KR20110077850A - Method for detecting edge shape using ultrasonic sensor - Google Patents

Abstract

PURPOSE: A method for detecting an edge shape using an ultrasonic sensor is provided to stably perform Simultaneous Localization and Mapping. CONSTITUTION: A method for detecting an edge shape using an ultrasonic sensor is as follows. Data about surrounding objects are collected using a sensor(S101). The union of two data among the collected data is extracted(S102). Imaginary circumscribed circles, which border the beam of the extracted data, are set(S103). Circumscribed circles, in which the data except the extracted data recognize, are compared with the imaginary circumscribed circles and are determined(S104). The determined circumscribed circles are classified(S105). The centers of the classified circumscribed circles are derived(S106).

Description

Edge shape detection method using ultrasonic sensor {METHOD FOR DETECTING EDGE SHAPE USING ULTRASONIC SENSOR}

The present invention relates to a method for detecting an object using an ultrasonic sensor, and more particularly, when data measured by ultrasonic sensors are continuously input, weights are measured in proportion to a shape recognition index. To combine these data to detect an object.

The mobile robot should be able to estimate its own position while creating an environment map if information on a given environment is not given in advance. This problem is known as simultaneous location and mapping. Robots need sensors to observe the environment for simultaneous positioning and mapping. Ultrasonic sensors, as is known, are suitable for commercialization of robots for several reasons. Ultrasonic sensors, however, are difficult to accurately determine the position of an object due to mirror reflection and wide beam width. Therefore, many methods have been proposed to reduce the direction uncertainty and to remove erroneous data for position estimation.

The types of models for creating environmental maps can be broadly divided into geometrical models, topological models, and semantic models. In particular, in the preparation of environmental maps, researches on extracting environmental shapes using a distance sensor such as an ultrasonic ring have been continuously conducted. Crowley first attempted a shape-based approach using ultrasonic sensors by introducing the concept of a composite local model. Leonard and Durrant-Whyte measured the data by rotating a single ultrasonic sensor to form data units called RCDs (Regions of Constant Depth), which extracted line and dot shapes for the experimental environment. Heale and Kleeman developed a scanner-type, real-time DSP ultrasonic echo processor, which measures and analyzes ultrasonic data, extracts line and point shapes to perform simultaneous localization and mapping (SLAM). Was performed.

Wijk and Christensen developed a concept called Triangulation based Fusion (TBF) using ultrasonic data to extract point shapes and estimate the robot's position. The Tardos team used the Hough Transform to classify the ultrasound data in the same parameter domain by interlocking the data measured from the fixed ultrasound sensor ring into a discrete parameter domain. In this study, we extracted the shape for the parameter with the largest number of selections by occlusal ultrasound data in two-dimensional parametric region of line and point shapes. Bank and Kampke arranged a plurality of sensors on the fixed ultrasonic sensor ring so that the directivity angles of the ultrasonic waves overlap each other, and extracted the linear shape by measuring and analyzing not only the reflected wave measured by the transmitting sensor but also the reflected wave measured by the neighboring sensor. Fazli and Kleeman also extracted line and point shapes by measuring and analyzing ultrasonic data using ultrasonic sensor rings with multiple sensors.

Some of the examples above attempted to maximize efficiency by measuring high-density data with just a few sensors using a rotating measurement system of ultrasonic sensors. However, this method is not only a obstacle for the robot to run smoothly because of the relatively long measurement time of the ultrasonic sensor, but also the operation of the device is not practical in terms of commercialization. In addition, most studies have simplified the use of environmental shapes by extracting line or point shapes from ultrasound data. However, in an indoor environment to which a mobile robot can be applied, there may exist a curved surface that is difficult to express only by line and point shapes due to inaccurate ultrasonic data. In addition, the edges tend to be hardly extracted due to the mirror reflection of the ultrasonic sensor. As a result, there is a need for a method that can effectively detect edge shapes using a simple stationary ultrasonic ring system.

