KR20150053603A - Intelligent surveillance system using robot and method therof - Google Patents

Abstract

The present invention relates to an intelligent surveillance system using a robot and a method thereof, which can improve reliability of target recognition by fusing information received from a plurality of robots. The present invention receives recognition reliability information about at least one target from the robots placed inside a predetermined surveillance area, generates standard reliability information by calculating recognition reliability information of a target which is repeated of the received recognition reliability information, and sets a target to each robot to monitor the related target by a robot wherein standard reliability about each target is the highest in accordance with the generated standard reliability information, thereby improving reliability about target recognition for a multiple robot to be effectively operated in an unmanned and autonomic combat system to improve combat performance.

Description

[0001] INTELLIGENT SURVEILLANCE SYSTEM USING ROBOT AND METHOD THEROF [0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention [0002] The present invention relates to a robot-based intelligent monitoring system and method capable of improving reliability of target recognition by fusing target information received from a plurality of robots.

The intelligent surveillance boundary robot system is in a trend of increasing demand in national infrastructure facilities such as airports, harbors and nuclear power plants due to the rapid growth of the security industry. In particular, the intelligent surveillance boundary robot system is required to improve the efficiency of the normal boundary mission in the military field.

Therefore, various unmanned devices that can perform 3D (Dangerous, Dirty, Dull) missions have been developed and deployed in order to enhance the safety of soldiers and the efficiency of combat. The introduction of unmanned robots based on artificial intelligence technology, As well as to increase military combat power. In particular, the surveillance and boundary robot system can play the most important role in the advancement of military tactics, and it can prevent human fatigue, . In addition, by precisely tracing and interacting with firearms, it is possible to develop high accuracy hitting ability at the exhibition by cultivating the ability to trace and counteract instantaneous.

However, the conventional monitoring boundary system uses a thermal camera or a general surveillance camera to monitor and warn, or to operate a plurality of monitoring boundary systems separately. It is a system for integrating / intelligently recognizing images and automatically tracking It was not a system that could do it, but a system that introduced a manual or rudimentary level of automation.

In addition, when a thermal imaging camera is used in a conventional monitoring system, there is a problem that an unnatural image is input in the daytime because it is an expensive and night-time dedicated camera, and when a general monitoring camera is used, low- There is a problem that is difficult to utilize. In addition, there is a problem in that tracking performance of an object is deteriorated due to inability to intelligently and automatically track the movement of the object.

The background art of the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2005-0056092.

Therefore, in order to effectively combine multiple robots in unmanned and autonomous combat systems to improve combat performance, a technique to improve reliability of target recognition by fusing target information generated by the detection / recognition / tracking process of a unit robot It is a fact that is demanded.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an intelligent surveillance system and method using robots that can effectively combine multiple robots in an unmanned and autonomous combat system to improve combat performance.

It is another object of the present invention to provide a robot-based intelligent monitoring system and method capable of improving reliability of target recognition by fusing target information generated by a detection / recognition / tracking process of a unit robot.

According to an aspect of the present invention, there is provided an intelligent monitoring system using a robot, including: a plurality of targets located within a predetermined monitoring area; A plurality of robots for monitoring a plurality of targets and generating and transmitting identification reliability information for each target using a probability distribution for each distance according to a target type; And an information processing apparatus that processes the identification reliability information of the overlapped target among the identification reliability information transmitted from the plurality of robots and allocates robots having the best identification reliability to each target.

The plurality of targets include an SUV vehicle, a truck, a tank, and a long distance value as a movable object.

Each of the robots represents the presence or absence of a target at a specific distance as a detection probability value by using a detection probability distribution according to the distance in the detection mode. If the probability value exceeds the reference value, the robot changes from the detection mode to the monitoring mode, To generate the identification reliability information for the type of the target to be tracked.

Each of the robots represents the presence or absence of a target as a probability value by using one of four methods of MWIR (Mid-Wave Infrared) optical clock / narrow-angle clock and LWIR (Long-Wave Infrared) optical clock / narrow-angle clock.

