KR101664575B1 - Method to obstacle avoidance for wheeled mobile robots - Google Patents

Abstract

According to an embodiment of the present invention, there is provided an obstacle avoidance system for a mobile robot, comprising: a sensing unit for sensing an obstacle located around a mobile robot traveling toward a target point; A position acquiring unit capable of acquiring position information of the mobile robot, an obstacle, and the target point; A signal generator for generating a switching signal in consideration of a distance between the mobile robot and the obstacle and a relative angle when the obstacle is detected by the sensing unit; A vector computing unit for computing a repulsive force vector between the mobile robot and the obstacle and a force vector between the mobile robot and the target point through position information obtained from the position obtaining unit; And a path determination unit for determining a movement path of the mobile robot using the avoidance vector, the attraction vector, and the switching signal, and the traveling of the mobile robot can be controlled in real time by the switching signal.

Description

TECHNICAL FIELD [0001] The present invention relates to an obstacle avoidance system and method for a mobile robot,

BACKGROUND OF THE INVENTION 1. Field of the Invention [0002] The present invention relates to an obstacle avoidance method of a mobile robot, and more particularly, to a mobile robot obstacle avoidance method that enables a mobile robot to avoid a obstacle stably by generating a switching signal in real time.

 Obstacle avoidance has been studied by many researchers for a long time, but it is still active research area and it is one of the main issues of mobile robot for industrial practical use, safe motion planning and autonomous driving. The obstacle avoidance has been studied for a long time, such as the artificial potential field (APF) approach, the road map approach, and the behavior based approach.

The potential field technique is quick to implement and demonstrates acceptable results in many situations.

However, since the motion and control inputs of the robot are basically determined in the direction of lowering the gradient of the artificial potential field (APF) generated by the target points and obstacles, the result of the local minimum ≪ / RTI > Moreover, when the robot moves close to obstacles or long and narrow passages, it causes chattering and oscillation problems in terms of control. To overcome this oscillation, the sampling period of the robot controller must be lowered or the robot must move at a lower speed. This may take a longer time to complete the obstacle avoidance.

 Another well-known obstacle avoidance technique is the roadmap technique. The roadmap consists of all the segments in the free-form space (C-space) that connect the vertices of each other but not the obstacle. Several techniques for modeling freeform space have been studied.

For example, the visibility graph technique provides a minimal length path solution as a result, but it does not guarantee an optimal solution in a higher dimension space. And the voronoi diagram maximizes the free space between the path in space and the obstacle, thus creating a safe path. Most of these techniques require knowing the entire spatial information in advance and also require considerable computational complexity or time.

Finally, the behavior-based approach is a technique to solve the obstacle avoidance problem based on the convergence of different behaviors such as avoiding the obstacle or moving to the target point. The main issue of behavior-based approaches is how to effectively combine or mediate different types of behaviors in order to achieve good performance.

The obstacle avoiding method of the mobile robot has been studied in various ways. For example, in the prior art document 2012-0027794 filed on March 19, 2012, a method of avoiding obstacles of the mobile robot is disclosed.

The object of the present invention is to provide a stable switched-system approach for obstacle avoidance of a mobile robot, which can perform stable switching between two situations by separately considering a situation in which there is an obstacle and a situation in which there is no obstacle The present invention also provides a system and method for avoiding obstacles of a mobile robot.

An object of the present invention is to provide a system and method for avoiding obstacles of a mobile robot that can avoid static circular or non-circular obstacles and prevent mutual collision through coordination between robots.

An object of the present invention is to provide a mobile robot capable of avoiding autonomous obstacles of a mobile robot and capable of obtaining robustness and reliability of the algorithm, which is one of the main issues of a mobile robot for industrial practical use, safe motion planning, And an obstacle avoidance system and method for a mobile robot.

An object of an embodiment is to provide a system and method for avoiding an obstacle of a mobile robot capable of generating different switching signals according to the shape of an obstacle.

An object of the present invention is to provide a system and a method for avoiding obstacles of a mobile robot that can prevent chattering or oscillation that may occur when an obstacle is avoided by a mobile robot and can shorten the time required for obstacle avoidance .

An object of an embodiment is to provide a system and method for avoiding obstacles of a mobile robot in which obstacle avoidance can be performed in a relatively simple manner through obstacle recognition around the mobile robot.

According to an aspect of the present invention, there is provided an obstacle avoidance system for a mobile robot, including: a sensing unit for sensing an obstacle located around a mobile robot traveling toward a target point; A position acquiring unit capable of acquiring position information of the mobile robot, an obstacle, and the target point; A signal generator for generating a switching signal in consideration of a distance between the mobile robot and the obstacle and a relative angle when the obstacle is detected by the sensing unit; A vector computing unit for computing a repulsive force vector between the mobile robot and the obstacle and a force vector between the mobile robot and the target point through position information obtained from the position obtaining unit; And a path determination unit for determining a movement path of the mobile robot using the attraction vector and the switching signal, wherein the traveling of the mobile robot can be controlled in real time by the switching signal.

According to one aspect of the present invention, in the vector operation unit, a avoidance vector is further calculated from the repulsive force vector, and the avoidance vector may have an inner product with the attracting vector equal to or greater than zero.

According to one aspect of the present invention, in the path determination unit, an unified velocity vector indicating a path through which the mobile robot is to be moved to avoid the obstacle is calculated,

The integrated velocity vector is calculated by the following equation,

는 통합 속도 벡터이고,

는 스위칭 신호이고,

는 회피 벡터이고,

는 인력 벡터일 수 있다.At this time,

Is an integrated velocity vector,

Is a switching signal,

Is an avoidance vector,

May be a force vector.

According to one aspect, a virtual target point is calculated from the integrated velocity vector and the position information of the mobile robot in the path determination unit, and the velocity information calculated from the virtual target point can be corrected.