In particular, the uncertain and complex situational conditions of the home environment cause confusion in the ultrasound data and make it difficult to recognize the exact position of the object. This phenomenon causes serious problems in recognizing the interior of the actual home environment. In other words, expressing a home environment using a shape of a line or a point extracted from ultrasonic data may cause simultaneous location estimation and mapping failure.

FIG. 1 is a photograph showing a confusion causing situation by linear ultrasonic data, and FIG. 2 is a photograph showing a confusion causing situation of ultrasonic data with respect to a point shape.

Referring to FIG. 1 and FIG. 2, the above-described problem can be considered in two respects as follows: data occlusion and map efficiency.

First, in terms of data occlusion, since the linear shape can provide directionality, the direction error of the mobile robot can be reduced by using the linear shape. However, in fact, a situation that causes confusion can easily occur due to the nature of the ultrasound as shown in FIG. This phenomenon is similar to the situation where the blind detects an object using an umbrella. Therefore, all the shape of the line by ultrasonic data is not helpful for simultaneous location estimation and mapping. This incorrect line shape makes data occlusion very difficult with the already registered surrounding shapes. Point shapes also cause the same problems in complex home environments. Usually, since the ultrasonic beam is easily reflected by the minute gaps present on the wall, meaningless point shapes can be extracted very easily and in large quantities as shown in FIG.

From the standpoint of map efficiency, the phenomena mentioned in this may make the shape map very messy. Experience has shown that even if a robot travels the same area several times over and over again, the environmental map, which consists of ultrasonic lines or point shapes, continues to become more complex and unstable over time. As a result, ever-increasing geometries make computation for simultaneous location and mapping difficult.

The amount of data obtained from the stationary ultrasonic sensor is relatively small and the uncertainty of the direction due to the wide beam width makes it difficult to estimate the exact position of the object. Ultrasonic beams that face the object relatively well detect the object, and many data often give incorrect distance values due to multiple reflections. Therefore, in general, ultrasonic data are accumulated and used to obtain accurate shape information. However, this accumulation is delayed in extracting the shape and estimating the location. Therefore, the stability of simultaneous location estimation and mapping cannot be guaranteed.

The present invention provides a method for stably performing simultaneous location estimation and mapping (SLAM) by extracting data that recognizes edge shapes such as edges from data measured by an ultrasonic sensor to determine edge shapes. Is in.

In the edge shape detection method according to an embodiment of the present invention, a data collection step of collecting data on a surrounding object using a sensor and a data extraction step of extracting two data combinations from the collected data and a beam of data in the data combination Virtual circumscribed circle setting step of extracting virtual circumscribed circle which is in contact with each other, circumscribed circle confirmation step of specifying circumscribed circle by comparing circumscribed circle recognized by data other than extracted data and virtual circumscribed circle, and a plurality of confirmed circled circle A group classification step of classifying, and a representative point forming step of deriving a center of the classified circumscribed circles.

The sensor may be made of an ultrasonic sensor, and the ultrasonic sensor may be mounted on a mobile robot. The data extracting step may include a width division step of dividing a beam width of data, and the group classification step may include a noise removal step of removing circumscribed circles having a lower reliability than a set reliability.

As described above, according to an embodiment of the present invention, the mobile robot can perform simultaneous position estimation and mapping more easily and stably using an ultrasonic sensor in an environment having a strong shape such as an edge.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and the same or similar components are denoted by the same reference numerals throughout the specification.

Figure 3 is a perspective view of a mobile robot equipped with an ultrasonic sensor according to an embodiment of the present invention.

The mobile robot 10 includes a moving part 1 capable of moving by magnetic force and an ultrasonic sensor system 5 mounted on the moving part 1.

The ultrasonic sensor system 5 includes eight ultrasonic sensors mounted along the outer circumference of the supporting plate 4 at intervals of π / 4 with respect to the center of the supporting plate 4 and the supporting plate 4 fixed to the upper portion of the moving part 1. Field 3.

4 is a flowchart illustrating a process of detecting an object using an ultrasonic sensor according to an embodiment of the present invention.

