Narrow road planning method and device, storage medium, autonomous vehicle, remote terminal and computer program product

Abstract

The invention discloses a narrow road planning method and device, a storage medium, an automatic driving vehicle, a remote terminal and a computer program product. The method comprises the following steps: determining a working mode of the automatic driving vehicle in the last round; trying to generate a low-precision track and a high-precision track under the condition that the working mode of the last round is narrow path detection; if the low-precision track is not successfully generated, or the low-precision track is successfully generated and has a collision risk, judging a generation result of the high-precision track; if the high-precision track is successfully generated and there is a collision risk, sending a first indication signal to a remote terminal; if a first feedback signal from the remote terminal is received, the working mode is switched from narrow road detection to narrow road passing and serves as the working mode of the current wheel, automatic driving is carried out based on the successfully generated high-precision track, and the first feedback signal is used for indicating that safe passing can be carried out on the high-precision track. According to the scheme, invalid and repeated high-precision trajectory planning can be avoided in a narrow road scene.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a narrow road planning method and device, a storage medium, an autonomous driving vehicle, a remote terminal, and a computer program product. Background Art

[0002] In the field of autonomous driving, unmanned vehicles face a variety of road conditions. For autonomous sanitation vehicles (such as street sweepers), narrow roads are a common scenario in their daily operations. For example, in parks and non-motorized vehicle lanes where sanitation vehicles operate, vehicles and debris often accumulate and block the road. In some cases, they can pass safely, while in others, the road is completely blocked and global routing must be redefined. Furthermore, compared to other types of unmanned vehicles, sanitation vehicles have several sweeping brushes that extend beyond the vehicle body for cleaning, making them less user-friendly for planning in narrow road scenarios. Autonomous driving systems need to comprehensively consider road condition information to provide reasonable vehicle driving plans.

[0003] In order to avoid collision risks, existing autonomous vehicle path planning schemes may, on the one hand, maintain high-precision trajectory planning all the time when facing narrow road scenarios. However, high-precision trajectory planning is more time-consuming and resource-intensive in data acquisition, and is less efficient. On the other hand, when it is confirmed that the high-precision trajectory has a collision risk, it will continue to enter the narrow road detection mode and perform trajectory generation and trajectory judgment operations (including whether the trajectory is successfully generated and whether there is a collision risk, etc.). However, when encountering extremely narrow roads or even completely blocked roads and other scenarios where the road is impassable, the vehicle will stay in place and continue to perform trajectory generation and trajectory judgment operations in a loop. This not only causes ineffective consumption of computing resources, but also leads to unpredictable dangers due to road occupation. Summary of the Invention

[0004] The technical problem solved by the embodiments of the present invention is how to ensure the safety of autonomous driving in narrow road scenarios while avoiding invalid and repetitive high-precision path planning as much as possible, thereby reducing computing time and resource overhead.

[0005] To solve the above technical problems, an embodiment of the present invention provides a narrow road planning method, comprising the following steps: determining an operating mode of an autonomous driving vehicle in the previous round, wherein the operating modes include at least narrow road detection and narrow road passage; if the operating mode in the previous round was narrow road detection, attempting to generate a low-precision trajectory and a high-precision trajectory; first determining the generation result of the low-precision trajectory; if the low-precision trajectory was not successfully generated, or if the low-precision trajectory was successfully generated but there is a collision risk, then determining the generation result of the high-precision trajectory; if the high-precision trajectory was successfully generated and there is a collision risk, sending a first indication signal to a remote terminal, the first indication signal being used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the perceived environment data used to generate the high-precision trajectory; if a first feedback signal is received from the remote terminal, switching the operating mode from narrow road detection to narrow road passage as the operating mode for the current round, and performing autonomous driving based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0006] Optionally, in the case that the working mode of the previous round is narrow road detection, the method further includes: if the high-precision trajectory is not successfully generated, or a second feedback signal is received from the remote terminal, determining the target parking area, and controlling the autonomous driving vehicle to drive to the target parking area and park, wherein the second feedback signal is used to indicate that the high-precision trajectory is not safe to pass; if the high-precision trajectory is successfully generated and there is no collision risk, switching the working mode from narrow road detection to narrow road passing as the working mode of the current round, and performing autonomous driving based on the successfully generated high-precision trajectory.

