Route generating apparatus and method

Abstract

The object of this invention is to generate a movement path for a mobile body not only by relying on information obtained from sensors, but also by using information indicating the probability of collision with obstacles, so that the mobile body can avoid obstacles naturally in more situations. The path generation apparatus of this invention includes: a movement state estimation unit that uses information obtained from sensors to estimate the movement state of a mobile body; a collision risk judgment unit that outputs information indicating the judgment result of the mobile body's attempt to avoid obstacles, i.e., judgment information, based on the movement state and information indicating the probability of collision between the mobile body and an obstacle, i.e., collision risk information; and a path generation unit that generates a movement path for the mobile body based on the judgment information.

Description

Technical Field

[0001] This disclosure relates to a path generation apparatus and method. Background Technology

[0002] For the purpose of autonomous movement of mobile bodies such as cars, ships, and airplanes, a path generation device is proposed to generate the movement path of the mobile body. The autonomously moving body uses sensors to detect obstacles in its surroundings. Moreover, when there are obstacles on the pre-planned movement path that may collide with the mobile body, the path generation device changes the movement path, thereby allowing the mobile body to avoid the obstacles.

[0003] For example, Patent Document 1 discloses a method that uses fine and natural-feeling braking based on the reliability of obstacle detection results performed by sensors, i.e., detection reliability (the length of time it takes to detect the same obstacle), so that the moving body avoids the obstacle with a natural feeling.

[0004] Patent Document 1: Japanese Patent Application Publication No. 2012-183868

[0005] However, the prior art described in Patent Document 1 has the following problems.

[0006] For example, in environments where sensor sensitivity is reduced, such as at night or in fog, the detection distance of sensors for obstacles decreases. Because the detection distance is shorter, the accuracy of detection reliability decreases. Therefore, even with braking based on detection reliability, there are limitations to a moving body's ability to avoid obstacles naturally. Summary of the Invention

[0007] This disclosure is made to solve the aforementioned problems, with the aim of generating a movement path for a moving body not only by relying on information obtained from sensors, but also by using information representing the probability of collision with obstacles, so that the moving body can avoid obstacles with a natural sense in more situations.

[0008] The path generation apparatus disclosed herein includes:

[0009] The movement state estimation unit uses information obtained from sensors to estimate the movement state of the moving body;

[0010] The collision risk assessment unit, based on the movement state and information indicating the probability of the moving body colliding with an obstacle (i.e., collision risk information), outputs information indicating the assessment result of the moving body's attempt to avoid the obstacle (i.e., assessment information); and

[0011] The path generation unit generates the movement path of the moving body based on the judgment information.

[0012] The path generation method disclosed herein,

[0013] The motion state estimation unit uses information obtained from sensors to estimate the motion state of the moving body.

[0014] Based on the movement state and information indicating the probability of the moving body colliding with an obstacle (i.e., collision risk information), the collision risk assessment unit outputs information indicating the assessment result of the moving body's attempt to avoid the obstacle (i.e., assessment information).

[0015] The path generation unit generates the movement path of the moving body based on the judgment information.

[0016] According to this disclosure, in most situations, a moving body can avoid obstacles with a natural sense of movement. Attached Figure Description

[0017] Figure 1 This is a functional block diagram showing the structure of the path generation device in Implementation Method 1.

[0018] Figure 2 This is a diagram showing the hardware structure of the path generation device in Embodiment 1.

[0019] Figure 3 This diagram is used to illustrate the specific operation of existing path generation devices.

[0020] Figure 4 This diagram is used to illustrate the specific operation of existing path generation devices.

[0021] Figure 5 This diagram illustrates the specific actions involved in path generation and placement in Implementation Method 1.

[0022] Figure 6 This diagram illustrates the specific actions involved in path generation and placement in Implementation Method 1.

[0023] Figure 7 This is a flowchart illustrating the operation of the path generation device in Embodiment 1. Detailed Implementation

[0024] The following is a reference to the appendix. Figure 1 The embodiments will be described below. The following embodiments are merely examples, and various modifications can be made within the scope of this invention. In the description of the embodiments and the accompanying drawings, the same elements and corresponding elements are labeled with the same reference numerals. Descriptions of elements labeled with the same reference numerals are appropriately omitted or simplified. In the following embodiments, the term "part" may also be appropriately replaced with "circuit," "process," "sequence," or "processing."

[0025] Implementation method 1.

[0026] Structural description

[0027] Figure 1 This is a functional block diagram illustrating the structure of the path generation apparatus of this embodiment. The path generation apparatus 10 is an apparatus capable of implementing the path generation method of this embodiment. Figure 1 In this system, the mobile body 100 includes a path generation device 10, various sensors 11, and a control unit 21. The path generation device 10 includes a movement state estimation unit 12, a collision risk assessment unit 13, a path generation unit 16, a map information storage unit 17, an event information storage unit 18, a risk map information storage unit 19, and a movement path information storage unit 20. The collision risk assessment unit 13 includes an estimation unit 14 and a assessment unit 15.

