JP2012243029A - Traveling object with route search function - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that although a technology of, when a self-position and a final target point are specified on map data, calculating a global route extending from the self-position via a movable range to the final target point is known, and a technology of, when any obstacle is detected, calculating a route to the final target point while avoiding the obstacle is also known, a technology of fusing the both technologies is unsolved, and a route to return to the global route after avoiding the obstacle is not calculable well.SOLUTION: An intermediate target point is used instead of a final target point to be used by an obstacle avoidance route calculation means. When a point on a global route which is far from an obstacle is set to the intermediate target point, a route to return to the global route after avoiding the obstacle is calculable.

Description

  The present invention relates to a moving body that moves from its own position at the start of movement to a final target point while searching for a route.

When the self-location and final target point are identified on the map data, a technology has been developed that calculates the route from the self-position to the final target point that extends within the movable range and moves along the calculated route. ing. Such a technique is disclosed in Patent Document 1. By the above technique, it is possible to move from its own position to the final target point while following a movable range determined from the map data. However, the above technique does not assume a situation where an unknown obstacle appears.
A technique for calculating a route that avoids an obstacle has also been developed, and is disclosed in Non-Patent Document 1. In this technology, multiple route candidates that can avoid obstacles are generated, and for each generated route candidate, an evaluation value is calculated based on the angle that is the final target point, the movable speed, and the distance to the obstacle. Adopt the candidate that yields the maximum evaluation value.

Description and drawing of Japanese Patent Application No. 2010-187044

Mobile robot online avoidance action planning based on pedestrian behavior prediction, Fumiko Kubota, Kei Tanaka, Yuji Tsusaka, 15th Robotics Symposia Proceedings, 2010, pp. 1-6

According to the technique disclosed in Patent Document 1, a route extending from the self-position within the movable range to the final target point can be calculated from the map data, the self-location, and the final target point, along the calculated route. Can move. However, if an obstacle that is not stored in the map data appears while moving along the calculated route, the obstacle cannot be avoided.
According to the technique disclosed in Non-Patent Document 1, a route for avoiding an obstacle can be calculated.
Therefore, if the technique of Patent Document 1 and the technique of Non-Patent Document 1 are used in combination, an obstacle while moving along the path extending from the self-position at the start of movement within the movable range to the final target point. When it appears, it seems that the obstacle can be avoided.

  However, in practice, it is difficult to use the technique of Patent Document 1 and the technique of Non-Patent Document 1 in combination. In the technique of Non-Patent Document 1, an evaluation value is calculated based on an angle formed by a direction along a path that avoids an obstacle and a direction toward a final target point. In other words, by using the position of the final target point when calculating the avoidance route, the route to the final target point is calculated while avoiding the obstacle. When the technique of Patent Document 1 and the technique of Non-Patent Document 1 are used together, after avoiding an obstacle, the route information extending within the movable range is ignored, and a route that goes straight to the final target point is calculated. It may deviate from the movable range. That is, it may come into contact with a wall or the like that is the boundary line of the movable range. Alternatively, the moving body may stop to avoid contact with a wall or the like.

  Instead of the final target point used in the technique of Non-Patent Document 1, if an intermediate point on the route extending in the movable range is set as a temporary target point, obstacles can be avoided and You can imagine both moving along a path that extends.

  However, it is actually difficult. When you actually try, both the element that avoids the obstacle and the element that returns to the path act together, and the movement direction that avoids the obstacle and the movement direction that approaches the obstacle trying to return to the path alternate It is calculated that. In order to use the technology of Patent Literature 1 and the technology of Non-Patent Literature 1 in combination, a technology for merging both is required.

  In the present invention, the map data is used to calculate and move a route extending from the self-position within the movable range to the final target point, and when an obstacle appears, a route for avoiding the obstacle is calculated. Fusing the moving technology, move along the route until an obstacle appears, move along the route avoiding the obstacle if an obstacle appears, and return to the route after avoiding the obstacle A moving body that moves along a route is provided. In this specification, a route calculated from map data and extending within the movable range from its own position to the final target point is referred to as a global route. Distinguish from the route returning to the route.

