KR101140984B1 - Safe path generating method considering appearance of invisible dynamic obstacle which is visibly occluded - Google Patents

Abstract

PURPOSE: A safe route creation method capable of considering appearance of a visually blocked dynamic obstacle which uses estimated risk costs and a mobile robot using the same are provided to compute a visually restricted region generated by a registered obstacle, thereby creating a safe route in which moving speed and direction of the mobile robot are considered. CONSTITUTION: A safe route creation method which considers appearance of a visually blocked dynamic obstacle is comprised of the following procedures. A restricted region is calculated with respect to an obstacle(S13). Estimated risk cost is calculated. The estimated risk cost is registered to a corresponding lattice(S16). A transport route to a target position is created(S20).

Description

Safe Path Generation Method Considering the Appearance of Visible Blocked Dynamic Obstacles and Mobile Robot Using the Same

The present invention relates to a method for generating a safety route in consideration of a predicted dynamic obstacle and a mobile robot using the same. In removing the present invention, the present invention relates to a safety route generation method and a mobile robot using the same.

In terms of autonomous navigation of autonomous mobile robots, safe driving to avoid collisions with dynamic obstacles in environments with dynamic obstacles such as humans is one of the most important problems. On the other hand, high speed driving is preferred to increase the efficiency of service of autonomous mobile robots.

In order to achieve high-speed driving of autonomous mobile robots, some limitations must be considered. These limitations include dynamic and mechanical limits, control and operational limits, and unexpected changes in the environment.

First, the dynamic and mechanical limitations mean a problem that a slip or changeover of the mobile robot occurs when the mobile robot corners or stops abruptly. However, in applying the mobile robot to the real environment, the dynamic and mechanical limit problem is not considered important because the high speed travel of the mobile robot causes more important problems than the dynamic and mechanical limit problem.

Control and computational limitations can be interpreted as a problem of avoiding obstacles in real time. The traveling speed of the mobile robot may be limited by sensing accuracy, processing speed, calculation amount, and motion control response. Kanayama, Y. Kimura, Y. Miyazaki, and F. Noguchi, T., et al. "A stable tracking control method for an autonomous mobile robot (Proc. IEEE International Conference on Robotics & Automation, Cincinnati, OH, USA, 13-18) May, 1990. pp. 384-389), proposes a trajectory tracking algorithm for a two-wheel differential robot that guarantees exponential convergence.

In addition, Macek, K., Petrovic, I. and Siegwart, R. also described a virtual vehicle approach in the paper "A control method for stable and smooth path following of mobile robots (In Proceedings of European Conference on Mobile Robots, 2005)". Based on Vehicle Approach, we proposed a control method for a stable and smooth path tracking algorithm.

And D. Fox, W. Burgard and S. Thrun wrote CVM (Curvature) in the paper "The Dynamic Window Approach to collision avoidance (IEEE Robotics and Automation Magazine, Vol. 4, no. 1, pp. 23-33, 1997). Velocity Method (DW)] is proposed as a useful control technique when the mobile robot travels in a dynamic obstacle environment.

In addition, Marija Seder and Ivan Petrovic, in the paper "Dynamic window based approach to mobile robot motion control in the presence of moving obstacles (Proc. IEEE International Conference on Robotics & Automation, 10-14 April, Roma, Italy, 2007, pp 1986-) In 1991, an improved DWA algorithm was proposed based on the integration of the 'Focused D * search algorithm', which considers moving obstacles.

In addition, Minguez's Nearness diagram (ND) technique performs exceptionally in very cluttered environments, while Quinlan's 'Real-time Modification of Collision-free paths' shows that the obstacle is in its original path. It proposes a technique for transforming a path by operating in real time.

Through the above techniques, the mobile robot is capable of autonomous driving while avoiding visible and dynamic obstacles, that is, unexpected obstacles detected by sensors installed in the mobile robot.