[0028] The mobile body 100 is a mobile body capable of autonomous movement. For example, the mobile body 100 may be an autonomous vehicle, a personal mobility vehicle, an autonomous robot, a ship, a railway, an airplane, or a drone. Furthermore, the mobile body 100 possesses various sensors 11 required for autonomous movement. These sensors 11 may include devices such as GPS (Global Positioning System), IMU (Inertial Measurement Unit), LiDAR (Light Detection and Range), millimeter-wave radar, ultrasonic sonar, cameras, and beacons. Alternatively, the sensors 11 may not possess all of the aforementioned devices. The selection or omission of the various sensors 11 can be appropriately chosen based on the type of mobile body 100, the mobile environment, and the cost of the equipment.

[0029] Furthermore, the various sensors 11 do not necessarily need to be present on the mobile body 100. For example, sensors located outside the mobile body 100 (e.g., surveillance cameras, beacons, roadside devices, etc.) can also obtain the information required for the autonomous movement of the mobile body 100. In other words, as long as the information required for the autonomous movement of the mobile body 100 can be provided to the mobile body 100 from the outside, the various sensors 11 of the mobile body 100 can be omitted.

[0030] The mobile body 100 includes a power source and power unit required for movement, and a driven device driven by the power unit. For example, the power source may be battery power, fuel, etc. For example, the power unit may be a motor, engine, etc. For example, the driven device may be wheels, propellers, screws, etc. The driven device may also include a steering control device for the mobile body 100 to perform direction changes and braking. Additionally, in Figure 1 The diagrams of the power source, power unit, driven device, and steering control device are omitted.

[0031] Various sensors 11 acquire motion information D1, which represents the state of the moving body 100 itself. For example, motion information D1 includes information such as the moving speed, moving direction, and posture of the moving body 100.

[0032] In addition, various sensors 11 acquire obstacle information D2, which represents information about obstacles existing around the mobile body 100. For example, obstacle information D2 includes information such as the type, location, size, shape, color, and whether the obstacle is moving. An obstacle is an object or state that hinders the movement of the mobile body 100. For example, if the mobile body 100 is an autonomous vehicle, obstacles include fallen objects, holes, puddles, other mobile bodies, people, animals, etc. For example, if the mobile body 100 is a ship, obstacles include drifting objects, reefs, animals, other mobile bodies, etc. For example, if the mobile body 100 is an aircraft, obstacles include turbulent air currents, thunderclouds, animals, other mobile bodies, etc. In addition, obstacle information D2 can also be acquired by various sensors on external infrastructure devices or other mobile bodies, such as roadside devices, surveillance cameras, beacons, radar, etc. The obstacle information D2 acquired by external infrastructure devices or other mobile bodies can also be provided to the path generation device 10 of the mobile body 100 via a wireless communication device (not shown).

[0033] The map information storage unit 17 is a storage unit that stores map information D3 representing the environment of the moving body 100's movement path. For example, the map information D3 can include information such as the state of the ground or space, including roads (paved, unpaved, road slope, etc.), tracks, waterways, intersections, grade-separated intersections, number of lanes and lane width, road markings, signs, and traffic lights. In addition, the map information D3 may also include information such as the type, location, width, and size of static features that may obstruct the movement of the moving body 100.

[0034] The movement state estimation unit 12 uses the movement body information D1 and map information D3 to estimate the movement state of the movement body 100. The estimated movement state is output as movement state information D4. For example, the movement state is the movement body 100's current position, speed, acceleration, direction of movement, posture, etc. on the map.

[0035] The event information storage unit 18 is a storage unit that stores event information D5, which represents information about a near-miss incident (a narrow escape) that occurred in the past, such as a collision. For example, event information D5 can be set to the time of the incident, the coordinates of the location, surrounding information, weather, the type of moving object involved, and information about obstacles. Furthermore, event information D5 can be predetermined before the moving object 100 moves, or it can be changed in real time during the movement of the moving object 100 through information update operations from outside the moving object 100.

[0036] The risk map information storage unit 19 is a storage unit that stores risk map information D6, which represents information about the potential hazards around the moving body 100. More specifically, the risk map information D6 can be set as information that records the frequency of the risk of collision between the moving body and obstacles and the degree of impact of such risks on a map, for example, based on the blind spots of the moving body 100, the location of the terrain features around the moving body 100, and the shape of the terrain features. In addition, the risk map information D6 can be predetermined before the moving body 100 moves, or it can be changed in real time during the movement of the moving body 100 through information update operations from outside the moving body 100.

[0037] The movement path information storage unit 20 is a storage unit that stores movement path information D7, which represents information about the movement path, including the movement speed and direction of the moving body 100. Furthermore, the movement path information D7 can be predetermined before the moving body 100 moves, or it can be appropriately changed during the movement of the moving body 100 through information update operations from outside the moving body 100.

