Control system and control method

Abstract

The control system (10) comprises: an acquisition unit (11) for acquiring sensing data output by a sensor, wherein the sensor is a sensor for sensing the outside of a moving body; a detection unit (12) for detecting the position of an object outside the moving body based on the sensing data; a movement prediction unit (13) for predicting the movement of the object based on the sensing data; a cost map generation unit (14) for generating a first cost map based on the position of the detected object and a second cost map based on the predicted movement of the object; a path generation unit (15) for generating a path of the moving body based on the first cost map; a judgment generation unit (16) for generating a movement judgment of the moving body based on the second cost map and the generated path; and a control unit (17) for controlling the movement of the moving body according to the generated path and the movement judgment.

Description

Technical Field

[0001] The present disclosure relates to a control system and a control method for controlling movement of a moving object. Background Art

[0002] Conventionally, there has been disclosed a driving support device (for example, Patent Document 1) that predicts the movement of an object such as an obstacle outside the vehicle and enables the vehicle to travel safely by avoiding a collision with the object outside the vehicle.

[0003] In addition, it is possible to use a cost map that represents objects outside the vehicle as costs, generate a vehicle path that avoids high-cost areas, and drive the vehicle safely along the generated path (for example, Non-Patent Document 1).

[0004] (Prior art literature)

[0005] (Patent Document)

[0006] Patent Document 1: Japanese Patent No. 5617677

[0007] Non-patent literature 1: A novel navigation scheme in dynamic environment using layered costmap, https: / / ieeexplore.ieee.org / abstract / document / 7978468

[0008] However, a path generated using a cost map that takes the movement of objects into consideration may sometimes impair the safety or comfort of movement of moving objects such as vehicles. Summary of the invention

[0009] Therefore, the present disclosure provides a control system and the like that can improve the safety or comfort of movement of a moving object using a cost map.

[0010] The control system involved in the present disclosure comprises: an acquisition unit, which acquires sensing data output by a sensor, wherein the sensor is a sensor for sensing the outside of a moving body; a detection unit, which detects the position of an object outside the moving body based on the sensing data; a movement prediction unit, which predicts the movement of the object based on the sensing data; a cost map generation unit, which generates a first cost map based on the detected position of the object, and generates a second cost map based on the predicted movement of the object; a path generation unit, which generates a path of the moving body based on the first cost map; a judgment generation unit, which generates a movement judgment of the moving body based on the second cost map and the generated path; and a control unit, which controls the movement of the moving body according to the generated path and the movement judgment.

[0011] In addition, these general or specific solutions may be implemented by systems, methods, integrated circuits, computer programs, or computer-readable recording media such as CD-ROMs, or by any combination of systems, methods, integrated circuits, computer programs, and recording media.

[0012] By using a control system and the like according to one aspect of the present disclosure, it is possible to improve the safety or comfort of movement of a moving object using a cost map. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1A This is a diagram for explaining the problem when the motion prediction of an object is reflected in the cost map.

[0014] Figure 1B This is a diagram for explaining the problem when the motion prediction of an object is reflected in the cost map.

[0015] Figure 2 This is a block diagram showing an example of a control system according to the first embodiment.

[0016] Figure 3 This is a flowchart showing an example of the operation of the control system according to the first embodiment.

[0017] Figure 4 This is a diagram showing an example of the first cost map and the second cost map.

[0018] Figure 5 This is a diagram showing an example of a path generated based on the first costmap.

[0019] Figure 6 This is a diagram showing an example of movement judgment generated based on the second costmap and the generated path.

[0020] Figure 7 This is a diagram showing an example of a path generated based on a part of the second costmap.

[0021] Figure 8 This is a block diagram showing an example of a control system according to the second embodiment. DETAILED DESCRIPTION

[0022] (The process of obtaining a solution of the present disclosure)

[0023] First, regarding the process of obtaining a solution of the present disclosure, using Figure 1A as well as Figure 1B Provide explanation.

[0024] Figure 1A as well as Figure 1BThis is a diagram for explaining the problem when the motion prediction of an object is reflected in the cost map.

[0025] As mentioned above, cost maps often involve static environments. In the case of dynamic environments, it is possible to consider reflecting the movement prediction of objects such as obstacles outside the moving body (such as a vehicle) into the cost map to generate the vehicle's path. For example, it is possible to consider increasing the cost of the area where the object is moving based on the predicted movement of the object, and generating the vehicle's path in a way that avoids the high-cost area. However, when the object's movement prediction is reflected in the cost map, Figure 1A as well as Figure 1B Description of the problem.