However, when the mobile robot is applied to a real environment, only the visible dynamic obstacles do not exist, and thus the risk of collision due to the visually blocked dynamic obstacles cannot be solved through the aforementioned techniques. For example, when driving along a narrow road and approaching an intersection while driving a daily vehicle, the driver generally decreases the speed of the vehicle. This is due to the expectation that a dynamic obstacle, for example another vehicle or person, that is visually blocked by the building walls forming the intersection will enter the intersection, in order to avoid blocking the visibility and consequently unexpected collisions.

Several approaches have been proposed to block this visibility and avoid unexpected collisions. One approach, M. Sadou, V. Polotski, and P. Cohen's paper, "Occlusions in obstacle detection for safe navigation (IEEE Intelligent Vehicle Symposium, pp. 716-721, 2004), focuses on obstacles that are visually blocked. In this paper, the range of visually blocked obstacles is proposed as blocked obstacles on the path, and the path is always fixed.

Other approaches include M. Bennewitz, W. Burgard, G. Cielniak, and S. Thrun's article "Learning motion patterns of people for compliant motion" (International Journal of Robotics Research, 24 (1), 2005). In this paper, we use the driving experience to predict human walking patterns, which shows that mobile robots can provide appropriate mobile services by monitoring and using human patterns.

However, the above-described conventional approaches for eliminating the risk of collision with the visually blocked dynamic obstacles have disadvantages that do not consider both the path planning and the speed control of the mobile robot. For example, in the case of humans, when entering an intersection environment as described above, it is common to slow down the vehicle and at the same time change the direction of the vehicle away from the position where it is expected to see a visually blocked dynamic obstacle. The above approach to robots has the disadvantage of lacking such complex considerations.

In addition, existing path planning focuses only on detectable obstacles, and it is common to approach mobile robots in terms of the distance traveled in creating an optimal path. However, in many driving environments in practice, it cannot be asserted that the shortest distance is inevitably safe, which cannot guarantee its safety in the presence of visible obstructed dynamic obstacles.

Accordingly, the present invention has been made to solve the above problems, in eliminating the risk of collision with the dynamic obstacle that is visually blocked by the registered obstacle present in the running environment of the mobile robot, route planning and mobile robot It is an object of the present invention to provide a method for generating a safety route in consideration of the appearance of a dynamically blocked dynamic obstacle capable of generating a safety route in which all driving speeds of the vehicle are reflected, and a mobile robot using the same.

According to the present invention, there is provided a safety path generation method considering the appearance of a visually blocked dynamic obstacle, the method comprising the steps of: (a) calculating a field of view restriction area by the obstacles already registered in each grid unit of the grid map; (b) calculating an expected risk cost based on a predetermined obstacle speed for the dynamic obstacle and a distance between the grating and the viewing limited area when the viewing limited area exists for the corresponding grid; (c) registering the estimated risk cost calculated in step (b) for the grid; and (d) generating a moving path to a target location by using a predetermined path generation technique by using the estimated risk cost registered in each of the grids.

Here, the path generation technique applied in the step (d) includes a gradient method using an intrinsic cost and an adjacent cost; The gradient method may generate the travel path by applying the estimated risk cost to the inherent cost.

In addition, in the step (b), when there is no view restriction area for the grating, the estimated risk cost for the grating may be set to a preset reference set value.

And, step (a) is (a1) calculating the minimum collision distance based on the obstacle speed and the robot speed of the mobile robot; calculating a reachable area for the lattice within the range of the minimum collision distance in the lattice unit; (a3) calculating the visible area of the mobile robot within the minimum collision distance within the grid; (a4) Based on the deviation between the reachable region and the visible region, the viewing restriction region for the grid may be calculated.

In addition, the minimum collision distance is

d _collision = d _delay + d _break + d _obs

d _delay = t _delay × (v _robot + v _obs )

d _break = v _robot ² / (2 × a _robot )

d _obs = v _obs × v _robot / a _robot

(Where d _collision is the minimum collision distance, t _delay is the motion delay time of the mobile robot, v _robot is the robot speed, v _obs is the obstacle speed, and a _robot is the acceleration of the mobile robot) Can be calculated by

Here, the reachable region may be calculated by a wavefront propagation algorithm, and the visible region may be calculated by a ray tracing method.