[0038] The collision risk assessment unit 13 uses obstacle information D2, movement status emotion D4, event information D5, risk map information D6, and movement path information D7 to estimate the collision risk between the moving body 100 and the obstacle, and determines whether the moving body 100 needs to avoid the obstacle, and outputs information indicating the judgment result of the moving body 100 in avoiding the obstacle, namely judgment information D9.

[0039] Next, the internal structure of the collision risk assessment unit 13 will be explained.

[0040] The estimation unit 14 uses obstacle information D2, event information D5, risk map information D6, and movement path information D7 to estimate collision risk information D8, which represents the probability that the moving body 100 may collide with an obstacle while moving according to the movement path information D7. More specifically, for example, the collision risk information D8 can be set as a numerical value representing the probability of the moving body 100 colliding with an obstacle at any position on the movement path shown by the movement path information D7. For example, the collision risk information D8 can also be represented by a percentage value from 0% to 100%. Alternatively, the collision risk information D8 can also be represented by a standardized value from 0 to 1. For example, when the collision risk information D8 is represented by a percentage value, the larger the value, the higher the risk of the moving body 100 colliding with the obstacle.

[0041] The estimation unit 14 can, for example, use a learned model to estimate collision risk information D8. If obstacle information D2, event information D5, risk map information D6, and movement path information D7 are input to the learned model, the learned model outputs a response to these inputs, i.e., the collision risk information D8. Furthermore, the learned model can be created using machine learning methods such as deep learning or support vector machines. Additionally, the estimation method for collision risk information D8 can be a rule-based method, such as referring to a pre-determined risk map table.

[0042] Furthermore, either event information D5 or risk map information D6 indicates the probability of collision with obstacles (e.g., other moving objects, fallen objects, etc.) on the movement path. Therefore, the estimation unit 14 may use only event information D5 to estimate collision risk information D8. Alternatively, the estimation unit 14 may use only risk map information D6 to estimate collision risk information D8.

[0043] The determination unit 15 uses the movement state information D4 and the collision risk information D8 to determine whether the moving body 100 needs to avoid an obstacle, and outputs information indicating the determination result of the moving body 100's attempt to avoid the obstacle, namely, determination information D9. More specifically, for example, the determination unit 15 can compare the collision risk information D8 at the position shown in the movement state information D4 with a predetermined threshold, and if the probability of a collision is higher than the threshold, output determination information D9 to cause the moving body 100 to begin an avoidance action.

[0044] In order to perform safety confirmation when changing the travel route, the judgment unit 15 can use the movement status information D4 and collision risk information D8 to determine the congestion situation around the moving body 100. For example, the congestion situation is the number of moving bodies per unit distance around the movement path.

[0045] The determination unit 15 can determine not only whether the moving body 100 avoids an obstacle, but also whether other actions are performed. For example, the determination unit 15 can also determine whether the moving body 100 is preparing to avoid an obstacle, that is, making a preparatory action. A preparatory action means applying light braking, stopping acceleration, or other actions that cause the moving body 100 to decelerate slowly.

[0046] The determination unit 15 can also determine whether the moving body 100 performs an avoidance action, such as changing the travel path of the moving body 100 to avoid an obstacle, performing a strong deceleration action of the moving body 100 to prevent a collision with an obstacle, or stopping. Avoidance actions include actions such as applying brakes to forcefully decelerate the moving body 100 to stop, or actions that change the travel path of the moving body 100.

[0047] Hereinafter, the structure and operation of the path generation device 10 of this embodiment will be described in more detail with reference to the mobile body 100 as an autonomous vehicle.

[0048] First, the acceleration generated when the moving body 100 decelerates or changes its course to avoid obstacles is explained, and the movement path that the moving body 100 can avoid with a natural feeling is specifically defined.

[0049] First, for example, let's explain the acceleration (deceleration) generated in the moving body 100 due to deceleration based on braking operation. For example, in the case of gentle braking, the deceleration is approximately 0.1G. In the case of normal braking, the deceleration is approximately 0.2G. In the case of forceful braking, the deceleration is approximately 0.3G. Here, G is the unit representing gravitational acceleration. Generally, if the deceleration exceeds 0.3G, a person will feel uncomfortable. Therefore, considering a margin, the deceleration is preferably no more than approximately 0.25G.

[0050] Furthermore, when the moving body 100 changes its course to avoid an obstacle, a lateral acceleration is generated within the moving body 100. The strength of this lateral acceleration is based on the value of the deceleration acceleration described above. Hereinafter, both deceleration acceleration and lateral acceleration will be collectively referred to as obstacle avoidance acceleration.

[0051] Based on the above, a movement path in which the mobile body 100 can avoid obstacles naturally refers to a movement path consisting of deceleration or a change in travel route with an avoidance acceleration not exceeding approximately 0.25G, or a movement path consisting of both deceleration and a change in travel route. Furthermore, the value of 0.25G for avoidance acceleration is an example; for instance, the value of avoidance acceleration can be appropriately changed depending on the type of mobile body 100 and the attributes of the occupants (driver, passenger, etc.).