According to one aspect of the present invention, the switching signal includes a first switching signal generated upon detection of a circular obstacle,

The first switching signal is determined as follows,

는 제1 스위칭 신호이고, d_ro는 상기 모바일 로봇과 상기 장애물 사이의 거리이고, θ는 상기 모바일 로봇의 주행각도(heading angle)이며, At this time,

D _ro is a distance between the mobile robot and the obstacle, θ is a heading angle of the mobile robot,

이고,The d _m

ego,

In this case, R is the sensing range of the sensing unit, r is the avoidance range, and? _Ro may be the difference between the traveling angle of the mobile robot and the obstacle azimuth angle.

According to one aspect of the present invention, the second switching signal for compensating the operation of the mobile robot by the first switching signal is determined as follows,

는 제2 스위칭 신호이고,

는 상기 모바일 로봇의 현재 스위칭 신호일 수 있다.At this time,

Is a second switching signal,

May be the current switching signal of the mobile robot.

According to one aspect of the present invention, the switching signal includes a third switching signal generated upon detection of the non-circular obstacle,

The third switching signal is determined as follows,

는 제3 스위칭 신호이고,

는 상기 모바일 로봇과 상기 목표 지점을 잇는 선이고,

는 상기 장애물 상의 모든 점의 집합일 수 있다.At this time,

Is a third switching signal,

Is a line connecting the mobile robot and the target point,

May be a set of all points on the obstacle.

According to another aspect of the present invention, there is provided an obstacle avoidance method for a mobile robot, including: obtaining an initial state of a mobile robot; Obtaining an environment of the mobile robot; Detecting an obstacle located around the mobile robot; Obtaining a position of an obstacle around the mobile robot and a position of a target point toward the mobile robot; Generating a switching signal by considering a distance between the mobile robot and the obstacle and a relative angle; And calculating a repulsive vector formed vertically from the repulsive force vector and a repulsive force vector between the mobile robot and the obstacle, a attracting vector between the mobile robot and the target point, and the repulsive force vector, The running of the robot can be controlled in real time.

According to one aspect, after the step of calculating the repulsive vector formed vertically from the repulsive force vector between the mobile robot and the obstacle, the attracting vector between the mobile robot and the target point, and the repulsive force vector, An attraction vector, and an integrated velocity vector representing a path through which the mobile robot is to be moved by the switching signal in order to avoid the obstacle.

According to one aspect of the present invention, after the step of calculating the integrated velocity vector indicating the path by which the mobile robot is to be moved by the mobile robot in order to avoid the obstacle by the avoiding vector, the attraction vector and the switching signal, Calculating a target point and a target direction of the target point; Calculating a target speed from the virtual target point; Correcting the target speed; And moving the mobile robot according to the corrected target speed.

According to the obstacle avoidance system and method of the mobile robot according to the embodiment, it is possible to perform stable switching between the two situations by separately considering the obstacle-free situation and the obstacle free situation, A system-switched approach (Stable Switched-System Approach) can be implemented.

According to the obstacle avoidance system and method of the mobile robot according to one embodiment, it is possible to avoid static circular or non-circular obstacles and to prevent mutual collision through cooperation between the robots.

According to the obstacle avoidance system and method of the mobile robot according to the embodiment, autonomous obstacle avoidance of the mobile robot is enabled, and the algorithm, which is one of the main issues of the mobile robot for industrial practical use, safe motion planning, Robustness and reliability can be obtained.

According to the obstacle avoidance system and method of the mobile robot according to one embodiment, it is possible to prevent chattering or oscillation that may occur in obstacle avoidance of the mobile robot, and shorten the time required to avoid an obstacle.

According to the obstacle avoidance system and method of the mobile robot according to the embodiment, the obstacle avoidance can be performed in a relatively simple manner through the recognition of the obstacle around the mobile robot.

1 shows an obstacle avoidance system of a mobile robot according to an embodiment.
2 shows a geometry for explaining an obstacle avoidance of the mobile robot.
Figures 3 (a) - (d) illustrate examples of obstacle types.
Fig. 4 shows a avoidance region according to the relative angle and distance between the mobile robot and the obstacle.
FIG. 5A shows a case where there is no obstacle between the mobile robot and the target point, and FIG. 5B shows a case where there is a circular obstacle between the mobile robot and the target point.
FIG. 6 illustrates a method of avoiding obstacles of a mobile robot according to an embodiment of the present invention.
7 (a) and 7 (b) show simulation and experimental results using the first and second switching signals and the first control law based on the zero-space projection method, and FIG. 7 (c) And the results of experiments are shown.
8 (a) to 8 (c) show simulation and experimental results using the third switching signal and the first control law based on the rotation of the repulsive force vector, and Figs. 8 (d) 8 (g) to 8 (i) show obstacle avoidance simulations of a general mobile robot based on a potential field approach, and FIGS. 8 The experimental results are shown.
Figures 9 (a) - (d) show simulation and experimental results to avoid multiple static obstacles.
Figures 10 (a) - (d) show simulation and experimental results for collision avoidance.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. Like reference symbols in the drawings denote like elements.

Fig. 1 shows an obstacle avoidance system of a mobile robot according to an embodiment, Fig. 2 shows a geometry for explaining obstacle avoidance of a mobile robot, Figs. 3 (a) to 3 (d) 5 (a) shows a case where there is no obstacle between the mobile robot and the target point, and FIG. 5 (b) shows a case where there is no obstacle between the mobile robot and the target point. Shows a case where there is a circular obstacle between the mobile robot and the target point.

1, an obstacle avoidance system 10 of a mobile robot according to an embodiment includes a sensing unit 100, a position acquiring unit 200, a signal generating unit 300, a vector computing unit 400, (500).

First, a robot system applied in the obstacle avoidance system 10 of the mobile robot according to an embodiment will be described.

The robotic system is a nonholonomic mobile robot which can be restricted from moving sideways by a nonholonomic constraint.