The information processing apparatus comprising: an identification reliability information receiver for receiving identification reliability information for each type of target from a plurality of robots; A standard reliability information generation unit for generating standard reliability information by calculating identification reliability information of an overlapping target among the received identification reliability information; A target setting unit for setting target targets to each robot so that the robot having the highest standard reliability for each target monitors the target in the generated standard reliability information; And a target information transmitting unit for transmitting the set target target information to the corresponding robot.

The standard reliability information generator generates standard reliability information by performing an average value, a multiplication value, and a Dempster-Shafer value operation on the received identification reliability information.

According to another aspect of the present invention, there is provided an intelligent monitoring boundary method using a robot, comprising: receiving identification reliability information of at least one target from a plurality of robots located within a predetermined monitoring area; Generating standard reliability information by calculating identification reliability information of an overlapping target among the received identification reliability information; Setting target targets to each robot so that the robot having the highest standard reliability for each target monitors the target according to the generated standard reliability information; And transmitting the set target target information to the corresponding robot.

The identification reliability information indicates whether each robot has a target at a specific distance as a detection probability value by using a probability distribution based on the distance in the detection mode, and then switches to the monitoring mode when the detection probability value exceeds the reference value, And generates the identification reliability information on the type of the target to be tracked using the probability distribution of the distance according to the type of the target.

The detection-versus-probability distribution according to distance in the detection mode is characterized by being one of four methods of MWIR (Mid-Wave Infrared) optical clock / narrow-angle clock and LWIR (Long-Wave Infrared) optical clock / .

The step of generating the standard reliability information generates standard reliability information by calculating at least one of an average value, a multiplication value, and a Dempster-Shafer value with respect to the identification reliability information of duplicated targets among the identification reliability information received from the plurality of robots.

The present invention is based on the idea that by calculating versatile standard reliability information using identification reliability information for a target received from a plurality of robots and setting the target to be monitored for a robot having the highest standard reliability (having the best identification reliability) The reliability of the target recognition can be improved. Accordingly, by improving the reliability of the target recognition by fusing the reliability information of the target generated by the detection / recognition / tracking process of the unit robot, it is possible to effectively combine multiple robots in the print and autonomous combat systems to improve combat performance It is effective.

1 is a block diagram of an intelligent monitoring system using a robot according to the present invention;
2 is a detailed configuration diagram of the information processing apparatus of FIG.
3 is a flow chart of an intelligent monitoring boundary method using a robot according to the present invention.
FIG. 4 is an exemplary view for explaining a detection / recognition / identification probability distribution according to a detection distance of the robot in FIG. 3;

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, wherein like reference numerals are used to designate identical or similar elements, and redundant description thereof will be omitted. The suffix "device" and "equipment" for the components used in the following description are given or mixed in consideration of ease of specification, and do not have their own meaning or role.

1 is a block diagram of an intelligent monitoring system using a robot according to an embodiment of the present invention.

Referring to FIG. 1, a target information processing system using a robot according to the present invention includes a plurality of robots 100 and an information processing apparatus 200.

The plurality of robots 100 are mobile or stationary robots, and monitor at least one target 10 within a predetermined area by using an infrared sensor, a camera, or the like. The target 10 is used to include a mobile object such as an SUV vehicle, truck, tank, and armored vehicle.

Each robot 100 monitors the target 10 in a detection mode and a monitoring mode. The detection mode is a mode for determining the presence or absence of the surrounding target 10 while rotating 360 degrees every predetermined period, for example, in a panoramic manner. The monitoring mode is a mode for tracking the target 10 and predicting the type of the target 10 when the target 10 is present.

In addition, the plurality of robots 100 can switch from the detection mode to the monitoring mode by using the probability distribution of the detection contrast for the target 10 according to the distance. For example, each robot 100 may use any one of the four methods of a Mid-Wave Infrared (" MWIR ") optical clock / narrow clock and a LWIR (Long-Wave Infrared) optical clock / The presence or absence may be represented by a probability value, which may vary depending on the user setting.