[0026] exist Figure 1A The diagram shows that the time has moved forward by, for example, 0.1 seconds in the order of the left diagram, the middle diagram, and the right diagram. Figure 1A At the moment of the left figure of , the action prediction of vehicle 200 is predicted to be straight, and the prediction result is reflected in the cost map. Based on the cost map reflecting the prediction result, a path is generated in which the direction of travel of vehicle 100 is deviated to the left in order to avoid vehicle 200 going straight. However, action prediction is unstable in reality. For example, Figure 1A At the moment of the left figure, the vehicle 200 is predicted to go straight, but at the moment of the middle figure 0.1 seconds later, the vehicle 200 is predicted to turn to the right. Figure 1A At the time of the middle figure, based on the cost map reflecting the prediction result, a path is generated for the vehicle 100 to turn its traveling direction to the right in order to avoid the vehicle 200 that has turned its traveling direction to the right. Figure 1A At the time of the middle figure of , the vehicle 200 is predicted to be moving to the right, and at the time of the right figure 0.1 seconds later, the vehicle 200 is predicted to be moving to the left. Figure 1A At the time of the figure on the right side of , a path is generated for the vehicle 100 to deviate to the left in order to avoid the vehicle 200 that deviates to the left in the direction of travel, based on the cost map reflecting the prediction result. Therefore, a path that changes from time to time is generated, and the vehicle 100 frequently zigzags, causing problems with the driving safety or comfort of the vehicle 100.

[0027] also, Figure 1BAs shown below, it is predicted that vehicle 200 will go straight, and the prediction result is reflected in the cost map. Based on the cost map reflecting the prediction result, a path is generated for vehicle 100 to deviate the traveling direction to the right in order to avoid vehicle 200 going straight. In this case, it is safer for vehicle 100 to park, but vehicle 100 actively avoids vehicle 200, so that vehicle 100 approaches vehicle 200, causing problems in the driving safety of vehicle 100.

[0028] Therefore, a control system involved in an embodiment of the present disclosure comprises: an acquisition unit, which acquires sensing data output by a sensor, wherein the sensor is a sensor for sensing the outside of a moving body; a detection unit, which detects the position of an object outside the moving body based on the sensing data; a movement prediction unit, which predicts the movement of the object based on the sensing data; a cost map generation unit, which generates a first cost map based on the detected position of the object, and generates a second cost map based on the predicted movement of the object; a path generation unit, which generates a path of the moving body based on the first cost map; a judgment generation unit, which generates a movement judgment of the moving body based on the second cost map and the generated path; and a control unit, which controls the movement of the moving body according to the generated path and the movement judgment.

[0029] Through the above, the first cost map based on the position of the detected object and the second cost map based on the predicted movement of the object are generated respectively, and each cost map is used in generating the path of the moving body and generating the movement judgment of the moving body. Specifically, in the generation of the path of the moving body, the first cost map that does not reflect the movement prediction of the object is used, and in the generation of the movement judgment of the moving body, the second cost map that reflects the movement prediction of the object is used. In the generation of the path of the moving body, the first cost map that does not reflect the movement prediction is used, and even in the case of unstable movement prediction, it is difficult to generate a path that changes at all times, which can improve the safety or comfort of the movement of the moving body. In addition, in the generation of the movement judgment of the moving body, the second cost map that reflects the movement prediction of the object and the generated path are used, so as to avoid stationary objects, and the moving objects are not forcibly avoided. According to the movement prediction of the object, for example, the moving body can be stopped or decelerated, which can improve the safety of the movement of the moving body. In this way, on the premise of unstable movement prediction, a stable path is generated, and it is stopped or decelerated when necessary. Therefore, through this scheme, the driving safety or comfort of the vehicle using the cost map can be improved.

[0030] Furthermore, when an area assigned a cost greater than a predetermined cost in the second cost map overlaps with the generated path, the movement judgment may include a judgment of stopping the moving body or decelerating the moving body.

[0031] When an area assigned a cost greater than a predetermined cost in the second cost map overlaps with the generated path, the moving body may collide with a moving object when moving along the generated path. Therefore, in this case, a movement judgment is generated to stop or slow down the moving body, thereby improving the safety of the moving body.

[0032] Furthermore, the path generating unit may generate the path of the moving body based on a portion of the second cost map.

[0033] For example, when the moving body moves to avoid a moving object, the rider of the moving body is alert to the moving object and preferably moves farther away from the moving object than when avoiding a non-moving object. Therefore, when generating the path of the moving body, a part of the second cost map reflecting the motion prediction of the object is used, so that when the moving body avoids a moving object, the moving body moves farther away from the object than when avoiding a non-moving object.

[0034] In addition, it can be a part of the second cost map, which is determined according to the attributes of the object.

[0035] For example, sometimes the accuracy of motion prediction decreases depending on the properties of the object. For example, a person can suddenly change the direction of travel, so when the object is a person, the accuracy of the person's motion prediction tends to decrease. Therefore, for example, when the property of the object is a property with low accuracy of motion prediction, a part of the second cost map can be made smaller to reduce the influence of motion prediction on path generation.