The estimated risk cost may be calculated based on the inverse of the maximum allowable speed of the mobile robot calculated based on the obstacle speed and the distance between the grid and the viewing limited area.

Here, the maximum allowable speed is a quadratic equation

v _max ² / (2 × a _robot ) + (t _delay + v _obs / a _robot ) × v _max + t _delay × v _obs = d _dist

(Where v _robot is the maximum allowable speed and t _delay is the operation delay time of the mobile robot, v _obs is the obstacle speed, a _robot is the acceleration of the mobile robot, and d _dist is the grid and the field of view limitation) Distance between the regions).

On the other hand, according to another embodiment of the present invention, the path generation unit for generating a moving path in accordance with the above-mentioned safe path generation method; It is also achieved by a mobile robot, characterized in that it comprises a travel control unit traveling along the movement path generated by the path generation unit.

Here, the driving controller may control the driving speed based on the maximum allowable speed.

Through the above configuration, according to the present invention in eliminating the risk of collision with the dynamic obstacle that is visually blocked by the registered obstacle present in the driving environment of the mobile robot, both the route planning and the traveling speed of the mobile robot are reflected Provided are a safety path generation method and a mobile robot using the same, considering the appearance of a dynamically blocked dynamic obstacle capable of generating a safety path.

1 is a control flowchart illustrating a method for generating a safe route according to the present invention;
2 is a view for explaining the risk of collision with an operating obstacle that is visually blocked by an obstacle,
3 is a diagram illustrating an example of displaying a collision risk grid on an environment map of a method for generating a safe route according to the present invention;
4 is a view showing an example of the distribution of the maximum allowable speed for the collision risk grating of the safety path generation method according to the present invention,
5 to 11 are views for explaining the effect of the safety path generation method according to the present invention,
12 is a diagram illustrating a configuration of a mobile robot to which a method for generating a safety route according to the present invention is applied.

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

The safety path generation method according to the present invention generates a safety path that eliminates the possibility of collision with the visually blocked dynamic obstacle DO, in consideration of the appearance of the visually blocked dynamic obstacle DO (see FIG. 2). Here, the safe path generation method according to the present invention generates a safe path in consideration of both path planning and speed limitation of the mobile robot 100.

1 is a view for explaining a safe path generation method according to the present invention. Referring to FIG. 1, a view limitation area DA (see FIG. 2) for an obstacle (hereinafter, referred to as a “fixed obstacle”) previously registered in each grid unit of a grid map is calculated. Here, the calculation process of the field of view restriction area DA is as follows.

First, a minimum collision distance is calculated based on the speed of the dynamic obstacle DO (hereinafter referred to as 'obstacle speed') and the speed of the mobile robot 100 (hereinafter referred to as 'robotic speed') (S10). . Here, the minimum collision distance is used to describe the risk of loyalty with the dynamic obstacle DO which is visually blocked by the fixed obstacle. 2 is a view for explaining the risk of collision with the dynamic obstacle (DO) visually blocked by the obstacle.

Referring to FIG. 2, when a map of a surrounding environment in which the mobile robot 100 travels is registered in the mobile robot 100, the area DA that is blocked by the fixed obstacle may be found. Here, from the perspective of the mobile robot 100, it means an area in which detection of the laser distance sensor 13 (see FIG. 12) is blocked by the fixed obstacle.

At this time, the collision distance with the dynamic obstacle DO, which is visually obscured by the fixed obstacle, that is, the minimum collision distance, is based on the motion performance of the mobile robot 100 and the obstacle speed of the dynamic obstacle DO. Equation 1] to [Equation 4] can be calculated.

[Equation 1]

d _collision = d _delay + d _break + d _obs

[Equation 2]

d _delay = t _delay × (v _robot + v _obs )

&Quot; (3) "

d _break = v _robot ² / (2 × a _robot )

&Quot; (4) "

d _obs = v _obs × v _robot / a _robot

Here, d _collision is the minimum collision distance, t _delay is the operation delay time of the mobile robot 100, v _robot is the robot speed of the mobile robot 100, v _obs is the obstacle speed of the dynamic obstacle (DO), a _robot is the acceleration of the mobile robot 100.