The kinematic equation of the non-holonomic robot is expressed as follows.

은 모바일 로봇의 상태이고,

는 모바일 로봇의 선속도와 각속도(linear and angular velocity)이다.here,

Is the state of the mobile robot,

Is the linear velocity and angular velocity of the mobile robot.

2, the geometry used in the present invention is as follows.

: 모바일 로봇을 나타내고,(One)

: Indicates a mobile robot,

: 좌표계를 나타내고,(2)

: Indicates a coordinate system,

(3) q: The pose of the mobile robot

(4) p: Position of mobile robot

(5) [theta]: The heading angle of the mobile robot

(6) θ _o : Bearing angle for the obstacle

(7) θ _ro : θ - θ _o

(8) P _g : Position of target point

(9) P _o : Location of the obstacle

(10) P _d : virtual target point

: 회피 벡터(11)

: Avoidance vector

: UVV(통합 속도 벡터)(12)

: UVV (integrated speed vector)

(13) R: Sensor detection area

(14) r: avoidance area

With reference to the above-mentioned contents, the configuration of the obstacle avoidance system 10 of the mobile robot according to one embodiment will be described in detail below.

The sensing unit 100 may sense an obstacle located around the mobile robot traveling toward the target point. The sensing unit 100 may directly sense an obstacle or indirectly sense an obstacle through environmental information around the mobile robot.

The sensing unit 100 may be, for example, an onboard sensor mounted on a mobile robot. However, the sensing unit 100 is not limited thereto, and any sensing unit 100 can detect an obstacle.

In addition, various types of obstacles can be detected through the sensing unit 100.

The obstacles can be largely classified into static obstacles and dynamic obstacles (for example, mutual collision between robots).

In particular, the static obstacle can be divided into a circular obstacle and a large non-circular obstacle, with reference to Figs. 3 (a) to 3 (d).

In the case of Fig. 3 (a), it is an obstacle small enough to be able to be modeled as one circle or circle, or sufficiently small to be bounded by a circle. 3 (b) to 3 (c) are non-circular obstacles, and there may be a case where there are several continuous obstacles on the surface of the obstacle.

In addition, when another mobile robot is detected around the mobile robot, the other mobile robot can be modeled as a circle, and the obstacle shown in FIG. 3 (a) can be considered as a dynamic obstacle.

The position acquisition unit 200 may be connected to the sensing unit 100.

The position acquiring unit 200 may acquire position information of a mobile robot, an obstacle, and a target point.

Here, the location information of the mobile robot indicates the current state of the mobile robot, the location information of the obstacle indicates the current state of the obstacle to the mobile robot, the location information of the target point is a location .

Therefore, the obstacle avoidance system of the mobile robot according to an embodiment does not need to acquire the entire spatial information of the mobile robot in advance, detects the obstacle by the sensing unit 100 while the mobile robot is traveling, The obstacle location information can be obtained in real time via the network 200.

The signal generator 300 may be connected to the position acquiring unit 200.

The signal generator 300 may be related to the location information of the obstacle, and may generate the switching signal in consideration of the distance between the mobile robot and the obstacle and the relative angle.

An obstacle avoidance task can be activated by the switching signal.

For example, if an obstacle is detected, a switching signal may be generated and the obstacle avoidance task may be activated, but if no obstacle is detected, no switching signal is generated and the obstacle avoidance task may be disabled .

At this time, the switching signal may be differently applied depending on the type of the obstacle, and may include a first switching signal, a second switching signal, and a third switching signal.

First, in the case of a circular obstacle, the first switching signal can be expressed as follows.

(1)

(One)

는 제1 스위칭 신호이고, d_ro는 상기 모바일 로봇과 상기 장애물 사이의 거리이고, θ는 상기 모바일 로봇의 주행 각도(heading angle)이며, At this time,

D _ro is a distance between the mobile robot and the obstacle, θ is a heading angle of the mobile robot,

이고,The d _m

ego,

In this case, R is the sensing range of the sensing unit, r is the avoidance range, and? _Ro can be the difference between the traveling angle of the mobile robot and the obstacle azimuth angle, that is, the relative angle.

Further, the azimuth angle of the obstacle can be expressed as follows.

(2)

(2)

Therefore, the first switching signal may be as follows.

(3)

(3)

(4)

When the first switching signal is 0, it means that the robot does not hit an obstacle when moving toward the current driving direction of the mobile robot.

Referring to FIG. 4, an avoidance region and a safe region according to the first switching signal can be confirmed.

인 경우, 그리고 모바일 로봇과 장애물 사이의 거리가 감지 범위(detection range)보다 작다면,

이 될 수 있다. 즉, 제1 스위칭 신호는 불활성화될 수 있다.E.g,

, And if the distance between the mobile robot and the obstacle is less than the detection range,

. That is, the first switching signal may be inactivated.

rad까지 증가함에 따라, 회피 영역은

에서

로 감소된다. 그러나, 모바일 로봇과 장애물 간의 거리만 고려할 경우, 회피 영역은 항상

이다. 그리고 θ_ro 절대치 값은 회피 영역에서 항상 증가한다.Further, when &thetas; _ro is 0rad

rad, the avoidance area is < RTI ID = 0.0 >

in

. However, when considering only the distance between the mobile robot and the obstacle,

to be. And θ _ro The absolute value is It always increases in the avoidance area.

이 항상 증가하는 것을 담보하여, 상기 시스템이 스위칭 표면을 항상 회피하게 할 수 있다.At this time, the control law, which will be described later,

Is always increased, so that the system can always avoid the switching surface.

Second, in the case of a circular obstacle, a second switching signal may be generated to supplement the first switching signal.

Under the first switching signal, the mobile robot may avoid an obstacle, but chattering may occur during movement of the mobile robot. In other words, the mobile robot can often be switched between the avoidance zone and the safety zone.