[0052] In the judgment unit 15, as a method for using collision risk information D8 to determine the likelihood of collision with an obstacle and to decide the action of the moving body 100, a rule-based method can be used, for example. Hereinafter, the rule-based judgment method will be explained using an example where the collision risk information D8 is represented by a percentage value from 0% to 100%.

[0053] First, the probability of a collision is determined, and thresholds are defined to determine the action of the moving body 100. Predetermined thresholds are set as TH1 and TH2; for example, the values ​​of thresholds TH1 = 20% and TH2 = 80%, respectively. Furthermore, the values ​​of thresholds TH1 and TH2 are just examples; their values ​​can be appropriately changed based on factors such as the distance of the moving body 100 to the obstacle, the probability of the obstacle's presence, the movement state of the moving body 100, and the condition of the movement path.

[0054] If the collision risk information D8 is above the threshold TH2, the judgment unit 15 judges that the possibility of colliding with the obstacle is high and outputs judgment information D9 so that the moving body 100 performs a formal action (braking or changing the route).

[0055] If the collision risk information D8 is less than the threshold TH2 and greater than the threshold TH1, the judgment unit 15 judges that the possibility of colliding with the obstacle is moderate and outputs judgment information D9 so that the moving body 100 can perform a preparatory action.

[0056] If the collision risk information D8 is less than the threshold TH1, the judgment unit 15 judges that the possibility of colliding with the obstacle is low and outputs judgment information D9 so that the moving body 100 maintains its current travel route and speed.

[0057] Furthermore, the threshold values ​​used to determine the likelihood of a collision are not limited to two. For example, when there are multiple types of preparatory actions (e.g., dividing actions into gentle braking and stopping acceleration), the threshold values ​​can be set to three or more stages.

[0058] Furthermore, the probability of a collision can also be set as a continuous value using a function y = f(x) that takes the collision risk information D8 as input x and the probability of collision as output y. For example, the function f(x) can be obtained using data from past accidents or near-accidents through statistical methods such as regression analysis.

[0059] Alternatively, the decision unit 15 can also use the fully learned model to determine the probability of a collision. Specifically, if the collision risk information D8 is input into the fully learned model, the fully learned model predicts the probability of a collision and determines the decision information D9. For example, the fully learned model can be created using machine learning such as deep learning or support vector machines.

[0060] The path generation unit 16 uses the movement state information D4, the movement path information D7, and the judgment information D9 to generate a movement path that allows the moving body 100 to avoid obstacles naturally. The generated (updated) movement path is input again into the movement path information storage unit 20 as movement path information D7 and output as control information D10.

[0061] The control unit 21 uses control information D10 to control the power source, power unit, driven device, and steering control device of the moving body 100. Moreover, the moving body 100 moves along the generated movement path, thereby being able to avoid obstacles with a natural feel.

[0062] As described above, regarding the estimation of the collision risk between the moving body 100 and the obstacle, and the method by which the moving body 100 avoids the obstacle, the collision risk judgment unit 13 not only relies on information obtained from various sensors (i.e., obstacle information D2), but also uses information indicating the probability of collision with the obstacle and other moving bodies (i.e., event information D5, risk map information D6) to make a comprehensive judgment. Therefore, the moving body 100 can avoid the obstacle with a natural feeling.

[0063] Hardware structure description

[0064] Figure 2 This diagram illustrates the hardware structure of the path generation apparatus 10 according to this embodiment. The path generation apparatus 10 is, for example, a computer, a dedicated computing device, or a device combining a computer and a dedicated computing device. The path generation apparatus 10 includes: a processor 31 as an information processing unit, a memory 32 as a storage unit, a storage device 33 as a non-volatile storage unit, an interface 34, and a communication unit 35.

[0065] Processor 31 is connected to other hardware via a system bus and controls the aforementioned other hardware. Processor 31 is an IC (Integrated Circuit) that performs processing. Specific examples of processor 31 include CPU (Central Processing Unit), DSP (Digital Signal Processor), GPU (Graphics Processing Unit), or FPGA (Field Programmable Gate Array).

[0066] The processor 31 executes the program stored in the memory 32. The program includes the path generation program of this embodiment. The function of the path generation device 10 is implemented by the processor 31 that executes the program. The processor 31 is an example of a processing circuitry.

[0067] For example, the movement state estimation unit 12, the collision risk judgment unit 13, and the path generation unit 14 are implemented by the processor 31. For example, the map information storage unit 17, the event information storage unit 18, the risk map information storage unit 19, and the movement path information storage unit 20 are implemented by the memory 32.

[0068] The CPU performs program execution and data processing. The DSP performs arithmetic operations and digital signal processing such as data manipulation. For example, the processing of sensor data obtained from millimeter-wave radar is preferably handled at high speed by a DSP instead of CPU-based processing.