The minimum collision distance means the minimum distance from which collision with the dynamic obstacle DO may be avoided. That is, when the mobile robot 100 detects the dynamic obstacle DO in a state in which the mobile robot 100 is separated from the dynamic obstacle DO than the minimum collision distance, the mobile robot 100 may safely stop.

d _delay is a control that is delayed according to the performance of the mobile robot 100 such as a sensor or a controller such as a laser distance sensor 13 installed in the mobile robot 100 after the mobile robot 100 detects a dynamic obstacle (DO). The distance reflects the time delay. d _delay is calculated by the sum of the control delay time t _delay and the robot speed of the mobile robot 100 and the obstacle speed of the dynamic obstacle DO, as shown in [Equation 2].

d _break means the braking distance of the mobile robot 100, it is calculated through the equation (3). In addition, d _obs is a distance that the dynamic obstacle DO moves during the deceleration time until the mobile robot 100 completely stops. As shown in Equation 4, the speed and movement of the dynamic obstacle DO are It is calculated by the speed and acceleration of the robot 100.

Here, in most cases, since the operation delay time and acceleration of the mobile robot 100 can be estimated as a constant, the minimum collision distance d _collision is determined by the robot speed of the mobile robot 100 and the obstacle speed of the dynamic obstacle DO. Can be estimated. In the present invention, the robot speed of the mobile robot 100 for calculating the minimum collision distance is set to the maximum speed of the mobile robot 100, and the obstacle speed is set according to the driving environment characteristic of the mobile robot 100. Yes. For example, when the mobile robot 100 is located in a busy driving environment, the obstacle speed with respect to the dynamic obstacle DO may be set to a normal walking speed of the person.

When the minimum collision distance is calculated based on the obstacle speed and the robot speed as described above (S10), the reachable area for the grid is calculated within the range of the minimum collision distance in units of grids (S11). That is, in the present invention, the field of view limited area DA is calculated within a minimum collision range for one grid, even if the dynamic obstacle DO appears in the field of limited view outside the minimum collision range. This is because the collision may be avoided through the emergency stop even if the mobile robot 100 detects the dynamic obstacle DO.

In the present invention, the reachable area for the grating is calculated by the wavefront propagation algorithm, thereby minimizing the computational cost.

As described above, when the reachable area within the minimum collision distance range is calculated for the grating, the visible area VA by the laser distance sensor 13 of the mobile robot 100 is calculated (S12). Here, the calculation of the visible area VA is an example calculated by the ray tracing method.

Then, based on the deviation between the reachable area and the visible area VA, the field of view restriction area DA for the corresponding grating is calculated (S13). 2 conceptually illustrates the calculation of the field of view restricted area DA according to the deviation between the reachable area and the visible area VA. The method as shown in FIG. 2 is performed within the range of the minimum collision distance, The field of view restriction area DA is calculated.

Here, it is determined whether the viewing restriction area DA exists for the grid (S14), and if it is determined that the viewing restriction area DA exists, the grid is a collision risk grid due to the dynamic obstacle DO ( CDG, see FIG. 4), to calculate the estimated risk cost for the grid. The estimated risk cost according to the invention is calculated based on the obstacle speed and the distance between the grid and the field of view limited area DA.

Referring to FIG. 1, the maximum allowable speed of the mobile robot 100 is calculated based on the obstacle speed and the distance between the grating and the view restriction area DA (S15). Here, the maximum allowable speed of the mobile robot 100 refers to the maximum speed that the mobile robot 100 is allowed to avoid collision with the dynamic obstacle DO at a distance from the viewing limited area DA. In the present invention, it is assumed that the maximum allowable speed v _max of the mobile robot 100 is calculated through Equation 5, which is a quadratic equation derived from Equations 1 to 4 above.