To prevent or reduce such chattering, a second switching signal may be generated.

At this time, a piecewise continuous switching signal can be generated to feed forward the current switching signal and prevent the system from rapidly switching to infinity.

The second switching signal may be expressed as follows.

(5)

(5)

는 제2 스위칭 신호이고,

는 상기 모바일 로봇의 현재 스위칭 신호일 수 있다.At this time,

Is a second switching signal,

May be the current switching signal of the mobile robot.

Third, in the case of a non-circular obstacle, a third switching signal may be generated.

At this time, the non-circular obstacle may be at least larger than the detection range of the robot. Specifically, the form of the obstacle from which the third switching signal can be generated is shown in Figs. 3 (b) to 3 (d).

Therefore, it will be described below that the third switching signal is generated by a large non-circular obstacle.

The third switching signal may be expressed as follows.

(6)

(6)

는 제3 스위칭 신호이고,

는 상기 모바일 로봇과 상기 목표 지점을 잇는 선이고,

는 상기 장애물 상의 모든 점의 집합이며, 보통 장애물의 윤곽(contour) 상의 모든 점의 집합을 말한다.At this time,

Is a third switching signal,

Is a line connecting the mobile robot and the target point,

Is a set of all points on the obstacle, usually a set of all points on the contour of the obstacle.

In order to deal with the large non-circular obstacle more effectively, when considering the time difference (? T) updated by the third switching signal, the avoidance vector must satisfy the following condition.

(7)

(7)

조건을 확인하기 위해서는

와 장애물의 윤곽이 교차점이 존재하는지를 확인하면 된다.And

To check the condition

And the outline of the obstacle intersect.

. 그리고 모바일 로봇과 목표 지점을 연결한 선과 장애물 사이에 교차점이 있어야 한다

. 이 두 조건을 만족하는 경우, 제3 스위칭 신호는 1로 표시되고, 그렇지 아니한 경우에는 제3 스위칭 신호가 0으로 표시될 수 있다.Also, since it is difficult to model a large non-circular obstacle as a circular obstacle, the third switching signal only uses the distance information between the mobile robot and the obstacle without considering the orientation information of the robot

. And there must be an intersection between the line connecting the mobile robot and the target point and the obstacle

. When these two conditions are satisfied, the third switching signal is represented by 1, and if not, the third switching signal may be represented by 0.

In addition, the closest distance information between the mobile robot and the large non-circular obstacle is used at each moment and the corresponding point can be considered as an obstacle.

도로 회전시킬 때, 항상 같은 방향으로 회전시켜야 한다는 것이다. 예를 들면 항상

로 방향으로 회전을 시키거나, 항상 반대 방향으로 회전시키는 것이 바람직할 수 있다.In this case, note that the repulsive force vector is based on the z-axis

When you turn the road, you always have to rotate in the same direction. For example,

, Or it may be desirable to always rotate in the opposite direction.

Or, when using the zero space projection method, keep the same direction as the new vector generated at time t and time t + 1. If the direction of the vector generated at time t and time t + 1 is opposite to each other, the direction of the new vector at time t + 1 must match the direction of vector at time t. At this time, it is possible to match the normal vector directions by rotating the new vector at time t + 1 by?.

Thus, depending on the type of obstacle, the switching signal can be determined by different equations.

In addition, a vector operation unit 400 may be connected to the position acquisition unit 200.

In the vector operation unit 400, a repulsive vector between the mobile robot and the obstacle, a pulling force vector between the mobile robot and the target point, and an avoidance vector formed vertically from the repulsion force vector can be calculated.

For example, in the circular obstacle avoidance problem of the mobile robot, it is possible to consider a case where there is no obstacle as shown in Fig. 5 (a) and a case where there is an obstacle as shown in Fig. 5 (b).

For each case, there are task functions φ _rg and φ _ro , and the corresponding vector can be called an attractive vector and a repulsive vector.

If there is no obstacle as shown in Fig. 5 (a), the attraction vector is generated from the mobile robot by the target point, and the mobile robot can move toward the vector direction to reach the target point.

On the other hand, if there is an obstacle as shown in FIG. 5 (b), an obstacle is detected between the mobile robot and the target point, and a repulsion vector may be generated between the mobile robot and the obstacle. The repulsive force vector may prevent the mobile robot from striking the obstacle.

At this time, it may be preferable that the mobile robot moves in a direction perpendicular to the repulsive force vector and moves in a direction close to the target point.

)와 회피 벡터(

)의 내적이 0이 되어 두 벡터가 수직이어야 하며, 또한 인력 벡터(

)와 회피 벡터(

)의 내적은 0과 같거나 0보다 큰 값을 취하며 그 방향이 장애물에서 모바일 로봇으로 향하는 벡터이어야 한다.In other words, a new vector perpendicular to the repulsion vector must satisfy the following two conditions. The repulsion vector (

) And the avoidance vector

) Must be zero so that the two vectors are perpendicular, and the excitation vector (

) And the avoidance vector

) Should take a value equal to or greater than 0 and be a vector whose direction is from the obstacle to the mobile robot.

This can be expressed as follows.

(8)

(8)

Then, a new vector perpendicular to the repulsion vector can be obtained by the following two methods.

First, according to the Null space projection approach,

(9)

(9)

At this time,

)가 0이 되게 할 것이다. 이러한 국소 최소값(local minima)은 불안정하며, 장애물을 회피하기 위해 모바일 로봇을 약간 회전시킬 필요가 있다.The zero space projection approach is to project a force vector onto the left zero space of the repulsion vector. This approach, when the P, P _o, P _g is in the same plane avoidance vector (

) To be zero. This local minima is unstable and requires a slight rotation of the mobile robot to avoid obstacles.

Then, the avoidance vector, which has a small nonzero value and is perpendicular to the repulsion vector, can move the mobile robot away from the local minimum. Turning towards the force vector may be the best option, but the direction of rotation may not affect the stability of the system.