Initially, the plurality of robots 100 operate in the detection mode to detect whether the target 10 exists in a predetermined area. When the target 10 to be detected is set to four types of SUV vehicles, trucks, tanks, and armored vehicles, the robot 100 determines the presence or absence of the target 10 through a probability distribution based on the distance. For example, if the detection probability exceeds a preset reference value at a specific distance and the recognition probability for the target 10 is greater than 25%, the robot 100 determines that the predetermined target 10 exists, .

In the monitoring mode, the plurality of robots 100 generate identification reliability information for the type of the target 10 to be tracked using the probability distribution of the distance according to the type of each target 10, 200). For example, the type of the target 10 generated in the robot 100, that is, the identification reliability information for SUV vehicles, trucks, tanks, and armored vehicles may be 0.6, 0.1, 0.2, and 0.1, respectively. From the identification reliability information, it can be known that the current target 10 has the highest probability of being an SUV vehicle.

The information processing apparatus 200 communicates with a plurality of robots 100 via a network and receives identification reliability information about the target 10 from each robot 100. [ In this case, the information processing apparatus 200 can receive the current position information of each robot 100 and the position information of the target 10 as coordinate information, And stores the identification reliability information on the database (DB).

The targets of the target 10 that can detect the positions of the plurality of robots 100 may be different. Therefore, the information processing apparatus 200 generates the standard reliability information by using the identification reliability information of the target 10 common among the identification reliability information of the target 10 received from the plurality of robots.

The information processing apparatus 200 can set a target of monitoring based on the generated standard reliability information and transmit it to each robot 100. Therefore, each of the robots 100 tracks the target to be monitored. The targets to be monitored set to the robots 100 may be the same or different.

Hereinafter, the operation of the information processing apparatus 200 will be described in detail with reference to the accompanying drawings.

2 is a detailed configuration diagram of the information processing apparatus 200 shown in FIG.

2, the information processing apparatus 200 includes a receiving unit 210 for receiving identification reliability information from a plurality of robots 100, a receiving unit 210 for receiving identification reliability information of a target 10 common among the plurality of received identification reliability information, A standard reliability information generation unit 220 for generating standard reliability information using the identification reliability information and a target setting unit 230 for setting a monitoring target to a plurality of robots 100 using the generated standard reliability information, .

3 is a flowchart of an intelligent monitoring boundary method using a robot according to an embodiment of the present invention.

Initially, the plurality of robots 100 operate in a detection mode to detect whether the target 10 exists in a predetermined detection area. When the target 10 to be detected is set to four types of SUV vehicles, trucks, tanks, and armored vehicles, the robot 100 determines whether or not the target 10 is present through a probability distribution based on the distance according to the distance, And switches to the monitoring mode when the target 10 exists in the area.

In the monitoring mode, the plurality of robots 100 generate identification reliability information for the type of the target 10 to be tracked using the probability distribution of the distance according to the type of each target 10, 200 (S310).

The identification reliability information receiving unit 210 of the information processing apparatus 200 receives the identification reliability information for at least one target (e.g., SUV vehicle, truck, tank, and armored vehicle) from the plurality of robots 100 (S320).

The received identification reliability information may be different depending on the position of each robot 100.

The standard reliability information generating unit 220 generates standard reliability information based on the identification reliability information of the overlapped target among the received identification reliability information (S330). The standard reliability information refers to information whose reliability is improved by the route calculated based on the identification reliability information of a plurality of targets generated by the plurality of robots 100. [ That is, when the likelihood that the target to be monitored by the robot 1 is A (e.g., a tank) is high and the probability of the target A to be monitored by the robot 2 is high, the reliability thereof is further improved, Reliability information.

The standard reliability information generation unit 220 generates at least one of an average value, a multiplication value, and a Dempster-Shafer value according to a plurality of standardization rules for the identification reliability information of duplicated targets among the identification reliability information received from the plurality of robots 100 To generate standard reliability information. That is, the standard reliability information generation unit 220 obtains a mean value or a multiplied value between the identification reliability information generated by the robot 1 and the identification reliability information generated by the robot 2, or calculates a fusion value using the Dempster-Shafer theory, Information can be generated. Standard reliability information according to the standardization rules can be expressed as shown in Table 1 below.