[0069] A GPU is a processor dedicated to image processing. GPUs process multiple pixels of data in parallel, enabling high-speed image processing. GPUs can handle template matching, a frequently used process in image processing, at high speed. For example, sensing sensor data from a camera is preferably processed by the GPU. If the CPU were to process the sensor data from the camera, the processing time would be enormous. Furthermore, GPUs are not only used for simple image processing; their computing resources can also be used for general-purpose computing (GPGPU). For example, using GPGPUs and deep learning for image processing can enable more accurate detection of obstacles and other moving objects.

[0070] An FPGA is a programmable processor with a logic circuit structure. FPGAs possess both dedicated hardware computing circuitry and programmable software. Complex calculations and parallel processing can be executed at high speed by FPGAs.

[0071] The memory 32 is, for example, a volatile memory. The volatile memory is capable of moving data at high speed when the path generation device 10 is in operation. Specific examples of volatile memory include RAM (Random Access Memory), SDRAM (Synchronous Dynamic Random Access Memory), etc.

[0072] Storage device 33 is a non-volatile memory that can continuously maintain the execution program and data even when the power to path generation device 10 is turned off. Specific examples of non-volatile memory include EEPROM (Electrically Erasable Programmable Read Only Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, etc. For example, non-volatile memory can also be portable storage media such as SD (Secure Digital) memory cards, CF (Compact Flash) memory, NAND flash memory, floppy disks, optical discs, compact discs, Blu-ray discs, and DVDs (Digital Versatile Discs).

[0073] Alternatively, the path generation program in this embodiment can also be provided by a portable storage medium. The memory 32 is connected to the processor 31 via a memory interface (not shown). The memory interface uniformly manages memory access from the processor and performs efficient memory access control. The memory interface is used for data transmission in the path generation apparatus 10.

[0074] Figure 1 The functions of the movement state estimation unit 12, collision risk assessment unit 13, and path generation unit 16 shown are implemented by the processor 31, which executes the program. The memory 32 stores programs that implement the functions of the movement state estimation unit 12, collision risk assessment unit 13, and path generation unit 16. The processor 31 reads these programs from the memory 32 and executes them. Additionally, the memory 32 is also used to temporarily store intermediate information for each program.

[0075] Furthermore, the functions of the movement state estimation unit 12, the collision risk judgment unit 13, and the path generation unit 16 can also be implemented by logic circuits as hardware. In this case, logic circuit information is stored in memory 32. The logic circuit information is read and executed by processor 31.

[0076] The processor 31 may also be composed of multiple processors. In this case, the multiple processors may also cooperate to execute programs that implement the functions of the movement state estimation unit 12, the collision risk judgment unit 13, and the path generation unit 16.

[0077] The memory 32 stores map information D3, event information D5, risk map information D6, and movement path information D7. Alternatively, the map information D3, event information D5, risk map information D6, and movement path information D7 can also be stored in an external storage device of the path generation device 10. In other words, the map information storage unit 17, event information storage unit 18, risk map information storage unit 19, and movement path information storage unit 20 can also be external storage devices of the path generation device 10.

[0078] The operation and effects of the path generation device 10 of this embodiment will be explained by comparing it with the operation of a conventional path generation device. Figure 3 and Figure 4 This diagram illustrates the specific operation of an existing path generation device. Figure 5 and Figure 6 This diagram illustrates the specific operation of the path generation device 10 in this embodiment. Figures 3-6 In the various figures, for example, fog is used as an example to illustrate an environment (weather) in which the sensor's detection sensitivity decreases, but this environment is not limited to fog. For example, environments in which the sensor's detection sensitivity decreases can include various situations such as rain, snow, nighttime, fire, smoke, smog, and dust caused by volcanic eruptions.

[0079] First, use Figure 3 The specific operation of existing path generation devices is explained. Figure 3In (a), the mobile body 100 moves on lane L1 and is located at point P1. Additionally, a shop S is located beside lane L1 at point P3. At point P1, the obstacle information D2 obtained by the mobile body 100 indicates that an obstacle H (e.g., a parked vehicle) may exist at point P3, 100-120 meters in front of the mobile body 100, and the detection reliability of obstacle H is 60%. A cloud-shaped symbol C, depicted around obstacle H, simulates the obstacle detection reliability; the larger the cloud-shaped symbol C, the lower the detection reliability. The dashed arrow R1 extending from the mobile body 100 indicates the planned movement path of the mobile body 100 within a certain time period from the current moment. Furthermore, the length of the arrow is proportional to the speed of the mobile body 100.

[0080] exist Figure 3 In (a), although the existing path generation device can infer that there may be an obstacle H near the location P3 at the location P1, its detection reliability is low, so it is determined that no obstacle avoidance action will be taken.