[Equation 5]

v _max ² / (2 × a _robot ) + (t _delay + v _obs / a _robot ) × v _max + t _delay × v _obs = d _dist

[Equation 5] is derived by substituting the maximum collision distance d _collision of [Equation 1] in the distance between the grid and the field of view restriction area DA and replacing V _robot with v _max .

As described above, when the maximum allowable speed of the mobile robot 100 for the grid is calculated (S15), the inverse of the maximum allowable speed is registered as an estimated risk cost for the grid (S16). In other words, if the maximum allowable speed to allow the mobile robot 100 to collide with the dynamic obstacle (DO) in the grid is large, this means that the risk of collision is small, and if the maximum allowable speed is small, the risk of departure is high. This means that the inverse of the maximum allowable speed will be registered as the estimated risk cost for that grid.

On the other hand, if it is determined in step S14 that the viewing limitation area DA for the grid does not exist, the estimated risk cost for the grid is set to a predetermined reference set value (S17). Here, in the present invention, setting the reference set value to '0' is taken as an example.

Through the above process, when the registration of the estimated risk cost for one grid is completed, it is determined whether the registration of the estimated risk cost is completed for all grids constituting the grid map (S18). If there is a grating that needs to be calculated, the grating is moved (S19) to perform steps S11 to S17.

Here, the grid from which the estimated risk cost is calculated does not include all grids on the grid map. For example, in the following gradient method applied to the path generation process, the C-space is applied to correspond to the mobile robot 100 as one point so that the size of the mobile robot 100 is reflected. Depending on the application, grids in close proximity to fixed obstacles, for example walls such as walls, are excluded from the grid where the estimated risk costs are registered through the expansion of the fixed obstacles. That is, in the safety path generation method according to the present invention, the grid reflecting the risk cost due to the fixed obstacle may be excluded from the grid to be registered for the estimated risk cost.

FIG. 3 is a diagram illustrating an example of displaying, on the environment map, a grid in which a field of view restriction area DA exists, that is, a collision risk grating CDG through the above process, and FIG. 4 is a collision risk grating CDG. Is a diagram showing an example of the distribution of the maximum allowable speed with respect to).

Through the above process, when the registration of the estimated risk cost for the grids is completed, a moving path to the target location is generated by using a predetermined path generation technique using the estimated risk cost registered for each grid ( S20).

Here, as a path generation technique for generating a moving path in the safe path generation method according to the present invention, a gradient method using an intrinsic cost and an adjacent cost is applied. At this time, the gradient technique generates the moving path by applying the estimated risk cost registered for the grid as the unique cost.

Thus, the inherent cost is reflected in the inherent cost of the collision risk with the visually blocked dynamic obstruction DO, ie the estimated risk cost associated with the margin of distance to the viewing restricted area DA, so that the intrinsic cost is proportional to the inverse of the maximum permissible speed. As a result, a minimum time path can be obtained.

Hereinafter, an experimental result for confirming that the safety path generation method according to the present invention is efficient will be described with reference to FIGS. 5 to 11. First, for the evaluation of the experimental results, the performance measurement for evaluating the safety of the driving uses [Equation 6]. Here, all the regions are defined in the dynamic window.

&Quot; (6) "

In Equation 6, I _collision is a collision risk, and A _collision and A _safe represent a collision region and an allowable velocity region, respectively. When the velocity range where no collision occurs becomes smaller, I _collision can vary from 0 to 1.

If I _collision is close to 1, most of the velocity in the dynamic window is hunted by collision. Therefore, when I _collision approaches 1, it is necessary to modify the driving control technique, for example, to reduce the speed of the mobile robot 100. Here, I _collision is used to monitor the collision risk in the dynamic obstacle (DO) environment, and I _collision should be close to zero for safe driving.