만큼 회전시키는 접근에 의하면 다음과 같다.Second, the repulsive force vector is calculated for the z-axis

According to the approach of turning as follows.

(10)

(10)

At this time,

Both of the approaches described above are similar in that they both yield an avoidance vector that is perpendicular to the repulsive force vector and pointing in the same direction, but the size of the avoidance vector may be different. And the two approaches may provide different feed forward terms for the control law that will be described later.

As described above, the attraction vector, the repulsive force vector, and the avoidance vector can be calculated through the vector operation unit 400.

The switching signal generated in the signal generating unit and the attraction vector and the avoidance vector calculated in the vector calculating unit may be transmitted to the path determining unit 500.

The path determining unit 500 can determine the movement path of the mobile robot using the avoidance vector, the attraction vector, and the switching signal.

At this time, in the path determining unit 500, a unified velocity vector (UVV) indicating a path to be moved by the mobile robot in order to avoid an obstacle can be calculated.

Concretely, the integrated speed vector can be calculated by the following equation.

(11)

(11)

는 통합 속도 벡터이고,

는 스위칭 신호이고,

는 회피 벡터이고,

는 인력 벡터이다.At this time,

Is an integrated velocity vector,

Is a switching signal,

Is an avoidance vector,

Is a force vector.

For example, when the switching signal is 0 (off), the integrated velocity vector is affected by the attraction vector, and when the switching signal is 1 (on), the integrated velocity vector may be affected by the avoidance vector.

In other words, if an obstacle is not detected around the mobile robot, the mobile robot can continue running toward the target point, and if an obstacle is detected around the mobile robot, the mobile robot moves in the direction of the avoidance vector formed vertically from the repulsion vector Can be driven.

의 끝점이 가상의 목표 지점(virtual goal position; P_d)이 될 수 있다. 그러므로, 현재 모바일 로봇의 위치 정보에 통합 속도 벡터의 합을 통해 가상의 목표 지점이 연산될 수 있다.Also,

May be the virtual goal position (P _d ). Therefore, a virtual target point can be calculated through the sum of the integrated velocity vectors in the current position information of the mobile robot.

Then, from the virtual target point P _d , the target velocity v _d And w _d may be computed. At this time, the target speed may be a feed forward term input to the controller to be described later.

Specifically, the target speed can be expressed as follows.

(12)

(12)

(13)

(13)

및

는 아래의 식들에 의해서 산출될 수 있으며, 영 공간 투사 접근 및 z-축에 대한 척력 벡터의

rad 회전 접근으로 될 수 있다.Gt;

And

Can be calculated by the following equations, and the zero-space projection approach and the repulsive vector of the z-axis

rad rotation approach.

You can also consider static and dynamic obstacles for each approach.

In the case of the dynamic obstacle, it can be considered as an extension of the circular static obstacle avoidance. Circular static obstacle avoidance does not change the position of the obstacle, but dynamic obstacle position changes in real time. Therefore, the repulsive vector can be determined by the mobile robot and the dynamic obstacle. Since the position of the dynamic obstacle always changes relative to the static obstacle, the feedforward term entering the controller can be different from the static obstacle.

A. Young Space Projection Approach

1) static obstacle

When the obstacle is stationary, the differential equation for the integrated velocity vector equation, ie, the time of equation (11), is:

(14)

(14)

(15)

(15)

2) Dynamic obstacles

When the obstacle is dynamic, the integrated velocity vector equation, ie, the differential equation for the time in equation (11) is:

(16)

(16)

(17)

(17)

rad 회전B. The repulsive force vector for the z-axis

rad rotation

1) static obstacle

When the obstacle is static, the differential equation for the time in Eq. (11) is as follows.

(18)

(18)

(19)

(19)

2) Dynamic obstacles

When the obstacle is dynamic, the differential equation for the time in Eq. (11) is as follows.

(20)

(20)

(21)

(21)

Thus, the target speed of the mobile robot can be obtained from the equation (11).

Further, the direction information? _D can be calculated from the equation (11).

Specifically, it can be calculated by the following equation.

(22)

(22)

The path determining unit 500 may include a controller to which the target speed and the target direction can be input.

In the controller, the target velocity can be corrected or controlled by a so-called control law for stabilization of the system.

The control law may include a first control law and a second control law.

The first control law allows forward and backward motions of the mobile robot while the second control law provides a bounded control input that allows only the forward motion of the mobile robot.

The obstacle avoidance system 10 of the mobile robot according to the embodiment can switch the operation between the obstacle avoidance mode and the obstacle-free modes by the switching signal generated in the signal generator 300, so that the hybrid switching system hybrid switching system.

The first control law and the stability analysis are as follows.

An equivalent error (ε) to represent the integrated velocity vector and an orientation error (θ _ε ) in the local coordinate system of the mobile robot can be used.

(23)

(23)

은 가역적이므로,

⇔

이다.In addition,

Is reversible,

⇔

to be.

The control input values v and w can then be selected as the feed forward term.

(24)

(24)

Where k _x , k _y, and k _[theta] are positive scalar values.

By differentiating [epsilon] in Eq. (23) and considering non-holonomic constraints, the error dynamics can be obtained as follows.

(25)

(25)

In addition, a Lyapunov candidate function can be proposed as follows.

(26)

(26)

The derivative of equation (26) is as follows.

(27)

(27)

Therefore, the control input can be rewritten as follows.

(28)

(28)

이라는 것을 알 수 있으며, 시스템은 안정화 여부를 판단할 수 있다.Through this expression

, And the system can judge whether or not it is stabilized.

Substituting Eq. (24) into Eq. (25) yields the following.

(29)

(29)

At this time, equation (29) can be simplified as follows through a simple assumption.

(30)

(30)

For example, by linearizing equation (30), it can be as follows.