[0036] For example, a portion of the second costmap may be determined according to the attributes of the interval included in the road on the path of the moving body. For example, the attributes of the interval may include the width of the road. For example, when the road is a road for moving bodies, the attributes of the interval may include the width of a sidewalk attached to the road. For example, the attributes of the interval may include an intersection on the path of the moving body. For example, the attributes of the interval may include the type of the road.

[0037] For example, when the width of a path of a moving body such as a vehicle (for example, the width of a driving lane) is narrow, it is sometimes difficult for the moving body to move farther away from a moving object when avoiding a moving object than when avoiding a non-moving object. Therefore, for example, when the width of the path of the moving body is narrow, a portion of the second cost map can be made smaller, so that the influence of motion prediction on the generation of the path is reduced, and when the moving body avoids the moving object, the moving body is prevented from greatly exceeding the driving lane. In addition, for example, at an intersection on the path of the moving body, the movement of objects around the moving body becomes complicated, so the accuracy of the object's motion prediction tends to decrease. Therefore, for example, when the moving body is at an intersection, a portion of the second cost map is made smaller, so that the influence of motion prediction on the generation of the path is reduced.

[0038] In addition, the cost map generation unit may further generate a third cost map based on the detected attributes of the object, and the judgment generation unit may further generate a movement judgment of the moving body based on the third cost map. For example, the judgment generation unit may generate a stop judgment of the moving body as the movement judgment based on the generated path and the third cost map.

[0039] The movement judgment is preferably determined according to the attributes of the external objects of the moving body. For example, depending on the attributes of the external objects of the moving body, it is sometimes preferred that the moving body stops. For example, there is a stationary person next to the crosswalk who wants to cross the crosswalk. In this case, in the movement judgment generated according to the second cost map, the person has not moved, so the moving body may not stop. Therefore, a third cost map is generated according to the attributes of the object (for example, whether the attribute of the object is a person), and a stop judgment of the moving body is performed according to the third cost map, so that the moving body can be stopped according to the attributes of the object, which can further improve the safety of the movement of the moving body.

[0040] In addition, for example, the control system further includes a lane prediction unit, which predicts the lane in which the moving body is scheduled to move, the cost map generation unit further generates a fourth cost map based on the predicted lane, and the judgment generation unit further generates a movement judgment of the moving body based on the fourth cost map. For example, the judgment generation unit can generate a deceleration judgment of the moving body as the movement judgment based on the generated path and the fourth cost map.

[0041] The movement judgment is preferably determined according to the planned lane of the moving body. For example, depending on the planned lane of the moving body, it is sometimes preferred that the moving body decelerates. For example, in the case where a moving body such as a vehicle wants to move to an adjacent lane, there is a rear vehicle behind the planned lane. In this case, in the movement judgment generated based on the second cost map, for example, even if there is a rear vehicle at a position behind the lane to a sufficient degree to move the lane, it is possible to stop the vehicle. Therefore, the fourth cost map is generated based on the predicted lane, and the deceleration judgment is performed based on the fourth cost map, so that the vehicle can be decelerated, the lane can be moved smoothly, and the driving safety or comfort of the vehicle can be further improved.

[0042] A control method involved in one scheme of the present disclosure is a control method executed by a computer, in which sensing data output by a sensor is obtained, and the sensor is a sensor that senses the outside of a moving body. Based on the sensing data, the position of an object outside the moving body is detected, and the movement of the detected object is predicted. A first cost map is generated based on the position of the detected object, and a second cost map is generated based on the predicted movement of the object. Based on the first cost map, a path of the moving body is generated, and based on the second cost map and the generated path, a movement judgment of the moving body is generated, and the movement of the moving body is controlled according to the generated path and the movement judgment.

[0043] Through the above, it is possible to provide a control method that can improve the safety or comfort of movement of a moving object using a cost map.

[0044] Hereinafter, embodiments and the like will be described with reference to the drawings.

[0045] In addition, the embodiments described below are general or specific examples. The values, shapes, materials, components, configuration positions of components, connection forms, steps, and the order of steps shown in the following embodiments are all examples, and their purpose is not to limit the present disclosure.

[0046] (Implementation Method 1)

[0047] use Figures 2 to 7 , implementation mode 1 is described.

[0048] Figure 2 This is a block diagram showing an example of the control system 10 according to the first embodiment.

[0049] The control system 10 is a computer for controlling the movement of a mobile body. The control system 10 is, for example, a vehicle-mounted device mounted on a mobile body (e.g., a vehicle 100, etc.). The control system 10 is a computer including a processor, a memory, and a communication interface, etc. The memory is a ROM (Read Only Memory) and a RAM (Random Access Memory), etc., which can store programs executed by the processor. The control system 10 includes an acquisition unit 11, a detection unit 12, a movement prediction unit 13, a cost map generation unit 14, a path generation unit 15, a judgment generation unit 16, and a control unit 17. The acquisition unit 11, the detection unit 12, the movement prediction unit 13, the cost map generation unit 14, the path generation unit 15, the judgment generation unit 16, and the control unit 17 can be implemented by a processor that executes a program stored in the memory, etc. Here, the mobile body is described as a vehicle, and the control system 10 is described as a computer for controlling the driving of the vehicle. In addition, in addition to the vehicle, the mobile body can also be a flying object such as a drone, a mobile robot, or a ship, etc.