[0007] Optionally, determining the target parking area includes: sending a second indication signal to the remote terminal, wherein the second indication signal is used to indicate that the target parking area is determined based on the current position of the autonomous driving vehicle and the perception environment data used to generate the high-precision trajectory and fed back to the autonomous driving vehicle.

[0008] Optionally, the working mode also includes non-narrow roads; in the case where the working mode of the previous round is narrow road detection, the method also includes: if the low-precision trajectory is successfully generated and there is no collision risk, the working mode is switched from narrow road detection to non-narrow road as the working mode of the current round, and automatic driving is performed based on the successfully generated low-precision trajectory.

[0009] Optionally, the working mode also includes non-narrow roads; when the working mode of the previous round is non-narrow roads, the method also includes: attempting to generate a low-precision trajectory; if the low-precision trajectory is successfully generated and there is no collision risk, the non-narrow road working mode of the previous round continues to be the working mode of the current round, and automatic driving is performed based on the successfully generated low-precision trajectory; if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a collision risk, the working mode is switched from non-narrow road to narrow road detection as the working mode of the current round.

[0010] Optionally, when the working mode of the previous round is narrow road passage, the method further includes: attempting to generate a low-precision trajectory and a high-precision trajectory; first judging the generation result of the low-precision trajectory, if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a collision risk, then judging the generation result of the high-precision trajectory; if the high-precision trajectory is not successfully generated, or the high-precision trajectory is successfully generated and there is a collision risk, then switching the working mode from narrow road passage to narrow road detection as the working mode of the current round.

[0011] Optionally, in the case that the working mode of the previous round is narrow road passage, the method further includes: if the low-precision trajectory is successfully generated and there is no collision risk, the working mode is switched from narrow road passage to narrow road detection as the working mode of the current round, and automatic driving is performed based on the successfully generated low-precision trajectory; if the high-precision trajectory is successfully generated and there is no collision risk, the narrow road passage working mode of the previous round continues to be the working mode of the current round, and automatic driving is performed based on the successfully generated high-precision trajectory.

[0012] Optionally, the high-precision trajectory and the low-precision trajectory are recorded as target trajectories, and the target trajectory includes multiple predicted positions of the autonomous driving vehicle; whether the target trajectory has a collision risk is judged in the following manner: obtaining perception environment data used to generate the target trajectory, the perception environment data including the position of each obstacle within a preset area around the autonomous driving vehicle; judging whether the distance between the position of each obstacle and the nearest predicted position on the target trajectory is greater than a preset safety distance threshold; if the distance between the positions of all obstacles involved in the judgment and the nearest predicted position on the target trajectory is greater than the safety distance threshold, it is confirmed that the target trajectory has no collision risk.

[0013] Optionally, the working mode of the previous round is different, and the corresponding safety distance threshold is different; among them, the safety distance threshold corresponding to the non-narrow road working mode of the previous round, the safety distance threshold corresponding to the narrow road detection working mode of the previous round, and the safety distance threshold corresponding to the narrow road passage working mode of the previous round are in an increasing relationship.

[0014] An embodiment of the present invention also provides a narrow road planning method, including: receiving a first indication signal; obtaining a successfully generated high-precision trajectory and perception environment data used to generate the high-precision trajectory from an autonomous driving vehicle; judging whether the high-precision trajectory is safe to pass based on the obtained high-precision trajectory and the perception environment data; if it is confirmed that the high-precision trajectory is safe to pass, sending a first feedback signal to the autonomous driving vehicle, the first feedback signal being used to indicate that the high-precision trajectory is safe to pass.