Referring to FIG. 4, the object detection method according to the present exemplary embodiment includes a data collection step (S101) of collecting data on a surrounding object using a sensor and a data extraction step of extracting two data from the collected data ( S102) and the virtual circumscribed circle setting step (S103) for specifying a virtual circumscribed circle in contact with the beam width of the data from the two data and the circumscribed circle recognized by comparing the circumscribed circle recognized by the data other than the extracted data with the virtual circumscribed circle. A circumscribed circle determination step (S104), a group classification step (S105) for classifying a plurality of circumscribed circles determined through the extraction step to the circumscribed circle determination step (S105), and a representative point forming step (S106) for obtaining the center of the circumscribed circle; Include.

First, in the data collection step (S101), data is collected by recognizing an object in front of the plurality of ultrasonic sensors 3 while moving in a set space using the mobile robot 10. In the present embodiment, the shape of the edge is identified and determined as an arbitrary circle. Referring to FIG. 5, a method of collecting data about the circle shape will be described.

Ultrasonic sensors with measurement origins O ₁ and O ₂ show a circle with a center A and a radius R. The ultrasonic sensor emits a fan-shaped sound wave and uses the reflection of the fan-shaped sound wave to diagnose the shape of the object. The data measured at the measurement origin O ₁ is defined as the data whose measurement origin is O ₁ , the radius is Z ₁ , and the direction angle from the measurement origin (O ₁ ) to the circumscribed circle is Φ ₁ , and the data measured at the measurement origin O ₂ Is defined as data with the measurement origin O ₂ , the radius Z _2, and the direction angle Φ ₂ . At this time, the distance between the measurement origin (O _{1 ,} O ₂ ) is d.

The hatched effective direction angle area S refers to an area that simultaneously satisfies the effective direction angle range of the ultrasonic data generated at the respective midpoints O ₁ and O ₂ . When two ultrasound data originate from one circle shape, the direction angles Φ ₁ , Φ ₂ up to the circumscribed circle midpoint are included in the effective direction angle region S, so the direction angles Φ ₁ , Φ ₂ are circular. In the triangle formed by the center (A) of and the measurement points (O ₁ , O ₂ ) can be defined by Equation 1 by the second law of cosine.

[Equation 1]

for ˚,

for ˚,

과

는 i번째 초음파 데이터의 각각 최소 및 최대 유효 지향각을 나타내며 [수학식 2]와 같이 나타낼 수 있다.In Equation 1

and

Represents the minimum and maximum effective directivity angles of the i-th ultrasound data, respectively, and can be expressed as shown in [Equation 2].

[Equation 2]

는 i번째 초음파 데이터의 유효 지향각에 대한 중심각을 나타내며 ω는 초음파 데이터의 유효 빔 폭이다. 물리적으로 초음파 데이터의 유효 빔 폭은 탐지 거리와 물체의 종류에 따라 정의될 수 있지만, 실제 적용 시에는 초음파 센서에 의한 물체 탐지 거리에 따른 편차와 물체 표면 특성에 따른 영향이 크지 않기 때문에 탐지 거리와 물체의 종류에 상관 없이 사용되는 초음파 센서의 특성에 따라 상수 값으로 설정한다.In [Equation 2]

Is the center angle with respect to the effective directivity angle of the i-th ultrasound data, and ω is the effective beam width of the ultrasound data. Physically, the effective beam width of the ultrasonic data can be defined according to the detection distance and the type of the object. However, in actual application, the detection distance and the detection distance and the influence of the object surface characteristics are not significant because of the variation of the object detection distance by the ultrasonic sensor. Regardless of the type of object, set the constant value according to the characteristics of the ultrasonic sensor used.

According to Equation 1 and Equation 2, the i-th data on the circular shape may be represented by a radius Z ₁ and Z ₂ and a direction angle φ ₁ and φ ₂ , respectively. Therefore, when the direction angles φ1 and φ2 satisfy the effective beam width conditions of the respective ultrasonic data, it can be determined that the data can be used morphologically.

As described above, a fan-shaped beam circumscribed to a circular shape having a radius and a direction angle can be derived through one data. Accordingly, a virtual circumscribed circle circumscribed to the data can be set through two data.