(31)

(31)

이고, 식 (31)의 고유 값이 양의 값이라는 것을 알 수 있으며,

는 불안정한 평형 상태이다. At this time,

, And that the eigenvalue of equation (31) is a positive value,

Is an unstable equilibrium state.

Therefore, any points around the equilibrium point will converge to the largest invariant set.

Thus, it can be seen that according to LaSalle's invariant principle, the equilibrium point (0; 0; 2kπ) can be asymptotically stable through the control input.

는 특별하고 컴팩트하며, 모든 서브 시스템들에 대하여 약간 안정적인 불변 집합이고, 식 (26)은 임의의 스위칭 신호 하에서 일반적인 리아푸노프 함수이다. Besides,

(26) is a generic Riapunov function under any switching signal, and is a constant set that is special and compact, slightly stable for all subsystems.

Therefore, the switching system with the switching signal described above can become asymptotically stable, according to the Bacciotii's theorem, and the mobile robot will track the desired pose (q _d ).

As a result, the desired pose error of the mobile robot can be asymptotically stabilized while guiding the mobile robot to the target point.

In this case, the inner product of the integrated velocity vector and the attraction vector should be equal to or greater than 0, whether the mobile robot is in the obstacle-free area or the avoidance area.

The dot product of the two vectors is

(32)

(32)

As a result, the mobile robot can eventually reach the target point by the first control law.

The second control law and stability analysis are as follows.

The direction error of the mobile robot is as follows.

(33)

(33)

Here, θ _d is the preferred direction information defined in equation (22).

The second control law can be expressed as follows.

(34)

(34)

Where k ₁ and k ₂ are positive scalar values.

To analyze the stability, a Lyapunov candidate function can be proposed as follows.

(35)

(35)

 (35) can be differentiated with respect to time as follows.

(36)

(36)

의 제2 텀(term)은

일 때마다 음수(negative)라는 것을 알 수 있다.

의 제1 텀(term) 또한

이고 m이 짝수일 때 음수(negative)이다.At this time,

The second term of < RTI ID = 0.0 >

It can be seen that each time it is negative.

The first term of < RTI ID = 0.0 >

And is negative when m is an even number.

Therefore, the system can be asymptotically stable by the second control law. The remainder of the proof follows the first control law and can establish stability under certain switching and always positive projections to the target point.

With the two control laws designed in this way, the corrected target velocities v and ω can be transmitted to the mobile robot by correcting the target velocities v _d and ω _d calculated from the integrated velocity vectors.

As a result, the mobile robot can be moved by the corrected target speed, and at this time, the state of the mobile robot can be localized.

Therefore, the obstacle avoidance system of the mobile robot according to the embodiment can perform the stable switching between the two situations by separately considering the situation in which the obstacle exists and the situation in which there is no obstacle, and the real time stable system switching approach (Stable Switched-System Approach).

FIG. 6 illustrates a method of avoiding obstacles of a mobile robot according to an embodiment of the present invention.

Referring to FIG. 6, an obstacle avoiding method of a mobile robot according to an embodiment can be performed.

Hereinafter, a case where a static obstacle is detected will be described as an example.

However, when a dynamic obstacle is detected, it can also be performed in a similar manner, but only the formula for obtaining the target velocity can be applied differently.

First, the initial state is obtained (R _i) of the mobile robot (S10).

The initial state (R _i) of the mobile robot can be obtained through the pickup position.

Then, the environment around the mobile robot is acquired (S20).

At this time, the obstacle located around the mobile robot can be detected by the sensing unit.

When the obstacle is detected, the position (P _o ) of the obstacle and the position of the target point (P _g ) toward the mobile robot are obtained by the position obtaining unit (S30).

The position of the obstacle and the position of the target point toward which the mobile robot is directed can be obtained simultaneously or sequentially by the position obtaining section.

Then, it is determined whether the obstacle is a static obstacle or a dynamic obstacle (S40).

If the obstacle is a static obstacle, it is determined whether the obstacle is a circular obstacle or a large non-circular obstacle (S50).

Then, a switching signal is generated in consideration of the distance between the mobile robot and the obstacle and the relative angle (S60).

At this time, depending on whether the obstacle is a circular obstacle or a large non-circular obstacle, the switching signal? _Ri can be generated according to another equation. If no obstacle is detected around the mobile robot, the switching signal may be O.

Simultaneously or sequentially with the generation of the switching signal, the avoidance vector formed vertically from the repulsive force vector between the mobile robot and the obstacle, the attracting force vector between the mobile robot and the target point, and the repulsion force vector is calculated (S70).

The repulsive force vector can be obtained from the current state of the mobile robot and the position of the obstacle, and the attraction vector can be obtained from the current state of the mobile robot and the position of the target point. The avoidance vector may be formed such that the inner product with the repulsive force vector is equal to or greater than zero.

Then, an integrated velocity vector indicating a path through which the mobile robot is to be moved by the mobile robot in order to avoid the obstacle is calculated by the avoidance vector, the attraction vector, and the switching signal (S80).

At this time, if an obstacle is not detected around the mobile robot and the switching signal is 0, the integrated velocity vector can be expressed by only the attraction vector, and if the obstacle is detected around the mobile robot and the switching signal becomes 1, The integrated velocity vector can be expressed by an expression of the avoidance vector only.

Subsequently, a virtual target point and a target direction are calculated from the integrated velocity vector (S90).

Further, the target speed is calculated from the virtual target point (S100).

Then, the target speed is corrected (S110).

Specifically, the target speed and the target direction are input to the first control law or the second control law in the controller, and the feedforward control can be performed.

Finally, a command for the corrected target speed is transmitted to the mobile robot according to the corrected target speed (S120).

Then, the mobile robot is moved and the state of the mobile robot is localized (S130).

After the state of the mobile robot is localized, the current state of the mobile robot is calculated (S140).