[0015] Optionally, the high-precision trajectory includes multiple predicted positions of the autonomous driving vehicle, and the perception environment data used to generate the high-precision trajectory includes the position and contour information of each obstacle within a preset area around the autonomous driving vehicle; based on the acquired high-precision trajectory and the perception environment data, judging whether the high-precision trajectory is safe to pass, including: projecting the high-precision trajectory and the position and contour information of each obstacle within the preset area around the autonomous driving vehicle onto the map coordinate system of the driving scene to which the autonomous driving vehicle belongs, and visually displaying them; receiving a judgment result from the outside, and the judgment result is used to indicate whether the high-precision trajectory is safe to pass.

[0016] Optionally, the method further includes: if it is confirmed that the high-precision trajectory is not safe to pass, sending a second feedback signal to the autonomous driving vehicle, wherein the second feedback signal is used to indicate that the high-precision trajectory is not safe to pass.

[0017] Optionally, after sending a second feedback signal to the autonomous driving vehicle, the method further includes: receiving a second indication signal; obtaining the current position of the autonomous driving vehicle and the perception environment data used to generate the high-precision trajectory from the autonomous driving vehicle; determining the target parking area based on the obtained current position of the autonomous driving vehicle and the perception environment data, and feeding back the result to the autonomous driving vehicle.

[0018] An embodiment of the present invention further provides a narrow road planning device, comprising: an operating mode determination module, configured to determine the operating mode of the autonomous driving vehicle in the previous round, wherein the operating modes include at least narrow road detection and narrow road passage; a trajectory generation module, configured to attempt to generate a low-precision trajectory and a high-precision trajectory if the operating mode in the previous round was narrow road detection; a trajectory judgment module, configured to first judge the generation result of the low-precision trajectory, and if the low-precision trajectory is not successfully generated, or if the low-precision trajectory is successfully generated and there is a collision risk, then judge the generation result of the high-precision trajectory; a remote signal sending module, configured to send a first indication signal to a remote terminal if the high-precision trajectory is successfully generated and there is a collision risk, the first indication signal being used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the perceived environment data used to generate the high-precision trajectory; and an operating mode switching module, configured to switch the operating mode from narrow road detection to narrow road passage as the operating mode for the current round upon receiving a first feedback signal from the remote terminal, and perform autonomous driving based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0019] An embodiment of the present invention also provides a narrow road planning device, comprising: a remote signal receiving module for receiving a first indication signal; a data acquisition module for acquiring a successfully generated high-precision trajectory and the perception environment data used to generate the high-precision trajectory from an autonomous driving vehicle; a narrow road pass judgment module for judging whether the high-precision trajectory is safe to pass based on the acquired high-precision trajectory and the perception environment data; and a feedback module for sending a first feedback signal to the autonomous driving vehicle if it is confirmed that the high-precision trajectory is safe to pass, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0020] An embodiment of the present invention further provides a storage medium storing a computer program. When the computer program is executed by a processor, the steps of the narrow road planning method are performed.

[0021] An embodiment of the present invention also provides an autonomous driving vehicle, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of the above-mentioned narrow road planning method when running the computer program.

[0022] An embodiment of the present invention further provides a remote terminal comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of the above-mentioned narrow road planning method when running the computer program.

[0023] An embodiment of the present invention further provides a computer program product, comprising a computer program, which executes the steps of the above-mentioned narrow road planning method when executed by a processor.

[0024] Compared with the prior art, the technical solution of the embodiment of the present invention has the following beneficial effects:

[0025] In an embodiment of the present invention, the working mode is switched and the trajectory is planned according to the adaptability of the working mode of the autonomous driving vehicle in the previous round. Specifically, on the one hand, narrow road detection, an intermediate working mode between narrow road passage and non-narrow road, is set to enable dynamic switching between non-narrow road and narrow road passage, and to ensure that high-precision trajectories are calculated only when necessary, thereby taking into account both safe passage and reduced computational burden. On the other hand, when it is confirmed that the high-precision trajectory is successfully generated but there is a risk of collision, instead of cyclically entering the narrow road detection mode and performing trajectory generation and trajectory risk judgment operations as in the prior art, a remote terminal is introduced to further confirm whether the high-precision trajectory is safe to pass; if the feedback result from the remote cloud indicates that the high-precision trajectory is safe to pass, the working mode is switched from narrow road detection to narrow road passage as the working mode for the current round.