[0050] In addition, the control system 10 may not be mounted on the vehicle, but may be a server connected to the vehicle in a wireless communication manner. In addition, the components constituting the control system 10 may be distributed and arranged on a plurality of servers.

[0051] The acquisition unit 11 acquires sensing data output by a sensor that senses the outside of the vehicle. For example, the acquisition unit 11 acquires sensing data output by a sensor mounted on a vehicle that is a control target of the control system 10, vehicles around the vehicle, or roadside equipment around the vehicle. The sensor is, for example, a camera or a LiDAR (Light Detection and Ranging) sensor.

[0052] The detection unit 12 detects the position of an object outside the vehicle based on the sensing data obtained by the acquisition unit 11. For example, the detection unit 12 analyzes the sensing data using a learning model or the like to detect the position of the object.

[0053] The movement prediction unit 13 predicts the movement of an object outside the vehicle based on the sensing data obtained from the acquisition unit 11. For example, the movement prediction unit 13 predicts the movement of the object (specifically, the moving direction and speed of the object) based on the change in the position of the object detected by the sensing data obtained every moment.

[0054] The cost map generation unit 14 generates a first cost map based on the position of the detected object and generates a second cost map based on the predicted movement of the object. Details of the cost map generation unit 14 will be described later. Figure 3 as well as Figure 4 In the description.

[0055] The path generation unit 15 generates a path for the vehicle based on the first cost map. The details of the path generation unit 15 will be described later. Figure 3 as well as Figure 5 In the description.

[0056] The judgment generating unit 16 generates a movement judgment of the vehicle based on the second cost map and the generated path. The details of the judgment generating unit 16 will be described later. Figure 3 as well as Figure 6 In the description.

[0057] The control unit 17 controls the movement of the vehicle according to the generated path and the movement determination. For example, the control unit 17 moves the vehicle along the generated path and stops or decelerates the vehicle according to the movement determination.

[0058] Figure 3 This is a flowchart showing an example of the operation of the control system 10 according to the first embodiment.

[0059] As described above, the acquisition unit 11 acquires sensing data output by a sensor for sensing the outside of the vehicle (step S11), the detection unit 12 detects the position of an object outside the vehicle based on the sensing data (step S12), and the movement prediction unit 13 predicts the movement of an object outside the vehicle based on the sensing data (step S13).

[0060] Next, the cost map generation unit 14 generates a first cost map based on the position of the detected object (step S14), and generates a second cost map based on the predicted movement of the object (step S15). The cost map is a grid map that represents objects such as obstacles as costs in order to generate an optimal path. For example, a high cost can be assigned to an area where an object exists, and a path that avoids the area can be generated. In a cost map, typically, each grid has a value within a range determined by an upper limit and a lower limit, but the cost map is not limited to this, and the value in each grid can also be a binary value. In addition, a cost map may also be referred to as a potential map. Here, using Figure 4 The first cost map and the second cost map generated by the cost map generation unit 14 will be described.

[0061] Figure 4 is a diagram showing an example of the first cost map and the second cost map. Figure 4 In the following figures, a vehicle 200 is shown as an example of an object outside the vehicle.

[0062] The first cost map is a cost map generated based on the position of the detected object (e.g., vehicle 200). Figure 4The first cost map is generated in such a manner that the closer to the center of the vehicle 200, the higher the cost.

[0063] The second costmap is a costmap generated based on the predicted movement of an object (e.g., vehicle 200). Figure 4 As shown in FIG. 1 , the second cost map is generated in a manner that the cost is high in the predicted moving direction of the vehicle 200 (e.g., the direction of the arrow mark extending from the vehicle 200). In addition, the higher the speed of the predicted vehicle 200, the further the area with high cost extends from the vehicle 200 in the moving direction.

[0064] The cost map generation unit 14 generates Figure 4 The first cost map and the second cost map shown. Figure 4 In the diagram, the cost map is schematically represented by a circle, and the cost inside the circle is represented as fixed. However, in fact, the cost in the cost map is not fixed, and the cost decreases from the center of the circle to the outside.

[0065] Back to Figure 3 As described above, the path generation unit 15 generates a path for the vehicle based on the first cost map (step S16). Figure 5 To illustrate the path generated based on the first cost map.

[0066] Figure 5 FIG. 1 is a diagram showing an example of a path generated based on the first cost map. Figure 5 In the drawings described later, a vehicle 100 is shown as an example of a vehicle controlled by the control system 10 .

