherein the navigation cost information included in the navigation environment database comprises:

intrinsic cost information stored by converting the information regarding the static obstacle into a cost;

congestion cost information stored by converting the information regarding the dynamic obstacle into a cost; and

adjacency cost information stored by assigning an adjacency cost to the number of grids on a minimum movement path from the departure point to the destination point 4 based on the departure point and the destination point being set.

3. The method according to claim 2 , wherein the path planning step comprises:

a navigation cost calculation step of performing an arithmetic operation on the navigation cost information from the departure point to the destination point in response to an arithmetic control signal from the control unit to calculate the navigation cost; and

a navigation path calculation step of calculating the navigation path from the departure point to the destination point using a path calculation method in which the navigation cost calculated in the navigation cost calculation step is previously set and stored.

4. The method according to claim 3 , wherein the navigation cost calculation step comprises calculating the navigation cost using the congestion cost information calculated from a Bayesian network employing Bayes' theorem based on the navigation environment database.

5. The method according to claim 4 , wherein the navigation environment database comprises obstacle map information including information regarding a probability of existence or absence of an obstacle in the grid according to temporal information.

6. The method according to claim 5 , wherein the temporal information comprises information regarding weekday, date and time, and

wherein a unit number of the temporal information influenced by different temporal information with respect to the same grid in the Bayesian network is set as an influence time unit number, and the preset data comprises the influence time unit number.

7. The method according to claim 6 , wherein the navigation cost calculation step comprises:

a congestion cost calculation step of calculating a congestion cost using the congestion cost information for the grid, which is calculated through the Bayesian network according to the influence time unit number;

an intrinsic cost calculation step of calculating an intrinsic cost using the intrinsic cost information for each grid included in the grid information;

an adjacency cost calculation step of calculating an adjacency cost using the adjacency cost information stored by assigning an adjacency cost to the number of grids on a minimum movement path from a departure point to a destination point if the departure point and the destination point are set; and

a navigation cost calculation step of calculating a navigation cost using the congestion cost, the intrinsic cost and the adjacency cost, which are calculated for the grid.

8. The method according to claim 7 , wherein the path calculation method is a minimum path gradient method, and comprises the navigation path calculation step of calculating the navigation path formed at a minimum cost from the departure point to the destination point using a navigation cost for each grid, which is calculated in the navigation cost calculation step.

9. The method according to claim 8 , wherein the navigation step comprises:

a sensing step of sensing information for the navigation environment through the sensing unit;

a position estimation step of allowing the control unit to estimate the current position of the mobile robot using the sensing information and the preset data;

an obstacle map calculation step of calculating and updating obstacle information for the navigation environment and calculating the obstacle map information using the current position of the mobile robot obtained in the position estimation step and the sensing information;

a movement control step of allowing the control unit to apply a movement control signal to a robot driving unit of the mobile robot to cause the mobile robot to move to the destination point;

a navigation completion confirmation step of confirming whether or not the mobile robot moving through the movement control step reaches the destination point; and

a navigation environment database update step of updating the navigation environment database using the obstacle map information if it is determined in the navigation completion confirmation step that the navigation of the mobile robot has been completed.

10. The method according to claim 9 , wherein the temporal information is information regarding date and time, and

wherein the congestion map building and calculating step comprises:

a navigation environment database selection step of selecting and calculating a navigation environment selection database having information regarding a probability of existence or absence of an obstacle in the grid according to temporal information with respect to a current time and a current date from the navigation environment database using the temporal information and the influence time unit number;

a Bayesian network calculation step of constructing a Bayesian network and performing an arithmetic operation on the Bayesian network using the navigation environment selection database, and calculating the probability of existence or absence of an obstacle in the grid with respect to the current time and the current date; and

a congestion map building step of building or updating a congestion map using the probability of existence or absence of an obstacle in the grid, obtained in the Bayesian network calculation step.

11. The method according to claim 9 , wherein the Bayesian network calculation step comprises:

a Bayesian network construction step of constructing the Bayesian network using the navigation environment selection database; and

a Bayesian network arithmetic step of calculating the probability of existence or absence of an obstacle in the grid with respect to the current time and the current date using the Bayesian network constructed in the Bayesian network construction step.