[0026] In other words, this implementation is equivalent to a two-pronged process for determining whether a narrow road is safe to navigate. The first round is performed by the autonomous vehicle based on its high-precision trajectory. If the first round confirms that the high-precision trajectory poses a collision risk, the remote terminal performs a second round of judgment. This effectively avoids the problem of existing technologies relying solely on the autonomous vehicle's trajectory judgment, which results in looping into narrow road detection mode and performing repeated and ineffective trajectory planning and judgment. This not only increases the probability of normal narrow road navigation, prevents autonomous vehicles from being stuck and occupying the road, but also effectively reduces unnecessary resource expenditure. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of a narrow road planning method according to an embodiment of the present invention;

[0028] Figure 2 This is a flowchart of a narrow road planning method when the previous working mode of the autonomous vehicle is narrow road detection;

[0029] Figure 3 This is a flowchart of a narrow road planning method when the autonomous vehicle's previous working mode is a non-narrow road;

[0030] Figure 4 This is a flowchart of a narrow road planning method when the autonomous vehicle's previous working mode was narrow road traffic;

[0031] Figure 5 is a flow chart of another narrow road planning method according to an embodiment of the present invention;

[0032] Figure 6 This is a schematic diagram of the driving state of an autonomous driving vehicle in a typical application scenario of an embodiment of the present invention;

[0033] Figure 7 This is a schematic structural diagram of a narrow road planning device according to an embodiment of the present invention;

[0034] Figure 8 It is a structural diagram of another narrow road planning device in an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the above-mentioned objects, features and beneficial effects of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0036] Reference Figure 1 , Figure 1 The flowchart of a narrow road planning method according to an embodiment of the present invention is shown. The narrow road planning method can be applied to autonomous vehicles, particularly autonomous sanitation vehicles (e.g., autonomous road sweepers, autonomous road garbage loaders, autonomous road sprinklers, etc.).

[0037] The narrow road planning method may include steps S11 to S15:

[0038] Step S11: determining the operating mode of the autonomous driving vehicle in the previous round, wherein the operating mode includes at least narrow road detection and narrow road passage;

[0039] Step S12: When the working mode of the previous round is narrow road detection, attempt to generate a low-precision trajectory and a high-precision trajectory;

[0040] Step S13: first determine the generation result of the low-precision trajectory. If the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated but there is a collision risk, then determine the generation result of the high-precision trajectory.

[0041] Step S14: If the high-precision trajectory is successfully generated and there is a collision risk, a first indication signal is sent to the remote terminal, where the first indication signal is used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the sensed environment data used to generate the high-precision trajectory;

[0042] Step S15: If a first feedback signal is received from the remote terminal, the operating mode is switched from narrow road detection to narrow road passage as the operating mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0043] Furthermore, the above steps S11 to S15 can be executed in a loop and iterative manner. When the next round is executed, the next round is the current round. Accordingly, the working mode of the current round determined by step S15 is the working mode of the previous round described in step S11.

[0044] In this application, a narrow road is a road scenario that autonomous vehicles, especially autonomous sanitation vehicles, will inevitably encounter during actual operation, as opposed to conventional non-narrow roads. Narrow roads specifically include, but are not limited to, roads where the width between the two lane boundaries is less than a preset width threshold, roads in certain areas (e.g., non-motorized vehicle lanes, sidewalks in residential areas and industrial parks, sidewalks on both sides of urban roads, etc.), and various roads where vehicles, debris, etc. accumulate and occupy the road.