[0081] exist Figure 3 In (b), the moving body 100 from Figure 3 As shown in (a), the moving body 100 moves along the direction of arrow R1 and is located at point P2. Furthermore, at the moment of reaching point P2, the obstacle information D2 obtained by the moving body 100 indicates that an obstacle H may exist at a location 50 meters in front of the moving body 100, and the detection reliability is 99%. Figure 3 In (b), the moving body 100 did not take any evasive action while moving from point P1 to point P2, therefore the speed of the moving body 100 at point P2 is the same as its speed at point P1. Therefore, the moving body 100 definitely needs to evade at point P2. Thus, the moving body 100 needs to apply braking. However, due to the close proximity to obstacle H, the moving body 100 has to perform emergency braking or a sharp change in its course.

[0082] Therefore, the moving body 100 cannot avoid the obstacle H with a natural sense of movement.

[0083] Figure 4 This is another example concerning the operation of existing path generation devices. Figure 4 In (a), with Figure 3 Similarly, in (a) the mobile body 100 is located at point P1 on lane L1. Furthermore, at point P1, the obstacle information D2 obtained by the mobile body 100 indicates that an obstacle H may exist at point P3, 100-120 meters in front of the mobile body 100, and its detection reliability is 60%. Figure 3 The difference in the examples shown is that, in Figure 4 In the case of P1, although the detection reliability was low, a decision was made to avoid obstacle H by stopping.

[0084] Figure 4 (b) illustrates a situation where, at the moment the moving body 100 arrives at location P2, the obstacle H is almost certainly present (e.g., with a detection reliability of over 99%). In this case, the moving body 100 has already decelerated towards a stop and is thus able to stop in front of the obstacle H.

[0085] on the other hand, Figure 4 (c) illustrates the case where there is no obstacle H at the moment the moving body 100 arrives at location P2. In this case, the moving body 100 undergoes unnecessary deceleration. Furthermore, the deceleration of the moving body 100 is intended to bring it to a stop, resulting in a significant decrease in the speed of the moving body 100. To quickly return the speed of the moving body 100 to its original speed, it needs to be accelerated again, but the acceleration becomes abrupt.

[0086] In this situation, the moving body 100 is also unable to avoid the obstacle H naturally.

[0087] Next, the operation of the path generation device 10 in this embodiment will be explained. Figure 5 and Figure 6 This diagram illustrates the specific operation of the path generation device 10 in this embodiment.

[0088] exist Figure 5 In (a), with Figure 3 Similarly, (a) the mobile body 100 is located at location P1. Additionally, the shop S is located beside lane L1 at location P3. At location P1, the obstacle information D2 obtained by the mobile body 100 indicates that an obstacle H may exist at location P3, 100-120 meters in front of the mobile body 100, and the detection reliability of obstacle H is 60%. A cloud-shaped symbol C, depicted around obstacle H, simulates the obstacle detection reliability; the larger the cloud-shaped symbol C, the lower the detection reliability.

[0089] exist Figure 5 In (a), the collision risk assessment unit 13, at location P1, uses obstacle information D2, movement status information D4, event information D5, risk map information D6, and movement path information D7 to estimate the collision risk between the moving body 100 and the obstacle H. Furthermore, based on the estimated collision risk, the collision risk assessment unit 13 determines whether the moving body 100 needs to avoid the obstacle, and what the preparatory or formal actions should be.

[0090] exist Figure 5In case (a), the estimation unit 14 referred to the event information D5 and the risk map information D6, and confirmed that there was a store S along the road near location P3. Furthermore, it was confirmed that vehicles used for loading and unloading goods (i.e., obstacles H) frequently stopped on the road in front of store S, resulting in multiple rear-end collisions in the past. Based on the above results, the estimation unit 14 estimated the collision risk near location P3 to be moderate.

[0091] At location P1, the probability (detection reliability) of the presence of obstacle H is moderate at 60%, and whether moving body 100 must stop is uncertain. However, as Figure 5 As shown in (b), the collision risk is moderate, therefore the collision risk assessment unit 13 prepares for a situation where it is deemed very necessary to stop the moving body 100. In other words, the collision risk assessment unit 13 determines that a light braking (i.e., a preparatory action) is required in advance to allow the moving body 100 to stop smoothly. Furthermore, the path generation unit 16 generates a path, as shown by arrow R3, to decelerate the moving body 100 based on the assessment information D9.

[0092] then, Figure 5 (c) shows a situation where the moving body 100 has moved to location P2, thus almost confirming the presence of obstacle H. In this case, the collision risk assessment unit 13 determines that it needs to stop near obstacle H. Furthermore, the path generation unit 16 further modifies the movement path according to the assessment information D9, as shown by arrow R4, in order to stop the moving body 100 near obstacle H. In other words, the path generation unit 16 further decelerates the moving body 100 compared to the speed shown by the movement path of arrow R4, and stops it near obstacle H. By reducing the speed of the moving body 100 in stages, the deceleration acceleration of the moving body 100 until reaching obstacle H becomes slow.

[0093] Therefore, the moving body 100 is able to avoid (stop at) the obstacle H with a natural sense of movement.