[0067] In the first costmap, if Figure 5 As shown, with the position of vehicle 200 as the center, the cost is high, and the path generation unit 15 generates a path that avoids the high-cost area in the first cost map. When vehicle 100 moves along the path, sometimes there is a possibility that vehicle 100 and vehicle 200 will collide based on the movement of vehicle 200, but the movement prediction of vehicle 200 is not considered here, and the path of vehicle 100 is generated based on the first cost map generated based on the position of vehicle 200 at a certain moment.

[0068] Back to Figure 3 The judgment generating unit 16 generates a vehicle movement judgment based on the second costmap and the generated path (step S17). Figure 6 The movement judgment generated based on the second cost map and the generated path will be explained.

[0069] Figure 6This is a diagram showing an example of movement judgment generated based on the second costmap and the generated path.

[0070] like Figure 6 As shown, an area in the second cost map that is assigned a cost greater than a prescribed cost overlaps with the generated path. The prescribed cost is a cost that indicates that the probability of the existence of an object is high or will become high. In other words, when vehicle 100 moves along the generated path, there is a possibility that vehicle 100 and vehicle 200 will collide. Therefore, the movement judgment generated by the judgment generation unit 16 includes the following judgment, which is to stop or slow down vehicle 100 when an area in the second cost map that is assigned a cost greater than a prescribed cost overlaps with the generated path. Thus, it is possible to prevent vehicle 100 from colliding with vehicle 200, and to improve the safety of vehicle driving.

[0071] In addition, the judgment generating unit 16 may generate a movement judgment based on the first cost map in addition to the second cost map and the generated path. Alternatively, the cost map generating unit 14 may generate a second cost map based on the position of the detected object in addition to the predicted movement of the object, and the judgment generating unit 16 may generate a movement judgment based on the second cost map and the generated path. In other words, the movement judgment of the vehicle may also be generated by considering the position of the object.

[0072] Back to Figure 3 As described above, the control unit 17 controls the movement of the vehicle according to the generated path and movement judgment (step S18). For example, the control unit 17 calculates the tire angle or the pedal step amount according to the generated path and movement judgment, and notifies the vehicle of the calculation result via CAN (Controller Area Network) or the like, thereby controlling the movement of the vehicle.

[0073] In addition, the path generation unit 15 may also generate a path for the vehicle based on a portion of the second cost map. Figure 7 To explain.

[0074] Figure 7 This is a diagram showing an example of a path generated based on a part of the second costmap.

[0075] A portion of the second costmap may be, for example, Figure 6 The part of the high-cost area in the second cost map shown in the figure. In addition, the part of the second cost map can be a part in which the high-cost part in the second cost map is reduced, or it can be a part in which only the object side part is cut out of the high-cost part in the second cost map.

[0076] For example, when the vehicle 100 is traveling to avoid the moving vehicle 200, the passengers of the vehicle 100 are alert to the moving vehicle 200 and want to travel farther away from the vehicle 200 than when avoiding a non-moving object. Therefore, in the generation of the path of the vehicle 100, a part of the second cost map reflecting the motion prediction of the vehicle 200 is used, so that the vehicle 100 can travel farther away from the moving vehicle 200 than when avoiding a non-moving object. Figure 7 In the situation, the generated path avoids the high-cost area in a part of the second cost map in addition to the first cost map, so the vehicle 100 can travel as far to the right as possible from the vehicle 200.

[0077] Next, a method for determining a part of the second costmap will be described.

[0078] For example, a portion of the second cost map is determined according to the attributes of an object outside the vehicle. For example, the control system 10 may have the function of the attribute detection unit 18 described later, which detects the attributes of the object outside the vehicle, for example, by sensing data sensed by a sensor such as a camera. For example, according to the attributes of the object, the accuracy of the action prediction may become lower. For example, a person may suddenly change the direction of travel, so when the object is a person, the accuracy of the person's action prediction is likely to become lower. Therefore, for example, when the attribute of the object is an attribute with low accuracy of action prediction, making a portion of the second cost map smaller (in other words, the area of ​​the high cost area is smaller) can reduce the impact of the action prediction on the generation of the path. In this case, the vehicle's avoidance action to the object is suppressed. In addition, for example, when the attribute of the object is an attribute with high accuracy of action prediction, making a portion of the second cost map larger (in other words, the area of ​​the high cost area is larger) can increase the impact of the action prediction on the generation of the path. In this case, the vehicle's avoidance action to the object is recommended.

[0079] In addition, for example, a part of the second costmap is determined according to the attributes of the interval included in the road on the path of the vehicle. For example, the attributes of the interval may include the width of the road on the path of the vehicle, the width of the sidewalk attached to the road, the intersection on the path of the vehicle, or the type of road. An intersection is a place where two or more roads intersect, typically a crossroads, but not limited to this, it can also be a three-way intersection or a roundabout. In addition, the sidewalk can be a sidewalk parallel to the road, or it can be a crosswalk that crosses the road. The attributes of the interval can be determined, for example, by sensing data or map information of sensors such as cameras. In addition, the type of road can be a road of a specified type such as a lane or a sidewalk, or it can be a road similar to a specified type. Specifically, a road similar to a specified type can be determined based on the type of objects around the vehicle determined by the sensing data or the distance between the object and the vehicle. For example, when the density of people around the vehicle is above a threshold, it can be determined that the type of road is a sidewalk (or a suspected sidewalk). Suspected sidewalks refer to, for example, passages in shopping streets, lanes in pedestrian streets, etc.