The present invention relates to a moving body that moves along a route to a final target point while avoiding an obstacle when an obstacle not included in map data is detected. The mobile body includes a map database, a global route calculation means, an obstacle detection device, a local route calculation means, and a control device. The map database stores data that defines a movable range. The global route calculation means calculates a global route extending from the self-location specified on the map data to the final target point and extending within the movable range. The obstacle detection device specifies the position of the obstacle that appears at least when an obstacle appears in front of the moving direction of the moving body. The local route calculation means calculates a route that avoids the obstacle and returns to the intermediate target point. The control device controls the mobile device included in the mobile body to actually move the mobile body along the route calculated by the local route calculation means.
In the present invention, the intermediate target point used in the local route calculation means is the following condition:
1) It is on the global route calculated by the global route calculation means,
2) Set to a position that satisfies the distance to the intermediate target point measured along the global path> the distance to the obstacle measured along the global path.

  Known techniques can be used for the global route calculation means. For example, the technique described in Patent Document 1 can be used. You may utilize the route search technique etc. which are used with the navigation apparatus mounted in the motor vehicle. Known techniques can also be used for local route calculation means. For example, the technique described in Non-Patent Document 1 can be used. However, in the present invention, the local route is calculated using the intermediate target point instead of the final target point. The intermediate target point used here is a point on the global path. As a result, a local route that returns to the global route is calculated. Further, an intermediate target point that satisfies the condition of the distance to the intermediate target point measured along the global path> the distance to the obstacle measured along the global path is employed. In the above distance relationship, it is highly appreciated that the obstacle avoids the obstacle and passes the side of the obstacle, and the local route is calculated. After passing, it is highly evaluated that it is directed to the intermediate target point, and the local route is calculated. An avoidance route is calculated during a period in which an obstacle needs to be avoided, and a route to return to the global route is calculated after the need to avoid an obstacle has diminished. Global route calculation technology and technology to calculate local routes that avoid obstacles are well integrated.

When setting the intermediate target point used in the local route calculation means, the following conditions are further met:
3) It is preferable that the distance to the intermediate target point measured along the global path includes a distance proportional to the relative speed of the moving object and the obstacle measured along the global path.
In this case, when the moving object and the obstacle approach at high speed, the local route is calculated with the remote point as the intermediate target point, and when the moving object and the obstacle approach at low speed, The local route is calculated using the determined point as an intermediate target point. As a result, a natural route, that is, a local route that can be avoided without causing pedestrians to feel anxiety and returns to the global route without difficulty after the avoidance is calculated.
The distance to the intermediate target point along the global path is preferably 6 meters or more. In that case, when the mobile body moves avoiding the pedestrian, a local route that does not enter the personal space of the pedestrian can be calculated.

  According to the present invention, the technology for calculating a global route extending from the movable position to the final target point within the movable range and the technology for calculating a local route to avoid an obstacle are well integrated, and along the global route. A route that follows the global route while it can move, moves to a local route that avoids the obstacle if an obstacle appears and the obstacle needs to be avoided, and returns to the global route after avoiding the obstacle Is calculated.

The figure explaining the function of the various apparatuses with which a mobile body is provided, and various apparatuses. The figure which shows the detail of the various apparatuses with which a moving body is provided. The processing procedure figure which calculates a global route. The process sequence diagram repeatedly performed at the time of a movement. The figure explaining the path | route obtained by the process of FIG.

The main features of the embodiments described below are listed.
(Feature 1) The moving body is self-propelled.
(Feature 2) An external sensor for detecting a person is mounted on the moving body.
(Feature 3) The moving body is equipped with an input device for the passenger to input the final target point.
(Feature 4) The moving body is equipped with a computer and moves by calculating a route.
(Characteristic 5) The moving body calculates a route that avoids a personal space that is a space around a person and that the person feels uneasy when another person enters.

  FIG. 1 shows a self-propelled mobile body 2 that embodies the mobile body of the present invention. The moving body 2 includes a moving device 14 and a control device 12. The moving device 14 includes a left wheel and a left wheel motor, and a right wheel and a right wheel motor. The control device 12 calculates the rotation speed of the left wheel motor and the rotation speed of the right wheel motor from the values of the moving speed of the moving body 2 and the angular speed of the moving body 2 instructed by the control apparatus 12, and calculates the calculated rotation speed. Adjust to. The control device 12 controls the rotation speeds of the left wheel motor and the right wheel motor independently. The moving body 2 moves at a moving speed instructed by the control device 12 and changes the moving direction at an angular velocity instructed by the control device 12. The moving body 2 moves along a route instructed to the control device 12. In this embodiment, the moving direction is changed by independently controlling the rotation speeds of the left and right wheels. However, when the steering device is provided, the moving direction is changed by adjusting the direction of the steering wheel. Also good. The mechanical configuration of the moving body 2 is disclosed in Non-Patent Document 1.