는 로봇 i에서 생성한 k개의 타겟에 대한 식별 신뢰도 정보를 나타내고,

는 로봇 j에서 생성한 k개의 타겟에 대한 식별 신뢰도 정보를 나타내며,

는 k개의 타겟에 대한 표준 신뢰도 정보를 나타낸다. In Table 1,

Represents the identification reliability information for the k targets generated by the robot i,

Represents the identification reliability information for the k targets generated by the robot j,

Represents the standard reliability information for k targets.

)와 로봇 j의 식별 신뢰도 정보(

) 간의 평균값을 나타내고, 표준화 규칙 중 곱셈값(ψ2)는 로봇 i의 식별 신뢰도 정보(

)와 로봇 j의 식별 신뢰도 정보(

) 간의 곱셈값을 나타낸다. 또한, 표준화 규칙 중 Dempster-Shafer 값(ψ3)은 Dempster-Shafer 이론에 i의 식별 신뢰도 정보(

)와 로봇 j의 식별 신뢰도 정보(

)를 적용하여 구한 값을 나타낸다. 이 경우, Dempster-Shafer값은 다음의 수학식 1과 같이 표현할 수 있다.Among the standardization rules, the average value (? 1) is the identification reliability information of the robot i

) And the identification reliability information of the robot j (

), And the multiplication value (? 2) of the standardization rules represents the average reliability value of the identification reliability information (

) And the identification reliability information of the robot j (

). ≪ / RTI > In addition, the Dempster-Shafer value (ψ3) among the standardization rules can be used as the Dempster-

) And the identification reliability information of the robot j (

). ≪ / RTI > In this case, the Dempster-Shafer value can be expressed by the following equation (1).

In addition, the standard reliability information generation unit 220 may normalize a plurality of standard reliability information of Table 1 and display it as shown in Table 2 below.

In Table 2, when the multiplication value (? 2) is obtained or the Dempster-Shafer value is obtained in the normalization rule, the value becomes larger when the standard reliability information of the identification reliability information for the common target is normalized.

When the standard reliability information is generated, the target setting unit 230 sets the target of monitoring to the plurality of robots 100 using the generated standard reliability information (S340). That is, the target setting unit 230 of the standard reliability information generating unit 220 allocates targets to be monitored to each robot by setting the target to be monitored for the robot with the highest standard reliability. This is for selecting a robot best suited for monitoring a target among a plurality of robots based on the standard reliability information and intensively monitoring the target to a selected robot. The set target setting information is transmitted to each robot through the target information transmitting unit 240 of the standard reliability information generating unit 220 (S350).

Accordingly, each robot 100 receives information about the target to be monitored by itself from the target setting information transmitting unit 240 of the standard reliability information generating unit 220, and tracks and monitors the movement of the target.

4 is an exemplary diagram for explaining the detection / recognition / identification probability distribution according to the detection distance of the robot.

Referring to FIG. 4, (a) shows a detection / recognition / identification probability distribution according to distance in the case of a MWIR (Mid-Wave Infrared) optical clock, (b) / RTI > < RTI ID = 0.0 > detection / recognition / identification < / RTI >

In this case, the detection probability distribution is a distribution of probabilities that can determine whether or not a target exists depending on distances, and the recognition probability distribution is a distribution of probabilities that can determine the type of the target according to distances. Which means the probability distributions that can identify specific details of the target.

The plurality of robots 100 can detect the probability that the recognition probability for the target located at a certain distance in the initial detection mode for the target using the detection / recognition / identification probability distribution according to the distance is equal to the number of target types If the distribution probability is larger than the distribution probability, the monitoring mode is switched. For example, if the target is a type of four (SUV vehicle, truck, tank and armored vehicle) and the robot uses the probability distribution of the LWIR narrowing of (b), detection of the presence of the target If the probability is 95.50% and the recognition probability of the target is more than 25%, the detection mode is switched to the monitoring mode, and the target is tracked and monitored intensively. In this case, the robot can also track the target using the probability distribution for each type of target even in the monitoring mode.