[0080] For example, when the width of the road on the path of the vehicle (e.g., the width of the driving lane) is narrow, it is sometimes difficult for the vehicle to travel further away from the object when avoiding a moving object than when avoiding a non-moving object. Therefore, for example, when the width of the road on the path of the vehicle is narrow, a part of the second costmap is made smaller, thereby reducing the influence of the action prediction on the generation of the path. In this case, the avoidance action of the vehicle to the object can be suppressed, so that the vehicle does not greatly go out of the driving lane when avoiding the moving object.

[0081] In addition, for example, if the width of the sidewalk attached to the road on the path of the vehicle is wide, the vehicle may cross the sidewalk or get very close to the sidewalk when avoiding the moving object. Therefore, for example, if the width of the sidewalk on the path of the vehicle is wide, a part of the second costmap is made smaller, thereby reducing the influence of the action prediction on the generation of the path. In this case, the avoidance action of the vehicle to the object is suppressed, and the vehicle can be prevented from crossing the sidewalk or getting very close to the sidewalk when avoiding the moving object.

[0082] In addition, for example, at an intersection on the path of the vehicle, the movement of objects around the vehicle becomes complicated, and the accuracy of the object movement prediction is likely to decrease. Therefore, for example, when the vehicle is at an intersection, making a part of the second cost map smaller can reduce the influence of the movement prediction on the generation of the path. In this case, the vehicle's avoidance action of the object is suppressed.

[0083] In addition, for example, when the type of road on the path of the vehicle is a sidewalk, the density of objects (such as pedestrians) around the vehicle is high (in other words, the vehicle is close to the objects), so the vehicle may stop repeatedly. On the other hand, the moving speed of the vehicle is limited on the sidewalk, and the risk of collision is lower than that of the lane. Therefore, for example, when the type of road is a sidewalk, compared with the case where the type of road is a lane, a part of the second cost map is made larger, so that the influence of action prediction on the generation of the path becomes larger. Thereby, it is possible to suppress the stopping of the vehicle and enable the vehicle to avoid the object and continue to move.

[0084] In addition, a portion of the second costmap may be determined based on both the attributes of an object outside the vehicle and the attributes of a section included in a road on the path of the vehicle.

[0085] As described above, the control system 10 generates a first cost map based on the position of the detected object and a second cost map based on the predicted movement of the object, respectively, and each cost map is used separately in the generation of the vehicle path and the generation of the vehicle movement judgment. Specifically, in the generation of the vehicle path, the first cost map that does not reflect the motion prediction of the object is used, and in the generation of the vehicle movement judgment, the second cost map that reflects the motion prediction of the object is used. In the generation of the vehicle path, the first cost map that does not reflect the motion prediction is used, so that even if the motion prediction is unstable, it is difficult to generate a path that changes at all times, which can improve the safety or comfort of the vehicle's driving. In addition, in the generation of the vehicle's movement judgment, the second cost map that reflects the motion prediction of the object and the generated path are used, so that both the stationary object is avoided and the moving object is not forcibly avoided but is based on the motion prediction of the object, for example, the vehicle can be stopped or decelerated, which can improve the safety of the vehicle's driving. In this way, on the premise of unstable motion prediction, a stable path can be generated, and it can be stopped or decelerated when necessary. Therefore, according to this solution, the driving safety or comfort of a vehicle using a cost map can be improved.

[0086] (Implementation Method 2)

[0087] Next, use Figure 8 Implementation method 2 will be described.

[0088] Figure 81 is a block diagram showing an example of a control system 10a according to Embodiment 2. The control system 10a according to Embodiment 2 is different from the control system 10 according to Embodiment 1 in that it further includes an attribute detection unit 18 and a lane prediction unit 19, and includes a cost map generation unit 14a and a judgment generation unit 16a instead of the cost map generation unit 14 and the judgment generation unit 16. The other parts are the same as those in Embodiment 1, so the description is omitted. In addition, the cost map generation unit 14a, the judgment generation unit 16a, the attribute detection unit 18, and the lane prediction unit 19 are implemented by a processor or the like that executes a program stored in a memory.

[0089] The attribute detection unit 18 detects the attributes of an object outside the vehicle based on the sensing data of the sensor such as the camera obtained by the acquisition unit 11. For example, the attribute detection unit 18 analyzes the sensing data using a learning model, etc., to detect the attributes of the object. For example, the attribute detection unit 18 detects whether the object is a person, a roadside device, a car, or a bicycle.