  The moving body 2 moves by calculating the global path 18 that moves within the movable range 28 that is delineated by the boundary line 20 such as a wall and reaches the final target point 38. Within the movable range 28, for example, there are stationary obstacles such as a parked automobile 24, and there are moving obstacles such as pedestrians 32 and 36, and the mobile body 2 is connected to local routes 26, 30, Calculate 34 and move. As apparent from FIG. 1, the moving body 2 returns to the global path 18 that reaches the final target point 38 when the obstacle 2 is avoided.

The moving body 2 includes a map database 4 that stores a geometric pattern of the boundary line 20. The moving body 2 includes a current position detection device (self-position estimation unit in FIG. 2), and can determine where on the map. The moving object 2 can be taught a final target point, and it can be determined where the taught final target point is on the map. For example, it can be specified that the current position of the moving body 2 is 16 and the final target point is 38. When the self-position and the final target point are specified, the moving body 2 moves from the self-position 16 to the final target point 38 under the condition that it does not collide with the boundary line 20 (in other words, under the condition that it extends within the movable range 28). A calculation means for calculating the extending path 18 is provided. The path so calculated is a global path than the local path calculated to avoid detected obstacles. The moving body 2 includes a global route calculation means 6 that calculates a global route 18 from the map database 4, its own position, and the final target point.
The global path calculation means 6 may calculate the global path 18 from the self-position at the start of movement and the final target point, and may not calculate again thereafter. Alternatively, the global path 18 may be recalculated in response to the update of the self-position with movement. FIG. 3 shows processing executed by the global route calculation means 6 of the former method.

  The moving body 2 includes an obstacle detection device 8. The obstacle detection device 8 emits a laser and receives reflected light, and can specify the distance and direction to the reflector. Radio waves can also be used instead of lasers. Reference numeral 22 indicates a range in which the obstacle detection device 8 can detect the presence of an obstacle. It is not possible to detect even shadowed parts such as walls. In the case of FIG. 1, the parked automobile 24 is detected, and the pedestrians 32, 36, etc. are not detected.

  The moving body 2 includes a local route calculation means 10. The local route calculation means 10 calculates a local route that avoids an obstacle when an obstacle is detected. In the case of FIG. 1, since the obstacle 24 is detected, the local route 26 is calculated. When the obstacle 32 is detected, the local route 30 is calculated. When the obstacle 36 is detected, the local route 34 is calculated. This is an example. As illustrated in the local paths 26, 30, and 34, the local paths 26, 30, and 34 calculated in this embodiment return to the global path 18 immediately after the mobile body 2 avoids the obstacle. It is characterized by that. In the present embodiment, moving along the global route 18 moves to the final target point 38, and adopting the local routes 26, 30, and 34 that avoid the obstacles to move around the obstacles. Is compatible.

  FIG. 2 shows the system configuration of the moving body 2. The final target point designating unit 40, the self-position estimating unit 42, and the map database used by the operator to teach the final target point 38 to the moving body 2. 4 is provided. The final target point designating unit 40 may display a map and designate the position of the final target point on the displayed map. The self-position estimation unit 42 may be a GPS or the like, may be one that estimates the self-position from the number of rotations of the wheel, or the boundary line 20 such as a wall detected by the obstacle detection device 8. The self-position on the map may be estimated from the shape. Techniques such as dead reckoning and scan matching can be used. The global path calculation means 6 calculates a path from the self position to the final target point that does not interfere with the boundary line. In the case of FIG. 1, the global route calculation means 6 calculates a global route 18 that is a route from the self-position 16 to the final target point 38 and that does not interfere with the boundary line 20.