[0094] Figure 5 (d) illustrates an example where the moving body 100 moves to location P2, thereby determining that the obstacle H does not exist. In this case, the collision risk assessment unit 13 determines that the preparatory action for avoidance has ended. Furthermore, the path generation unit 16 generates an accelerated movement path to bring the moving body 100 back to the speed defined in the movement path.

[0095] In this situation, the moving body 100 only decelerates minimally, and can return to its original speed with only a slight acceleration. Therefore, the moving body 100 can avoid the obstacle H with a natural feel.

[0096] Next, use Figure 6Let me illustrate with another example. Figure 6 The example shown illustrates that the collision risk assessment unit 13 considers the distance to the obstacle H, the detection reliability, the speed of the moving body 100, etc., and retains the assessment until the distance gets slightly closer, even though the obstacle H may be present.

[0097] exist Figure 6 In (a), with Figure 5 Similarly, (a) the mobile body 100 is located at location P1. Additionally, the shop S is located beside lane L1 at location P3. At location P1, the obstacle information D2 obtained by the mobile body 100 indicates that an obstacle H may exist at location P3, 100-120 meters in front of the mobile body 100, and the detection reliability of obstacle H is 60%. A cloud-shaped symbol C, depicted around obstacle H, simulates the obstacle detection reliability; the larger the cloud-shaped symbol C, the lower the detection reliability.

[0098] Figure 6 (b) shows a situation where the moving body 100 moves to location P4 and gets closer to obstacle H, increasing the detection reliability from 60% to 75%. At this moment, the collision risk assessment unit 13 determines that a preparatory action is needed to avoid the obstacle. Next, the collision risk assessment unit 13 uses the movement status information D4, event information D5, risk map information D6, and movement path information D7 to determine the congestion situation around the moving body 100 (i.e., the probability of collision with other moving bodies).

[0099] exist Figure 6 In (b), there are no other moving bodies around the moving body 100, therefore the collision risk assessment unit 13 determines that the congestion of other moving bodies around the moving body is low. Furthermore, compared with the above... Figure 5 Unlike other examples, the collision risk assessment unit 13 determines that the obstacle should be avoided by changing the movement path instead of stopping the moving body 100. The path generation unit 16 generates the movement path shown by arrow R6 based on the judgment information D9 of the collision risk assessment unit 13, so that the moving body 100 changes its travel route to the adjacent lane L2.

[0100] exist Figure 6 In (c), it is shown that the moving body 100 has moved to location P5 and has almost confirmed the presence of obstacle H. In this case, the moving body 100 has changed its route to the adjacent lane L2, and thus continues to move as before, passing by obstacle H.

[0101] Figure 6(d) shows the case where there is no obstacle H. In this case, the collision risk assessment unit 13 makes a judgment to end the avoidance action. Moreover, the path generation unit 16 generates a path back to the original movement path as shown by arrow R7.

[0102] As described above, the path generation device 10 determines the avoidance method based on the surrounding conditions of the moving body 100. In the presence of an obstacle H, it changes the predetermined position of movement to avoid the obstacle H. In the absence of an obstacle H, it can quickly return to the original (before the change) predetermined movement path without affecting the movement of surrounding moving bodies. Therefore, the moving body 100 can avoid the obstacle H with a natural feeling.

[0103] exist Figure 5 , 6 In the example, the information indicating the probability of a collision between the moving body 100 and the obstacle H, i.e., the collision risk information D8, is inferred from the fact that vehicles used for loading and unloading goods frequently stop on the road in front of the store S near location P3, resulting in multiple rear-end collisions in the past. Based on the movement state of the moving body 100 and the collision risk information D8, the collision risk judgment unit 13 outputs information indicating the judgment result of the moving body 100 in avoiding the obstacle H, i.e., judgment information D9. Furthermore, the path generation unit 16 can generate a movement path for the moving body 100 based on the judgment information D9. Thus, the moving body 100 can avoid the obstacle H with a natural feel.

[0104] As described above, in the path generation apparatus 10 of this embodiment, the collision risk judgment unit 13 not only relies on information obtained from various sensors (i.e., obstacle information D2), but also uses information indicating the probability of collision with obstacles and other moving bodies (i.e., event information D5, risk map information D6) to make a comprehensive judgment. Therefore, when the possibility of the presence of an obstacle H arises, the moving body 100 can perform the necessary minimum preparatory actions. As a result, the moving body 100 can reliably avoid the obstacle H with a natural feeling, or reliably stop at the obstacle H. In addition, even when there is no obstacle H, the path generation apparatus 10 of this embodiment can make the moving body 100 quickly return to the speed defined in the movement path with only a minimum deceleration.

[0105] Therefore, in the path generation device 10 of this embodiment, the moving body 100 can avoid obstacles with a natural sense in various situations.

[0106] Explanation of the sequence of actions

[0107] Next, the operating sequence of the path generation device 10 in this embodiment will be explained. Figure 7This is a flowchart illustrating the operation of the path generation device 10.