The data extraction step S102 extracts any two data combinations from the collected data. In this case, any two data combinations may be extracted from tens or hundreds of combinations. The number of data combinations to be extracted may be variously set in consideration of required accuracy and arithmetic processing speed. The data extraction step S102 includes a width division step of dividing the beam width of the data. The collected data is set to have a sector-like improvement as described above, so that the beam width of the data can be divided into several or dozens in order to extract more specific circumscribed circles. By dividing the width of the data in this way, the circumscribed circle can be extracted more easily.

The virtual circumscribed circle setting step S103 extracts a virtual circumscribed circle circumscribed with a beam of data from the extracted data combination. As described above, since the data has a direction angle and a radius, the circumscribed circle circumscribed to two fan-shaped arcs can be extracted through the two data.

6 is a view for explaining a process of determining the circumscribed circle by using any three pieces of data.

Three ultrasonic data are considered to find the radius of the virtual circle shape. F ₁ ^θ is an arbitrary angle obtained by subdividing the effective beam width of the ultrasonic data F ₁ . The radius R of the virtual circle and the arbitrary angle F ₂ ^θ of the ultrasonic data F ₂ can be calculated in the relation between F ₁ and F ₂ as described above.

The circumscribed circle determination step (S104) compares the circumscribed circle recognized by the data other than the extracted data with the virtual circumscribed circle to determine whether there is a coincidence.

As shown in FIG. 6, if the arbitrary angle F ₂ ^θ satisfies the effective beam width condition, it is determined whether the virtual circle is morphologically compatible with the ultrasonic data F ₃ . If the arbitrary angle F ₃ ^θ satisfies the effective beam width condition of the ultrasonic data F3, this virtual circle is stored in the storage.

The ultrasound data accumulated in the storage space may include three data sets that detect the same edge.

For example, FIG. 7A shows ultrasound data accumulated in an arbitrary space. You can also see circular groups through corner circles.

In the group classification step S105, the determined plurality of circumscribed circles are classified into similar groups. The circumscribed circles specified among a plurality of data combinations are classified among similar circumscribed groups.

As shown in FIG. 7B, they are classified into each group through simple morphological conditions. Classify the other circles that occupy the circle around the largest circle in the region into a group. In FIG. 7B, circumscribed circles of group 1 are formed in the corner portion, and circumscribed circles of group 2 are formed in an object installed at the portion spaced from the corner.

The group classification step S105 includes a noise removing step of removing circumscribed circles having a reliability lower than the set reliability. The number of prototypes in a classified group is assessed by the reliability of that group, and groups with a confidence lower than the set point are removed. At this time, the evaluation of reliability can be made based on various items such as positional concept such as center of circumscribed circle and improved concept such as over radius.

Representative point forming step (S106) derives the center of the classified circumscribed circle to form a representative point. The representative point consists of the average center position of the circumscribed circle forming the group. As illustrated in FIG. 7C, representative points ρ1 and ρ2 may be formed through a plurality of circumscribed circles forming a group.

FIG. 8A is a photograph illustrating an ultrasound trace taken by a robot while moving a first space and a moving path of the robot by a method according to an embodiment of the present invention, and FIG. 8B is a method by the method according to an embodiment of the present invention. It is a photograph showing the ultrasonic trace taken while moving the second space and the movement path of the robot.

In addition, Figure 9a is a picture showing the movement path of the robot and the edge shape measured by the robot initially in the first space, Figure 9b is a picture showing the movement path of the robot and the edge shape measured by the robot initially in the second space to be.

The first space is a 10ⅹ8m home environment test bed, and the second space is a 12ⅹ10m home environment test bed. The first space and the second space are experimental environments having three rooms, a living room, and a kitchen.

The dotted line represents the actual environment, and the solid line represents the robot's movement path. In addition, a dot represents the representative point of a corner. Corners include not only the corners of the building's internal structure, but also objects installed inside the building.