The global route calculation means 6 calculates a global route by a known method. For example, Dijkstra search technology, A ^* search technology, or the like can be used. When there are a plurality of routes from the self-position 16 to the final target point 38, it is possible to divert the search conditions adopted by a commercially available navigation device. For example, the search condition can be set so that distance priority is designated or a route having a narrow road is avoided. The global path 18 is more detailed than the route, and specifies where in the road width to follow in order to avoid interference with the boundary line 20.

The moving body 2 includes an external sensor 8a. The external sensor 8a includes a light emitting / receiving element for laser light, and scans the light emitting direction of the laser light in a horizontal plane so as to detect an obstacle in the detection range 22 of FIG. When there is an obstacle in the detection range 22 shown in FIG. 1, the external sensor 8a receives reflected light from the obstacle. The distance to the obstacle can be calculated from the laser light emission timing and the light reception timing, and the direction in which the obstacle exists can be calculated from the laser light emission direction.
Among objects detected by the external sensor 8a, fixed objects such as walls and the like are included in the map database 4, and temporary objects such as parked vehicles and pedestrians are included in the map database 4. Contains no objects. In the present specification, the latter is referred to as an obstacle. Among obstacles, there are obstacles that are stationary such as parked vehicles and obstacles that are moving like pedestrians.

The obstacle extracting unit 8b extracts obstacles by excluding objects included in the map database 4 from objects detected by the external sensor 8a.
The obstacle trajectory calculation unit 8c extracts a characteristic pattern (future) for each obstacle from the spatial distribution pattern of the reflected light detected by the external sensor 8a. Detection by the external sensor 8a and extraction processing by the obstacle extraction unit 8b are repeatedly performed at predetermined time intervals. The position, moving speed, and moving direction for each obstacle are calculated from the positions of the features obtained every predetermined time. The obstacle trajectory calculation unit 8c calculates a movement trajectory for each obstacle. If it is the parked vehicle 24, it turns out that it is still, and if it is the pedestrians 32 and 36, it will be understood which position is moving in which direction at what speed. Details of a technique for extracting an obstacle and calculating a position, a moving speed, and a moving direction for each obstacle are disclosed in Non-Patent Document 1.
In the present embodiment, a laser light emitting / receiving element is used as the external sensor 8a. However, a radar that uses radio waves may be used, or a stereo camera that receives visible light may be used. Moreover, it is not necessarily mounted on the moving body 2 and may be a stationary camera that monitors the movable range 28.

The global route calculated by the global route calculation means 6 is input to the intermediate target point setting unit 10a. The intermediate target point setting unit 10a sequentially sets intermediate target points on the global path with the following logic.
1) When the global route extends straight, the distance between the intermediate target point and the adjacent intermediate target point is set as a predetermined distance. In this embodiment, an intermediate target point is set every 10 cm.
2) When the global path is curved, the distance between the intermediate target point and the adjacent intermediate target point is made shorter than the predetermined distance. At this time, the shorter the route, the shorter the distance.
If the intermediate target point is set as described above and moved to the next intermediate target point every predetermined time, the moving body 2 moves on the global path 18 at a high speed on a straight path, and decelerates on a curve. You can move while.

  Information on the set intermediate target point and the calculated trajectory of the obstacle are input to the local route calculation unit 10b. The local route calculation unit 10b calculates a local route that avoids an obstacle using the technique described in Non-Patent Document 1. At this time, the local route calculation unit 10b does not obtain a route to the final target point 38 while avoiding the obstacle, but calculates a local route that returns to the global route 18 when the obstacle is avoided. For this purpose, the local route calculation unit 10b uses the position of the return intermediate target point for calculation instead of the position of the final target point 38.

FIG. 5 shows that the moving body 2 is on the position P _{0 at the} time t ₀ (the position P ₀ is on the global path 18), and the obstacle 50 is at the position P ₂ and moves along the vector V _2. This is an example of a state that is found to be. The obstacle 50 is estimated to move to the positions indicated by t ₁ , t ₂ , t ₃ , and t ₄ over time. FIG. 5 illustrates a case where it is expected that when the moving body 2 continues to move along the global path 18, it interferes with the obstacle 50, and it is necessary to switch to a path that avoids the obstacle 50.