[0108] In step ST10, the movement state estimation unit 12 uses the movement body information D1 and map information D3 to estimate the movement state of the movement body 100 (step ST10).

[0109] In step ST11, the estimation unit 14 uses obstacle information D2, event information D5, risk map information D6, and movement path information D7 to estimate collision risk information D8, which indicates the possibility of collision with an obstacle, when the moving body 100 moves according to the movement path information D7 (step ST11).

[0110] In step ST12, the judgment unit 15 uses the movement state information D4 and the collision risk information D8 to determine whether the moving body 100 needs to avoid the obstacle, and outputs judgment information D9 indicating the avoidance judgment result (step ST12).

[0111] In step ST13, the path generation unit 16 uses the movement state information D4, the movement path information D7, and the judgment information D9 to generate the movement path information D7 based on the judgment information D9 (step ST13). Specifically, based on the judgment information D9, the path generation unit 16 generates a movement path in which the moving body 100 performs a preparatory action for avoiding obstacles, such as light braking (deceleration); a movement path in which the moving body 100 performs a formal action for avoiding obstacles, such as braking (deceleration) or changing the route; or a movement path in which the moving body 100 maintains its current route and speed.

[0112] In step ST14, it is determined whether the mobile body 100 has reached its destination. If the mobile body 100 has reached its destination (yes in step ST14), this process ends. If it has not reached its destination (no in step ST14), the process proceeds to step ST10 again. Furthermore, the estimation of the movement state of the mobile body 100, the estimation of the collision risk with obstacles, the output of the judgment information D9, and the generation of the movement path of the mobile body 100 are performed respectively.

[0113] <Effects of Implementation Method 1>

[0114] As explained above, according to the path generation apparatus of this embodiment, regarding the estimation of the collision risk between the moving body 100 and obstacles, and the method by which the moving body 100 avoids obstacles, the collision risk judgment unit 13 not only relies on information obtained from various sensors 11 (i.e., obstacle information D2), but also uses information indicating the probability of collision with obstacles and other moving bodies (i.e., event information D5, risk map information D6) to make a comprehensive judgment and output judgment information D9. Furthermore, the path generation unit 16 generates a movement path based on the judgment information D9.

[0115] As a result, the moving body 100 can determine the possibility of collision with an obstacle as early as possible and can move along a path that allows for appropriate avoidance maneuvers. Therefore, the moving body 100 can avoid obstacles with a natural sense of movement.

[0116] In the above embodiments, an autonomous vehicle was described as an example of a moving body, but the moving body is not limited to autonomous vehicles. For example, the moving body can also be a railway, a ship, an airplane, etc. For example, when the moving body is a railway, the path generation apparatus of this disclosure can consider the track as a moving path and thereby generate a moving path. In addition, for example, when the moving body is a ship, the path generation apparatus of this disclosure can consider a waterway or waterway as a moving path and thereby generate a moving path.

[0117] Industrial availability

[0118] The path generation apparatus disclosed herein can be applied to various mobile bodies such as automobiles, personal mobility vehicles, railways, ships, airplanes, and drones. In particular, the path generation apparatus disclosed herein is suitable for use in autonomous vehicles.

[0119] Explanation of reference numerals in the attached figures

[0120] 10…Path generation device; 11…Various sensors; 12…Movement state estimation unit; 13…Collision risk assessment unit; 14…Estimation unit; 15…Judgment unit; 16…Path generation unit; 17…Map information storage unit; 18…Event information storage unit; 19…Risk map information storage unit; 20…Movement path information storage unit; 21…Control unit; 31…Processor; 32…Memory; 33…Storage device; 34…Interface; 35…Communication unit; 100…Moving body.

Claims

1. A path generation device, characterized in that, have: The movement state estimation unit uses information obtained from sensors to estimate the movement state of the moving body; The collision risk assessment unit outputs information representing the assessment result of the moving body's attempt to avoid the obstacle, i.e., assessment information, based on the movement state and information indicating the possibility of the moving body colliding with the obstacle, i.e., collision risk information. as well as The path generation unit generates the movement path of the moving body based on the judgment information.

2. The path generation device according to claim 1, characterized in that, Based on the movement state and the collision risk information, the collision risk assessment unit determines the preparatory action or avoidance action of the moving body to avoid the obstacle. The path generation unit generates a movement path based on the preparatory action or the avoidance action.

3. The path generation device according to claim 1 or 2, characterized in that, The collision risk information includes at least one of the following: information indicating the state of a past accident or near-accident, i.e., event information, and information related to the potential danger around the moving body, i.e., risk map information.

4. A path generation method, characterized in that, The motion state estimation unit uses information obtained from sensors to estimate the motion state of the moving body. Based on the movement state and information indicating the probability of the moving body colliding with an obstacle (i.e., collision risk information), the collision risk assessment unit outputs information indicating the assessment result of the moving body's attempt to avoid the obstacle (i.e., assessment information). The path generation unit generates the movement path of the moving body based on the judgment information.

Images