Then, whether or not the above-described process is stopped is determined (S150).

For example, if the user selects to stop, the above-described process can be terminated, and if not selected to stop, the environment of the mobile robot during the above-described process can be recovered.

Although the system and method for obstacle avoidance of the mobile robot according to one embodiment described above have been described, simulation and actual experimental results of the obstacle avoidance system and method of the mobile robot according to one embodiment will be described below.

7 (a) and 7 (b) show simulation and experimental results using a zero-space projection method with first and second switching signals and a first control law, and Fig. 7 (c) Avoidance simulation, and experimental results.

In this case, the blue circle and the green point represent the obstacle and the target point, respectively, and the initial position of the mobile robot is the same as the target point (the number 4).

Referring to Figs. 7 (a) to 7 (c), the mobile robot includes [1; 0; 0] ^T and sequentially move counterclockwise toward the four intermediate points (waypoints). At this time, an obstacle was placed between all two intermediate points. The results in these simulations and experiments show that the mobile robot successfully avoids four obstacles and reaches four intermediate points.

Also, when comparing Figs. 7A to 7C, it can be seen that angular velocity using the second switching signal is less chattering than others. However, the tracking result using the first switching signal in Fig. 7 (a) is smaller than the second switching signal in Fig. 7 (b) and the obstacle avoidance simulation and experimental results of the general mobile robot in Fig. It can be seen that it has.

에서

으로 감소할 것이다.Since the first and second switching signals consider both the distance and the direction,

in

.

을 유지한다.In the case of obstacle avoidance of a general mobile robot, only the distance between the mobile robot and the obstacle is considered,

Lt; / RTI >

In addition, by considering the direction, the motions of the mobile robot are not affected by the four obstacles using the first and second switching signals, but in the case of the obstacle avoidance of a general mobile robot, .

8 (a) to 8 (c) show simulation and experimental results using the third switching signal and the first control law based on the rotation of the repulsive force vector, and Figs. 8 (d) 8 (g) to 8 (i) show obstacle avoidance simulations of a general mobile robot based on a potential field approach, and FIGS. 8 The experimental results are shown.

These simulations and experiments were performed with large non-circular obstructions.

에서 시작하고

에 도달해야 한다. 이때, 모든 시뮬레이션들 및 실험들에는 비원형 장애물들이 구비된다.All simulations and experiments are also based on the second control law, and the control gains are k ₁ = 0.5 and k ₂ = 4. The mobile robot

Start with

. At this time, non-circular obstacles are included in all simulations and experiments.

Referring to Figs. 8 (a) to 8 (c), the rotation of the repulsion vector including the third switching signal can efficiently deal with the obstacles.

However, as shown in Figs. 8 (d) to 8 (f), the first switching signal can be difficult to deal with large non-circular obstacles.

Further, as shown in Figs. 8 (g) to (i), the potential field approach may fail to avoid obstacles and the mobile robot may fall into the local minima.

Figures 9 (a) - (d) show simulation and experimental results to avoid multiple static obstacles.

만큼 회전시키는 것에 의해 통합 속도 벡터를 유도하고, 제2 제어 법칙을 이용하였으며, 모두 제1 스위칭 신호 및 제2 스위칭 신호를 이용하였다. In this simulation and experiment, the repulsive force vector is calculated for the z-axis

And the second control law was used. Both the first switching signal and the second switching signal were used.

라면, 척력 벡터는 z축에 대해

rad으로 회전될 것이고, 그렇지 않다면, 척력 벡터가 반대 방향으로 회전될 것이다. E.g,

, The repulsive force vector is calculated for the z-axis

rad, otherwise the repulsion vector will be rotated in the opposite direction.

Multiple obstacle avoidance scenarios can not be solved using the global path planar, so individual obstacle information can be considered locally.

라면, 새로운(또는 다음) 장애물을 회피하도록 고려될 수 있다.if

, Then it can be considered to avoid the new (or next) obstacle.

에서 시작하고, 예를 들어 20m x 30m이고 30개의 정적 장애물들로 덮인 회피 영역을 통과하여,

에서 도달할 필요가 있다.As shown in Figs. 9 (a) and 9 (b), the mobile robot

For example, 20m x 30m and passes through the avoidance zone covered by 30 static obstacles,

.

으로 변경시켜 시뮬레이션 및 실험을 다시 수행할 수 있다.Thereafter, as shown in Figs. 9 (c) and 9 (d), within the same obstacle avoidance region, the positions of the initial position and the target position of the mobile robot are And

The simulation and the experiment can be performed again.

Figures 10 (a) - (d) show simulation and experimental results for collision avoidance.

In this simulation and experiment, two robots point to each other and exchange their positions. This constitutes a local minima form for each simulation and experiment.

만큼 회전시키는 것에 의해 유도된 통합 속도 벡터와, 각각 제1 스위칭 신호 및 제2 스위칭 신호를 이용하는 것이다.10 (a) and 10 (b), the repulsive force vector is calculated for the z-axis

And a first switching signal and a second switching signal, respectively.

만큼 회전시키는 것에 의해 유도된 통합 속도 벡터를 이용하는 것이고, 도 10(d)는 일반적인 모바일 로봇의 장애물 회피 시뮬레이션 및 실험에서 포텐셜 필드 기법을 이용하는 것이다.On the other hand, Fig. 10 (c) shows the obstacle avoidance simulation and experiment of a general mobile robot,

Fig. 10 (d) illustrates the use of the potential field technique in the obstacle avoidance simulation and experiment of a general mobile robot.

From the distances and trajectories between the two robots, it is possible to confirm that the two robots have successfully avoided collision with each other with the approach proposed in the present invention as shown in Figs. 10 (a) to (c) , As shown in Fig. 10 (d), the two robots can reach a deadlock when using the potential field.