5 shows collision risk for unexpected dynamic obstacles (DO). FIG. 5A illustrates sensor data when the dynamic obstacle DO exists in front of the mobile robot 100, and a clearance object is shown in FIG. 5B in the dynamic window. If the clearance object is smaller than the threshold value of 0.2, it is classified as the collision velocity region. And, it can be seen from (c) of FIG. 5 that the mobile robot 100 should rotate to the right by accelerating the left wheel to avoid the dynamic obstacle DO. At this time, the collision velocity region becomes very wide, and the I _collision becomes 0.78. Therefore, the mobile robot 100 can conclude that the dangerous state possible collision.

Experiments were conducted in the same building as the office. The maximum speed of the mobile robot 100 was set to 0.5 m / s, and the maximum acceleration was set to 0.8 m / s ² . In addition, the sampling time of the laser range finder installed in the mobile robot 100 is set to 0.2 s, and it is known as 1 m / s, which is the maximum walking speed of a normal pedestrian. The speed of the dynamic obstacle (DO) was set to 1.5 ~ 3m / s.

The control scheme for comparison with the safe path generation method according to the present invention uses a gradient technique combined with a DWA technique. In addition, three experiments were conducted with hair-pin navigation, doorway navigation, and narrow area passage. In addition, in the mobile robot 100 to which the path generation method according to the present invention is applied, the collision avoidance for the visible dynamic obstacle (DO) while driving is applied to the same DWA technique.

The experimental situation is that a dynamic obstacle DO emerges from the field of view restriction area DA to obstruct the movement of the mobile robot 100. The dynamic obstacle DO is set to move across the path of the mobile robot 100 starting from the field of view restriction area DA.

Experiment 1: Hair-Pin Navigation

First, experimental conditions for driving the hairpin are as shown in FIG. 6. In the hairpin driving experiment, the dynamic obstacle (DO) was set to move 3 m / s from the A position to the B position. The path PP of the mobile robot 100 and the path GP of the mobile robot to be compared are as shown in FIG. 6.

Although the path GP of the mobile robot to be compared is relatively short, it can be seen that the path GP passes near the wall, which is a fixed obstacle. Even in the case of speed control, the mobile robot 100 is moving along the gradient path generated under the proposed safe speed limit. And it was confirmed that the comparison target mobile robot stopped near the corner (refer to the red colored display in FIG. 6).

As described above, since the comparison robot moves through a very dangerous area, the maximum speed of the comparison robot moves to zero. As a result, the mobile robot to be compared stops before arriving at the destination and failed to arrive at the destination when the speed limit was applied to the path generated by the gradient technique.

FIG. 7 illustrates the velocity and collision risk I _collision of the mobile robot 100 during the above experiment. FIG. 7A is a diagram illustrating the speed and the collision risk I _collision of the mobile robot to be compared, and FIG. 7B is the speed and the collision risk I _collision when the safety speed limit is set to the mobile robot to be compared. 7 (c) is a diagram illustrating the speed and the collision risk I _collision of the mobile robot 100 according to the present invention.

As shown in (a) of FIG. 7, when the comparison target mobile robot encounters the dynamic obstacle DO near 8 seconds, the collision risk rapidly increased, which means that the path GP set in the comparison target mobile robot is dynamic. This means that safety from collision cannot be guaranteed.

In addition, as shown in (b) of FIG. 7, it can be seen that the comparison target mobile robot performs driving according to the safety speed limit. On the other hand, due to the slow speed, the collision risk remains at zero. However, the comparison target mobile robot fails to reach the destination because of the stop phenomenon at the corner portion as described above.

On the other hand, as shown in (c) of FIG. 7, the mobile robot 100 according to the present invention generates a path PP bypassing the corner portion, as shown in FIG. 6, and thus, is near 31 seconds. The collision risk due to the dynamic obstacle (DO) increased to 0.42, but the collision with the dynamic obstacle (DO) was avoided and continued driving. In addition, since the mobile robot 100 moves along a safe path, the speed of the mobile robot 100 is also increased. Although the total travel time of the mobile robot 100 according to the present invention exceeds 40 seconds, it may be evaluated as a method of providing a minimum time for satisfying safety.