[0090] The lane prediction unit 19 predicts the lane that the vehicle is scheduled to travel. For example, the lane prediction unit 19 obtains information indicating the vehicle's scheduled travel from a car navigation system, and can predict the lane that the vehicle is scheduled to travel based on the information. In addition, for example, the lane prediction unit 19 obtains operation information for a direction indicator, and can predict the lane that the vehicle is scheduled to travel based on the information.

[0091] The cost map generation unit 14a generates a third cost map based on the attributes of the detected object. For example, when the object is an immobile object such as a roadside device, the third cost map is generated in such a way that the cost around the area corresponding to the object becomes smaller. In addition, for example, when the object is a person, the third cost map is generated in such a way that the cost around the area corresponding to the object becomes larger.

[0092] Furthermore, the costmap generation unit 14a generates a fourth costmap based on the predicted lane. For example, the fourth costmap is generated so that the cost of the area corresponding to the lane in which the vehicle is expected to travel becomes larger.

[0093] The judgment generating unit 16a also generates a movement judgment of the vehicle based on the third cost map. For example, the judgment generating unit 16a generates a stop judgment of the vehicle as a movement judgment based on the generated path and the third cost map. For example, the movement judgment includes the following judgment: when an area that is assigned a cost greater than a specified cost in the third cost map overlaps with the generated path, the judgment stops the vehicle. For example, depending on the attributes of an object outside the vehicle, it is sometimes preferred that the vehicle stops. For example, there is a situation where someone stops on the side of a crosswalk and the person wants to cross the crosswalk. In this case, in the movement judgment generated based on the second cost map, the person does not move, so it is possible that the vehicle will not stop. Therefore, a third cost map is generated based on the attributes of the object (for example, whether the attribute of the object is a person), and a stop judgment of the vehicle is performed based on the third cost map, so that the vehicle can be stopped according to the attributes of the object, which can further improve the safety of vehicle driving.

[0094] In addition, the judgment generating unit 16a also generates a movement judgment of the vehicle based on the fourth cost map. For example, the judgment generating unit 16a generates a deceleration judgment of the vehicle as a movement judgment based on the generated path and the fourth cost map. For example, the movement judgment includes the following judgment: when an area with a cost greater than a prescribed cost in the fourth cost map overlaps with the generated path, a judgment is made to decelerate the vehicle. For example, it is sometimes preferable to decelerate the vehicle according to the lane in which the vehicle is scheduled to travel. For example, when the vehicle wants to move to the lane next to it, there is a rear vehicle behind the lane in which it is scheduled to travel. In this case, in the movement judgment generated based on the second cost map, for example, even if there is a rear vehicle at a position behind the lane to a degree sufficient to move the lane, it is possible to stop the vehicle. Therefore, the fourth cost map is generated based on the predicted lane, and the deceleration judgment is made based on the fourth cost map, so that the vehicle is decelerated, the lane can be moved smoothly, and the driving safety or comfort of the vehicle can be further improved.

[0095] As described above, in addition to the first cost map based on the position of the object and the second cost map based on the movement of the object, a third cost map based on the attributes of the object and a fourth cost map based on the planned lane of travel are also generated. Based on the first cost map, the second cost map, the third cost map and the fourth cost map, a vehicle movement judgment is generated, which can further improve the vehicle's driving safety or comfort.

[0096] (Other embodiments)

[0097] The above describes the control system involved in one or more forms of the present disclosure based on the implementation methods, but the present disclosure is not limited to these implementation methods. Without departing from the scope of the present disclosure, forms obtained by performing various modifications that can be thought of by people in the field on each implementation method and forms formed by combining constituent elements in different implementation methods are also included in the scope of one or more forms of the present disclosure.

[0098] For example, in the above-mentioned embodiment 2, an example is described in which the cost map generation unit 14a generates both the third cost map and the fourth cost map, but the fourth cost map may not be generated. In this case, the control system 10a may not have the lane prediction unit 19, and the judgment generation unit 16a may not further generate the vehicle movement judgment based on the fourth cost map. Alternatively, the cost map generation unit 14a may not generate the third cost map. In this case, the control system 10a may not have the attribute detection unit 18, and the judgment generation unit 16a may not further generate the vehicle movement judgment based on the third cost map.

[0099] In addition, the present disclosure can be implemented not only as a control system but also as a control method including steps (processing) performed by each component constituting the control system.

[0100] Specifically, the control method is a control method that enables a computer to execute, such as Figure 3 The process shown includes: obtaining sensing data output by a sensor for sensing the outside of a moving body (step S11), detecting the position of an object outside the moving body based on the sensing data (step S12), predicting the movement of the detected object (step S13), generating a first cost map based on the position of the detected object (step S14), generating a second cost map based on the predicted movement of the object (step S15), generating a path for the moving body based on the first cost map (step S16), generating a movement judgment of the moving body based on the second cost map and the generated path (step S17), and controlling the movement of the moving body according to the generated path and the movement judgment (step S18). In addition, in Figure 3 In the example, the moving object is set to be a vehicle.