Therefore, the obstacle avoidance system and method of the mobile robot according to the embodiment can avoid static circular or non-circular obstacles, prevent collision between robots by cooperating with each other, and enable autonomous obstacle avoidance of the mobile robot , Robustness and reliability of algorithms, one of the main issues of mobile robots for industrial practical use, safe motion planning, autonomous driving, etc., can be obtained. In addition, it is possible to prevent chattering or oscillation that may occur in obstacle avoidance of the mobile robot, to shorten the time required for obstacle avoidance, and to avoid obstacles in a relatively simple manner by recognizing obstacles around the mobile robot Can be performed.

Although the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, And various modifications and changes may be made thereto without departing from the scope of the present invention. Accordingly, the spirit of the present invention should not be construed as being limited to the embodiments described, and all of the equivalents or equivalents of the claims, as well as the following claims, belong to the scope of the present invention .

10: Mobile robot obstacle avoidance system
100:
200:
300: Signal generator
400: Vector operation unit
500: Path determining unit

Claims (10)

A sensing unit for sensing an obstacle located around the mobile robot traveling toward the target point;
A position acquiring unit capable of acquiring position information of the mobile robot, an obstacle, and the target point;
A signal generator for generating a switching signal in consideration of a shape of the obstacle, a distance between the mobile robot and the obstacle, and a relative angle when the obstacle is detected by the sensing unit;
A vector computing unit for computing a repulsive force vector between the mobile robot and the obstacle and a force vector between the mobile robot and the target point through position information obtained from the position obtaining unit; And
A path determination unit for determining a travel path of the mobile robot using the attraction vector and the switching signal;
Lt; / RTI >
The traveling of the mobile robot can be controlled in real time by the switching signal,
Wherein the switching signal includes a first switching signal generated upon detection of a circular obstacle,
The first switching signal is determined as follows,

At this time,

D _ro is a distance between the mobile robot and the obstacle, θ is a heading angle of the mobile robot,
The d _m

ego,
Here, R is the sensing range of the sensing unit, r is the avoidance range,? _Ro is the difference between the traveling angle of the mobile robot and the obstacle azimuth angle,
Wherein the switching signal further comprises a second switching signal for compensating for the operation of the mobile robot by the first switching signal,
The second switching signal is determined as follows,

At this time,

Is a second switching signal,

Is the current switching signal of the mobile robot,
The second switching signal is determined in real time considering the distance between the mobile robot and the obstacle, the relative angle, and the current switching signal of the mobile robot, thereby reducing the avoidance area of the mobile robot, Obstacle avoidance system of mobile robot to prevent the phenomenon.
The method according to claim 1,
Wherein the avoidance vector is further calculated from the repulsive force vector in the vector operation unit, and the avoidance vector has an inner product with the attracting vector equal to or greater than zero.
3. The method of claim 2,
In the path determining unit, an unified velocity vector indicating a path to be moved by the mobile robot in order to avoid the obstacle is calculated,
The integrated velocity vector is calculated by the following equation,

At this time,

Is an integrated velocity vector,

Is a switching signal,

Is an avoidance vector,

Is a workforce vector.
The method of claim 3,
A virtual target point is calculated from the integrated velocity vector and the position information of the mobile robot in the path determination unit, and the velocity information calculated from the virtual target point can be corrected.
delete delete The method according to claim 1,
Wherein the switching signal further comprises a third switching signal generated upon detection of the non-circular obstacle,
The third switching signal is determined as follows,

At this time,

Is a third switching signal,

Is a line connecting the mobile robot and the target point,

Wherein the obstacle avoidance system is a set of all points on the obstacle.
Obtaining an initial state of the mobile robot;
Obtaining an environment of the mobile robot;
Detecting an obstacle located around the mobile robot;
Obtaining a position of an obstacle around the mobile robot and a position of a target point toward the mobile robot;
Generating a switching signal based on a shape of the obstacle, a distance between the mobile robot and the obstacle, and a relative angle; And
Calculating a repulsive force vector between the mobile robot and the obstacle, a force vector between the mobile robot and the target point, and an avoidance vector formed vertically from the repulsive force vector;
Lt; / RTI >
The traveling of the mobile robot can be controlled in real time by the switching signal,
Wherein the switching signal includes a first switching signal generated upon detection of a circular obstacle,
The first switching signal is determined as follows,

At this time,

D _ro is a distance between the mobile robot and the obstacle, θ is a heading angle of the mobile robot,
The d _m

ego,
Here, R is the sensing range of the sensing unit, r is the avoidance range,? _Ro is the difference between the traveling angle of the mobile robot and the obstacle azimuth angle,
Wherein the switching signal further comprises a second switching signal for compensating for the operation of the mobile robot by the first switching signal,
The second switching signal is determined as follows,

At this time,

Is a second switching signal,

Is the current switching signal of the mobile robot,
The second switching signal is determined in real time considering the distance between the mobile robot and the obstacle, the relative angle, and the current switching signal of the mobile robot, thereby reducing the avoidance area of the mobile robot, The obstacle avoiding method of the mobile robot prevents the phenomenon.
9. The method of claim 8,
After the step of calculating a repulsive vector formed vertically from the repulsive force vector and the repulsive force vector between the mobile robot and the obstacle, the attraction vector between the mobile robot and the target point,
Further comprising the step of calculating an integrated velocity vector indicating a path through which the mobile robot is to be moved by the mobile robot in accordance with the avoidance vector, the attraction vector, and the switching signal.
10. The method of claim 9,
After the step of calculating the integrated speed vector indicating the route by which the mobile robot should be moved by the mobile robot in order to avoid the obstacle by the avoiding vector, the attracting vector, and the switching signal,
Calculating a virtual target point and a target direction from the integrated velocity vector;
Calculating a target speed from the virtual target point;
Correcting the target speed; And
Moving the mobile robot according to the corrected target speed;
The obstacle avoidance method of the mobile robot further comprising:

Images