Experiment 2: Doorway Navigation

Experimental conditions for the doorway driving is as shown in FIG. As shown in FIG. 2, the dynamic obstacle DO is set to move 1.5 m / s in the A to B direction. In the experimental conditions for driving the doorway shown in FIG. 8, it can be seen that there is not much space for changing the path of the mobile robot 100. Thus, speed limitation in driving of the mobile robot 100 may be regarded as the only way to avoid a collision.

In general, when the mobile robot 100 passes through the door, it is common to empirically slow down the speed. However, in the mobile robot 100 to which the safe path generation method according to the present invention is applied, the safe path generation method itself controls the speed of the mobile robot 100 without any empirical learning theory.

Referring to FIG. 8, it can be seen that the safety path generation method according to the present invention and the paths PP and GP according to the existing gradient technique are similar. However, it is clear that the mobile robot 100 according to the present invention lowered the speed before entering the door due to the speed limitation.

9 (a) is a diagram illustrating the speed and the collision risk of the mobile robot to be compared. When the mobile robot to be compared encounters the dynamic obstacle DO in the vicinity of 6 seconds, the collision risk increases rapidly and the speed of the mobile robot to be compared decreases. At this time, in a real situation, a state in which the comparison target mobile robot and the dynamic obstacle DO near each other near the collision occurs. As shown in FIG. 9A, the sensing data sensed by the laser distance sensor 13 such as the laser range finder of the mobile robot to be compared when it is about 6 seconds. Thus, in actual application, the mobile robot to be compared is difficult to guarantee safety from collision.

9 (b) is a view showing a speed profile of the mobile robot 100 according to the present invention. When the mobile robot 100 moves to the danger zone between 0 and 4 seconds, it can be seen that the speed is approximately 0.5 m / s. In addition, it may be confirmed that the speed is reduced to about 0.1 m / s through the door due to the visual field limitation of the mobile robot 100.

In the above experiments, the speed of the dynamic obstacle (DO) was set at a relatively fast 1.5 m / s, and the maximum collision risk was 0.08 throughout the experiment. This fact means that speed control is important for safe driving in doorway driving. Therefore, when the method for generating a safety route according to the present invention is applied, it is possible to ensure safe driving even in doorway driving without additional setting.

Experiment 3: Narrow Aisle Driving

10 shows experimental conditions for narrow passageway driving. As shown in FIG. 10, the conventional gradient technique provides a path GP through a narrow area between the wall and the column. This is because the gradient technique also approaches the shortest path like other conventional path generation methods.

As shown in FIG. 10, when the dynamic obstacle DO moves from A to B direction, as shown in FIG. 11A, the moving object to be compared is about 11 seconds due to the dynamic obstacle DO. The collision risk will reach 1. This indicates that the path by the existing gradient technique is extremely dangerous. FIG. 11B illustrates the mobile robot 100 operating under the safety speed limit, and the mobile robot 100 stops before entering a narrow area.

On the other hand, the safe path generation method according to the present invention, as shown in Figure 10, to provide a bypass path (PP). If the mobile robot 100 attempts to pass through a narrow area, the speed should be drastically reduced to avoid collisions, which causes a problem that the shortest path consumes longer travel time for safe driving. . As a result, the bypass path PP can be a safer and faster path. This can be confirmed through a diagram showing a speed distribution of the mobile robot 100 according to the present invention shown in FIG.

Hereinafter, the mobile robot 100 to which the safe path generation method according to the present invention is applied will be described with reference to FIG. 12. As shown in FIG. 12, the mobile robot 100 according to the present invention may include a path generator 10, a driver 12, a laser distance sensor 13, and a driving controller 20.

The route generation unit 10 generates a movement route to the target position according to the safety route generation method according to the present invention. Here, the method of generating the moving path by the path generation unit 10 is as described above, and a detailed description thereof will be omitted.

The driving unit 12 rotates a wheel which is a driving means of the mobile robot 100. Here, the driving unit 12 may be provided in the form of a motor for the rotation of the wheel, in the present invention, for example, a pair of motors for each rotating a pair of wheels is provided. The laser distance sensor 13 detects an obstacle in front of the laser beam by scanning the laser at a predetermined angle toward the front of the mobile robot 100.