The local route calculation unit 10b calculates whether or not the mobile body 2 and the obstacle 50 interfere with each other, and if it is found that the mobile body 2 and the obstacle 50 interfere with each other, calculates the relative speed between the mobile body 2 and the obstacle 50. To do.
The relative velocity here is calculated from the velocity component V _2c along the global path 18 in the velocity vector V ₂ of the obstacle 50 and the velocity component V _1c along the global path 18 of the moving body 2. Yes, when the speed component V _2c approaches the moving body 2, it is a value obtained by adding both. When the speed component V _2c is in the same direction as the speed component V _1c , V _2c is subtracted from V _1c . Value.

Local route calculation section 10b, the relative speed V, to calculate the distance L ₂ to the intermediate target point by the following equation.
L ₂ = α + β × relative speed V (1)
The above α and β are values for calculating the distance L ₂ having a relationship of L ₂ > L _{1 with} respect to the distance L ₁ to the obstacle 50 that may interfere. α and β may be constants. Further, α may be a value that increases or decreases depending on the speed of the moving body. Alternatively, a larger α is adopted as the locality radius of the route is larger (moving at a high speed along a route close to straight traveling), and a smaller α as the locality radius of the route is smaller (moving along a sharp curve at a lower speed). May be adopted. When the distance L ₂ is calculated to extract the nearest intermediate target point in the calculated distance L _2, the extracted intermediate target point to the intermediate destination point for the return. The local route calculation unit 10b calculates a local route that returns to the global route at the return intermediate target point.
The relative speed V is a positive value when the moving body 2 and the obstacle 50 approach each other. As a result of approaching the final target point, if L ₂ > distance to the final target point, the final target point is set as an intermediate target point for return used for local calculation. When the moving body 2 and the obstacle 50 are separated and take a negative value, there is no possibility that the moving body 2 and the obstacle 50 interfere with each other.

  When the mobile object 2 and the obstacle 50 do not interfere with each other, the local route calculation unit 10b calculates a local route that follows the intermediate target point set by the intermediate target point setting unit 10a every predetermined time. When the mobile body 2 and the obstacle 50 interfere with each other, the local route calculation unit 10b calculates a route for avoiding the obstacle by using the technique of Dynamic Window Approach described in Non-Patent Document 1. At that time, instead of the final target point, a return intermediate target point in the vicinity of the distance L2 is used. As a result, the route is calculated with emphasis on avoiding the obstacle until it passes the side of the obstacle, and the route is calculated with an emphasis on returning to the global route when passing the side of the obstacle. The A local route that returns to the global route at the intermediate target point selected for return is calculated.

In the technique of Non-Patent Document 1, a plurality of travel direction candidates and travel speed candidates are created, a preference index is obtained for each candidate, and a travel direction candidate and a travel speed candidate that provide the highest index are adopted. Use logic. For example, when the moving body 2 is at the position indicated by P ₀ , candidates for the traveling directions exemplified by A to G (A to G differ from each other by a predetermined angle) and the level of traveling speed set in a plurality of stages A candidate for a progression vector obtained by a combination of a plurality of candidates (for example, three candidates of low speed, medium speed, and high speed) is set. Next, for each progression vector candidate, calculate the distance to the obstacle, the angle to be the vector pointing to the intermediate target point, the change angle from the current traveling direction, etc., and the distance to the obstacle The longer the value, the higher the score, and the smaller the angle, the higher the score given to the vector pointing to the intermediate target point, and the smaller the change angle, the higher the score. Use the candidate that gives the highest score. A route that has a long distance to the obstacle, is directed to the intermediate target point for return as much as possible, and has little change in the traveling direction is adopted. For Figure 5, it illustrates a case where the time t ₀ the vector V ₁₁ is the highest score. The moving body 2 proceeds while repeating the above calculation every predetermined time. FIG. 5 illustrates the case where the vector V ₁₂ has the highest score at time t ₁ and the case where the vector V ₁₃ has the highest score at time t ₂ .