[0101] For example, the present disclosure can be implemented as a program for causing a processor to execute the steps included in the control method. Furthermore, the present disclosure can be implemented as a non-transitory computer-readable recording medium such as a CD-ROM recording the program.

[0102] For example, when the present disclosure is implemented by a program (software), the program is executed by using the hardware resources such as the CPU, memory, and input / output circuit of the computer, thereby executing each step. In other words, the CPU obtains data from the memory or input / output circuit and performs calculations, or outputs the calculation results to the memory or input / output circuit, thereby executing each step.

[0103] In addition, in each of the above-mentioned embodiments, each component included in the control system may be formed by dedicated hardware, or implemented by executing a software program suitable for each component. Each component may be implemented by a program execution unit such as a CPU or a processor reading and executing a software program recorded in a recording medium such as a hard disk or a semiconductor memory.

[0104] Part or all of the functions of the control system involved in the embodiment are typically implemented as an integrated circuit, i.e., LSI. These can be monolithic, or monolithic in a manner that includes part or all. In addition, integrated circuitization is not limited to LSI, and can be implemented using a dedicated circuit or a general-purpose processor. It is also possible to use a programmable FPGA (Field Programmable Gate Array) after LSI manufacturing, or a reconfigurable processor that can reconfigure the connection and settings of the circuit unit inside the LSI.

[0105] Furthermore, various modifications of the embodiments of the present disclosure that can be conceived by those skilled in the art are also included in the present disclosure without departing from the gist of the present disclosure.

[0106] The present disclosure can be applied to a system for controlling movement of a mobile object such as a vehicle.

[0107] Explanation of symbols

[0108] 10, 10a Control system

[0109] 11. Acquisition Department

[0110] 12. Inspection Department

[0111] 13 Motion Prediction Department

[0112] 14, 14a Costmap Generation

[0113] 15 Path Generation Unit

[0114] 16, 16a Judgment generation unit

[0115] 17 Control Unit

[0116] 18 Attribute Detection Unit

[0117] 19 Lane Prediction Department

[0118] 100, 200 vehicles

Claims

1. A control system comprising: an acquisition unit that acquires sensing data output by a sensor that senses the outside of the moving body; a detection unit, which detects a position of an object outside the moving body based on the sensing data; a movement prediction unit, configured to predict the movement of the object based on the sensing data; A cost map generating unit generates a first cost map according to the detected position of the object, and generates a second cost map according to the predicted movement of the object; a path generating unit for generating a path for the moving body based on the first cost map that does not reflect the motion prediction of the object; a judgment generating unit that generates a movement judgment of the moving body based on the second cost map reflecting the movement prediction of the object and the generated path; as well as The control unit controls the movement of the moving body according to the generated path and the movement judgment.

2. The control system according to claim 1, When an area to which a cost equal to or greater than a predetermined cost is assigned in the second cost map overlaps with the generated path, the movement determination includes a determination of stopping the moving body or decelerating the moving body.

3. The control system according to claim 1, The path generation unit further generates the path of the moving object based on a portion of the second cost map.

4. The control system according to claim 3, A portion of the second cost map is determined according to the attributes of the object.

5. The control system according to claim 3, A portion of the second costmap is determined according to an attribute of a section included in a road on a path of the moving object.

6. The control system according to claim 5, The attributes of the section include the width of the road.

7. The control system according to claim 5, When the road is a road for moving objects, the attribute of the section includes a width of a sidewalk attached to the road.

8. The control system according to claim 5, The attributes of the section include intersections on the path of the moving body.

9. The control system according to claim 5, The attributes of the section include the type of the road.

10. The control system according to claim 1, The cost map generation unit further generates a third cost map based on the attributes of the detected object. The judgment generating unit further generates a movement judgment of the moving object based on the third cost map.

11. The control system according to claim 10, The judgment generating unit generates a stop judgment of the moving object as the movement judgment based on the generated path and the third cost map.

12. A control system according to any one of claims 1 to 11, The control system further includes a lane prediction unit that predicts a lane in which the moving object is scheduled to move. The costmap generating unit further generates a fourth costmap based on the predicted lane. The judgment generating unit further generates a movement judgment of the moving object based on the fourth cost map.

13. The control system according to claim 12, The judgment generating unit generates a deceleration judgment of the moving object as the movement judgment based on the generated path and the fourth cost map.

14. A control method is a control method executed by a computer, in which: obtaining sensing data output by a sensor that senses the exterior of the moving body, detecting a position of an object outside the moving body according to the sensing data, predicting the movement of the detected object, According to the detected position of the object, a first cost map is generated, generating a second cost map based on the predicted movement of the object, generating a path of the moving object based on the first cost map that does not reflect the motion prediction of the object, generating a movement judgment of the moving body based on the second cost map reflecting the movement prediction of the object and the generated path, and The movement of the moving object is controlled according to the generated path and the movement judgment.

Images