As described above, according to the present invention, more precise standard reliability information is calculated by using the identification reliability information of a target received from a plurality of robots, and then the target with the highest standard reliability (having the best identification reliability) The reliability of the target recognition can be improved. Accordingly, by improving the reliability of the target recognition by fusing the reliability information of the target generated by the detection / recognition / tracking process of the unit robot, it is possible to effectively combine multiple robots in the unmanned and autonomous battle system to improve combat performance .

The above-described intelligent monitoring boundary robot system and method according to the present invention can be applied to a configuration and a method of the above-described embodiments in a limited manner, but the embodiments can be implemented without changing their technical ideas or essential features It will be understood that the invention may be embodied in other specific forms. Therefore, it should be understood that the above-described embodiments are to be considered in all respects as illustrative and not restrictive.

10: Target 100: Robot
200: target information processing apparatus 210: identification reliability information receiving unit
220: Standard Reliability Information Generation Unit 230:
240: Target information transmission unit

Claims (10)

A plurality of targets located within a predetermined monitoring area;
A plurality of robots for monitoring a plurality of targets and generating and transmitting identification reliability information for each target using a probability distribution for each distance according to a target type; And
And an information processing device for processing the identification reliability information of the redundant target among the identification reliability information transmitted from the plurality of robots and allocating the robot having the optimum identification reliability to each target, . The method of claim 1, wherein the plurality of targets
An intelligent surveillance system using robots, characterized by comprising a SUV vehicle, a truck, a tank, and a long distance value as a movable object. The robot system according to claim 1, wherein each of the robots
The detection probability of the target at a specific distance is expressed as a detection probability value by using the detection probability versus probability distribution according to the distance in the detection mode, and when the probability value exceeds the reference value, the mode is switched from the detection mode to the monitoring mode, And generates identification reliability information on the type of the target to be tracked using the probability distribution. 4. The robot system according to claim 3, wherein each of the robots
Wherein the presence or absence of the target is represented as a probability value by using any one of four methods of MWIR (Mid-Wave Infrared) optical clock / narrow watch and LWIR (Long-Wave Infrared) optical clock / Intelligent surveillance boundary system. The information processing apparatus according to claim 1,
An identification reliability information receiving unit for receiving identification reliability information for each type of target from a plurality of robots;
A standard reliability information generation unit for generating standard reliability information by calculating identification reliability information of an overlapping target among the received identification reliability information;
A target setting unit for setting target targets to each robot so that the robot having the highest standard reliability for each target monitors the target in the generated standard reliability information; And
And a target information transmitting unit for transmitting the set target information to the robot. 6. The apparatus of claim 5, wherein the standard reliability information generator
And generating standard reliability information by performing an average value, a multiplication value, and a Dempster-Shafer value operation on the received identification reliability information. Receiving identification reliability information for at least one target from a plurality of robots located within a predetermined monitoring area;
Generating standard reliability information by calculating identification reliability information of an overlapping target among the received identification reliability information;
Setting target targets to each robot so that the robot having the highest standard reliability for each target monitors the target according to the generated standard reliability information; And
And transmitting the set target target information to the robot. 8. The method of claim 7, wherein the identification reliability information
Each robot displays the presence or absence of a target at a specific distance as a detection probability value by using a detection probability versus a probability distribution according to the distance in the detection mode. Then, when the detection probability value exceeds a reference value, it switches to the monitoring mode, And generating identification reliability information on the type of the target to be tracked using the probability distribution of stars. 9. The method of claim 8, wherein the detection-versus-probability distribution along the distance in the detection mode is selected from the group consisting of four methods: a Mid-Wave Infrared (MWIR) optical clock / Wherein the robot is one of a plurality of robots. 8. The method of claim 7, wherein generating the standard reliability information comprises:
And generating standard reliability information by calculating at least one of an average value, a multiplication value, and a Dempster-Shafer value with respect to identification reliability information of duplicated targets among the identification reliability information received from a plurality of robots. Way.

Images