In the case of the present embodiment, the trajectory of the obstacle 50 is calculated on the moving body 2 side, and the future position of the obstacle 50 can be predicted. Thus, at time _{t 0,} it is possible to predict the positional relationship at the time _t 1, _{t 2,} etc., at time _{t 0,} it is also possible to determine the progress vector definitive like time _t 1, _{t 2.} For example, when calculating the local route at time t ₀ , estimating the position where the obstacle 50 will exist at time t ₁ , the position where the obstacle 50 will exist at time t ₂ , etc. it can. Therefore, when calculating the local route at time t ₀ , the mobile unit 2 determines the possible travel direction candidates and travel speed candidates at time t ₀ , and possible travel direction candidates and travel speed at time t ₁ . Candidates, candidates for the direction of travel possible at time t ₂ , candidates for speed of travel, etc. are calculated, the candidates at time t ₀ and their scores, the candidates at time t ₁ and their scores, the candidates at time t ₂ and their scores, etc. Calculate and calculate the route that gives the highest total score. For figure adopts the time _{t 0} traveling vector _{V 11} In employs a time _{t 1} traveling vector _{V 12} In employs a time _{t 3} traveling vector _{V 13} In, adopted progression vector _{V 14} at time _{t 4} and, when by a route employing a time t ₅ traveling vector V ₁₅ in exemplifies a case where the total score is the highest. In this case, the local route calculation means 10 calculates a local route connecting the progress vectors V ₁₁ , V ₁₂ , V ₁₃ , V ₁₄ , and V ₁₅ . In the present embodiment, at time t ₀ , the local route with the highest total score over the time required for the moving body 2 to move to the return intermediate target point M ₅ set at that time is calculated. Calculating a local path on the basis of the movement locus of the obstacle between the up returning intermediate target point M _5.

The above calculation may be repeated every predetermined time. For example, at time _{t 0,} calculate the _{V 11, V 12, V 13} , V 14, V 15 in FIG. 5, again executes the computation of the same type at time _{t 1.} As a result, the velocity vectors V ₁₂ , V ₁₃ , V ₁₄ and V ₁₅ shown in FIG. 5 are recalculated, and V ₁₆ is newly calculated.
According to the latter method, it is possible to follow that the obstacle changes its moving speed or moving direction.

FIG. 4 shows a processing procedure executed by the local route calculation unit 10b when the mobile body 2 and the obstacle 50 interfere. While the mobile body 2 is moving, the process of FIG. 4 is repeatedly executed at predetermined time intervals.
In step S40, the self position of the moving body 2 is detected. In step S42, the self-speed (to be precise, the direction) is calculated from the amount of change from the self-position detected before a predetermined time. In step S44, the position of the obstacle is detected. The obstacle position detected in step S44 is a relative position with respect to the moving body 2, but since the self position of the moving body 1 is detected in step S40, the position of the obstacle on the map can be known. In step S46, the moving speed (to be exact, the moving direction) of the obstacle is calculated from the amount of change from the obstacle position detected a predetermined time ago.

In step S48, the relative speed in the direction along the global path 18 described with reference to FIG. 5 is calculated. In step S50, the distance to the intermediate target point used by the local route calculation means 10 is calculated using Equation 1. In step S52, the distance between the distance to the final target point and the distance calculated in step S50 are compared. When approaching the final target point, L ₂ > distance to the final target point. If the condition is satisfied (YES in S52), the final target point is set as a return intermediate target point used for local calculation (S54). During the distance to L ₂ <the final target point (while NO in S52), the intermediate target point at the distance L2 is set as the intermediate target point for return used for the local calculation (S56).
In step S58, a local route that avoids an obstacle is calculated using the technique of Non-Patent Document 1. Using the return intermediate target point set in step S54 or S56, the local route is calculated using the movement trajectory of the obstacle until the moving body reaches the return intermediate target point. When the local route is calculated, the rotation speed of the left motor and the rotation speed of the right motor that are the routes are calculated and adjusted to the rotation speed.

When the local route calculation means 10 uses the return intermediate target point instead of the final target point, the local route that returns to the global route 18 at the return intermediate target point is highly evaluated.
In order for the local route to avoid the obstacle to be calculated while it is necessary to avoid the obstacle, and to return to the global route once the obstacle has been avoided, the local route is calculated. The selection of the return intermediate target point used in the route calculation means 10 is important. If the return intermediate target point is too close, that is, if a return intermediate target point is set between the moving object and the obstacle, the local route that avoids the obstacle is calculated while the obstacle needs to be avoided. And the element that calculates the local route to return to the global route act simultaneously. The latter will calculate the local route towards the obstacle. When an intermediate target point is set between a moving object and an obstacle, while it is necessary to avoid the obstacle, calculate the local route to avoid the obstacle and the local route to the obstacle. To act simultaneously. Since the local route is calculated while the route avoiding the obstacle is highly evaluated and the route toward the obstacle is highly evaluated, the actually obtained local route is complicated and confusing.