The driving control unit 20 controls the driving unit 12 so that the mobile robot 100 travels according to the path generated by the path generation unit 10. Here, the driving controller 20 controls the driving unit 12 to travel by avoiding this when the obstacle is not registered on the grid map based on the sensing result detected by the laser distance sensor 13. In the present invention, the driving control unit 20 travels according to the DWA control technique.

In addition, the driving control unit 20 controls the traveling speed based on the maximum allowable speed described above. That is, as shown in FIG. 4, the maximum allowable speed is calculated for the danger area caused by the dynamic obstacle DO, and the driving controller 20 enters the danger area when the mobile robot 100 enters the danger area. By controlling the driving unit 12 so that the mobile robot 100 travels below the maximum allowable speed calculated for the above, the collision risk caused by the dynamic obstacle DO is avoided.

Although several embodiments of the present invention have been shown and described, those skilled in the art will appreciate that various modifications may be made without departing from the principles and spirit of the invention . It is intended that the scope of the invention be defined by the claims appended hereto and their equivalents.

100: mobile robot 10: path generation unit
11: driving control unit 12: driving unit
13: laser distance sensor

Claims (10)

In the safety path generation method considering the appearance of a dynamically blocked dynamic obstacle,
(a) calculating a field of view restriction area by an obstacle that is previously registered in each grid unit of the grid map;
(b) calculating an expected risk cost based on a predetermined obstacle speed for the dynamic obstacle and a distance between the grating and the viewing limited area when the viewing limited area exists for the corresponding grid;
(c) registering the estimated risk cost calculated in step (b) for the grid;
and (d) generating a movement route to a target position by using a predetermined route generation technique by using the estimated risk cost registered in each of the grids. The method of claim 1,
The path generation technique applied in the step (d) includes a gradient method using an intrinsic cost and an adjacent cost;
And the gradient method applies the estimated risk cost to the inherent cost to generate the travel path. The method of claim 2,
In the case of (b), if there is no view limit region for the grating, the estimated risk cost for the grating is set to a predetermined reference set value. The method of claim 3,
In step (a),
(a1) calculating a minimum collision distance based on the obstacle speed and the robot speed of the mobile robot;
calculating a reachable area for the lattice within the range of the minimum collision distance in the lattice unit;
(a3) calculating the visible area of the mobile robot within the minimum collision distance within the grid;
(a4) calculating the field of view restriction region for the grid, based on the deviation of the reachable region and the visible region. The method of claim 4, wherein
The minimum collision distance is
d _collision = d _delay + d _break + d _obs
d _delay = t _delay × (v _robot + v _obs )
d _break = v _robot ² / (2 × a _robot )
d _obs = v _obs × v _robot / a _robot
(Where d _collision is the minimum collision distance, t _delay is the motion delay time of the mobile robot, v _robot is the robot speed, v _obs is the obstacle speed, and a _robot is the acceleration of the mobile robot) Safe path generation method, characterized in that calculated by. The method of claim 4, wherein
And the reachable area is calculated by a wavefront propagation algorithm, and the visible area is calculated by a ray tracing method. The method of claim 3,
And the expected risk cost is calculated by reciprocal of the maximum allowable speed of the mobile robot calculated based on the obstacle speed and the distance between the grid and the viewing limited area. The method of claim 7, wherein
The maximum allowable speed is a quadratic equation
v _max ² / (2 × a _robot ) + (t _delay + v _obs / a _robot ) × v _max + t _delay × v _obs = d _dist
(Where v _robot is the maximum allowable speed and t _delay is the operation delay time of the mobile robot, v _obs is the obstacle speed, a _robot is the acceleration of the mobile robot, and d _dist is the grid and the field of view limitation) Safety distance generation method). A path generation unit for generating a moving path according to the safety path generation method according to claim 7 or 8;
And a traveling controller configured to travel along the movement path generated by the route generation unit. 10. The method of claim 9,
And the traveling controller controls the traveling speed based on the maximum allowable speed.

Images