8A, 8B, 9A and 9B show ultrasound data measured during the first 12 minutes. The robot passively traveled through the home environment at a speed of 0.2 m / s, and the measurement frequency of the ultra-amp sensor was 1 Hz while driving. Corner circlering to extract circumscribed circle was performed every 15 steps of the robot and the distance between each step is 2cm.

The driving path of the robot indicated by the thick line is seriously off. The gray arc represents the edge of the ultrasonic beam. As expected, the ultrasound data looks very confusing due to mirror reflection and wide beam width. In addition, it can be seen that the representative point of the corner is also incompletely generated.

Figure 10a is a picture showing the movement path of the robot and the edge shape measured by the robot after the entire driving in the first space, Figure 10b is a moving path and the edge shape measured by the robot after the entire driving in the second space The picture shown.

As shown in FIGS. 10A and 10B, 26 center points were generated in FIG. 10A and 18 center points were created in FIG. 10B. 10A and 10B, it was confirmed that simultaneous position estimation and mapping were performed stably for a long time in a complicated home environment, so that a robust shape can be effectively extracted.

Thus, despite the mirror reflection phenomenon by the method according to the present embodiment it can be seen that the robot grasped the corner shape almost the same as the actual, and moved along the set driving path.

Although the preferred embodiments of the present invention have been described above, the present invention is not limited thereto, and various modifications and changes can be made within the scope of the claims and the detailed description of the invention and the accompanying drawings. Naturally, it belongs to the scope of the invention.

1 is a diagram illustrating a situation of causing confusion caused by linear ultrasonic data.

FIG. 2 is a diagram illustrating a confusion causing situation of ultrasonic data with respect to a point shape. FIG.

Figure 3 is a perspective view of a mobile robot equipped with an ultrasonic sensor according to an embodiment of the present invention.

4 is a flowchart illustrating a process of detecting an object using an ultrasonic sensor according to an embodiment of the present invention.

FIG. 5 is a diagram for describing a morphological relationship between two single ultrasound data with respect to a circular shape.

6 is a view for explaining a process of determining the circumscribed circle by using any three pieces of data.

Figure 7a is a photograph showing the ultrasonic beam for the edge shape, Figure 7b is a photograph showing the circumscribed circle is classified into groups, Figure 7c is a photograph showing that the representative point of the circumscribed circle is formed.

FIG. 8A is a photograph illustrating an ultrasound trace taken by a robot while moving a first space and a moving path of the robot by a method according to an embodiment of the present invention, and FIG. 8B is a method by the method according to an embodiment of the present invention. It is a photograph showing the ultrasonic trace taken while moving the second space and the movement path of the robot.

FIG. 9A is a photograph showing the movement path of the robot and the edge shape measured by the robot initially in the first space, and FIG. 9B is a photograph showing the movement path of the robot and the edge shape measured by the robot initially in the second space.

Figure 10a is a picture showing the movement path of the robot and the edge shape measured by the robot after the entire driving in the first space, Figure 10b is a moving path and the edge shape measured by the robot after the entire driving in the second space The picture shown.

<Description of the symbols for the main parts of the drawings>

10: mobile robot 1: moving unit

3: ultrasonic sensor 4: support plate

5: ultrasonic sensor system

Claims (5)

A data collection step of collecting data about the surrounding object using a sensor; A data extraction step of extracting two data combinations of the collected data; A virtual circumscribed circle setting step of extracting a virtual circumscribed circle in contact with the beam of data in the data combination; A circumscribed circle determining step of comparing circumscribed circles recognized by data other than the extracted data with virtual circumscribed circles to specify a matching circumscribed circle; A group classification step of classifying the determined plurality of circumscribed members into similar groups; And Representative point forming step of deriving the center of the classified circumscribed circle; Edge shape detection method comprising a. The method according to claim 1, The sensor is an edge shape detection method of the ultrasonic sensor. The method according to claim 1, The ultrasonic sensor is a corner shape detection method mounted on a mobile robot. The method according to claim 1, The data extraction step includes a width division step of dividing the beam width of the data. The method according to claim 1, The group classification step may include a noise removing step of removing a circumscribed circle having a lower reliability than a set reliability.

Images