In the present embodiment, α and β are set to values at which a distance L ₂ larger than the distance L ₁ to the obstacle is calculated, and the distance calculated by the formula L ₂ = α + β × relative speed V is used. In order to calculate a local route using an intermediate target point, while avoiding an obstacle, the route avoiding the obstacle is highly evaluated and the local route is calculated, and the obstacle is avoided. Then, the route returning to the global route is highly evaluated and the local route is calculated.

As a result of the above,
1) Move toward the final target point along the global path.
2) When an obstacle is detected during movement, a local route is searched while highly evaluating a route that avoids interference with the obstacle. In this case as well, elements that return to the return intermediate target point on the global path are taken into account, but since the return intermediate target point is set behind the obstacle, Until it passes, the element that returns to the intermediate target point does not search for a local route that causes the moving body to enter the front of the obstacle.
3) After passing the side of the obstacle, a route for quickly returning to the return intermediate target point is searched. Never lose sight of the target while avoiding obstacles.
4) After returning to the global route, the robot moves to the final target point along the global route.
As a result of the above, the moving body can achieve both the movement to the final target point along the global path and the avoidance of the obstacle.

In the case of the technique of recalculating the local route while moving along the local route, as shown by the detection range 22 in FIG. Obstacles are no longer detected at the stage of rushing to the side of the obstacle. In this case, the relative speed V in the equation L ₂ = α + β × relative speed V is zero. When the obstacle is no longer detected, the value of the distance L2 changes discontinuously. In order to avoid this, it is effective to use a sensor that detects the back of the moving body 2. Alternatively, instead of setting the relative speed V to zero when no obstacle is detected, an artificial relative speed that changes to zero at a constant speed may be used.

  The obstacle may be a pedestrian. Pedestrians have a personal space that makes them feel anxious or uncomfortable when they are approached by others. 5 in FIG. 5 indicates a personal space of the pedestrian 50. In the case of the present embodiment, when an intermediate target point located 6 meters or more along the global route is selected as an intermediate target point for return, not only does the moving body 2 interfere with the pedestrian 50 but also the pedestrian 50 Local routes that do not enter the personal space 52 can be calculated.

  Specific examples of the present invention have been described in detail above, but these are merely examples and do not limit the scope of the claims. The technical elements described in this specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. In addition, the technology illustrated in the present specification or the drawings can achieve a plurality of purposes at the same time, and has technical utility by achieving one of the purposes.

2: Mobile body 4: Map database 6: Global route calculation means 8: Obstacle detection device 10: Local route calculation means 12: Control device 14: Mobile device 16: Current position 18: Global route 20: Boundary line 22: Detection Range 24: Parked vehicles (stationary obstacles)
26, 30, 34: Local route 28: Range of movement 32, 36: Pedestrian (moving obstacle)
38: Final target point

Claims (3)

It is a moving body that moves along the path to the final target point while avoiding the obstacle when it detects the obstacle,
A map database storing data defining a movable range;
A global route calculating means for calculating a global route extending from the self-location specified on the map data to the final target point and extending within the movable range;
An obstacle detection device for identifying the position of an obstacle that appears when an obstacle appears at least in front of the moving direction of the moving body;
A local route calculating means for calculating a route for avoiding an obstacle and returning to an intermediate target point;
A control device that controls the mobile device provided in the mobile body and moves the mobile body along the route calculated by the local route calculation means;
The intermediate target point used in the local route calculation means is the following condition:
1) It is on the global route calculated by the global route calculation means,
2) A moving object characterized in that it is set at a position satisfying the following: distance to intermediate target point along global path> distance to obstacle along global path.   2. The moving body according to claim 1, wherein the distance to the intermediate target point along the global path includes a distance proportional to a relative speed between the moving body and the obstacle along the global path.   The moving body according to claim 1 or 2, wherein the distance to the intermediate target point along the global path is 6 meters or more.

Images