[0045] In this application, the working modes of the autonomous driving vehicle may include "narrow road detection" and "narrow road passage", and may also include the three working modes of "non-narrow road". Among them, "non-narrow road" is used to represent common non-narrow road scenarios (for example, a road where the width between the two lane boundaries is greater than or equal to a preset width threshold, or a motor vehicle lane in a city / suburb / township area that is not occupied by vehicles / debris, etc.). In the "non-narrow road" working mode, usually only a low-precision trajectory (also called "normal trajectory") is required to keep the autonomous driving vehicle driving safely. "Narrow road detection" is an intermediate mode or state between "non-narrow road" and "narrow road passage". One of the purposes of setting up this intermediate working mode is to smoothly and safely transition to real narrow road passage, and secondly, to isolate some undesirable situations such as passing through a dead end that should not be allowed. "Narrow road passage" represents the passage stage of a real narrow road, and usually a high-precision trajectory is required to ensure that the autonomous driving vehicle passes the narrow road safely. Compared with low-precision trajectories, high-precision trajectories have higher accuracy. The amount of data required to generate the trajectory, the computational time and resource overhead are often greater, and the required travel speed is slower. These characteristics just meet the requirements for safe low-speed passage on narrow roads.

[0046] In the specific implementation of step S12, when the working mode of the previous round is narrow road detection, an attempt is made to generate a low-precision trajectory and a high-precision trajectory. The perception environment data used to generate the low-precision trajectory and the high-precision trajectory can be the same (for example, belonging to the same acquisition time or the same frame). The perception environment data can include but is not limited to: image data and / or video data within a preset area around the autonomous driving vehicle collected by an image sensor, and laser point cloud data within a preset area around the autonomous driving vehicle collected by a laser sensor.

[0047] In the specific implementation of step S13, if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated but there is a risk of collision, it means that the autonomous driving vehicle is likely to face a narrow road ahead, and the generation result of the high-precision trajectory can be further combined to confirm whether the narrow road ahead is safe to pass.

[0048] In the specific implementation of step S14, if a high-precision trajectory is successfully generated and there is a collision risk, one possibility is that the road ahead is extremely narrow or completely blocked, making it unsafe to pass. Another possibility is that the road ahead, while narrow, is actually safe to pass, but unstable perception (for example, perception errors caused by changes in vehicle posture) may lead to a misjudgment of the collision risk. To avoid the latter misjudgment, a remote terminal can be further introduced to determine whether the high-precision trajectory is actually safe to pass.

[0049] In the specific implementation of step S15, if it is confirmed that the remote terminal feedback indicates that the high-precision trajectory is safe to pass, the working mode is switched from narrow road detection to narrow road passing as the working mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory.

[0050] Furthermore, the autonomous driving vehicle is an autonomous driving sweeper, and the autonomous driving sweeper includes a retractable cleaning device. Before performing autonomous driving based on the successfully generated high-precision trajectory, the method also includes: retracting the cleaning device.

[0051] It should be pointed out that in the prior art, when it is confirmed that the high-precision trajectory has a collision risk, the narrow road detection working mode of the previous round is continued as the working mode of the current round, and the process returns to execute the trajectory planning operation of step S12 and the trajectory generation situation judgment operation of step S13. This method will cause the autonomous driving vehicle to continuously and cyclically perform narrow road detection, which on the one hand causes repeated and ineffective waste of resource overhead, and on the other hand causes the vehicle to stay in place and cause road occupation problems.

[0052] In contrast, in an embodiment of the present invention, when it is confirmed that the high-precision trajectory has been successfully generated and there is a risk of collision, instead of switching to narrow road detection again, a remote terminal is introduced to further confirm whether the high-precision trajectory is safe to pass. If the feedback result from the remote cloud indicates that the high-precision trajectory is safe to pass, the working mode is switched from narrow road detection to narrow road passage as the working mode for the current round. In other words, this embodiment is equivalent to a two-round confirmation of whether the narrow road is safe to pass. That is, the first round is first determined by the autonomous driving vehicle (which may be referred to as the "self-vehicle" in this application) based on the high-precision trajectory. If the first round confirms that the high-precision trajectory has a collision risk, the remote terminal performs a second round of judgment. In this way, the problem of the existing technology of cyclically entering the narrow road detection mode and performing repeated and invalid trajectory planning and trajectory judgment due to reliance on the trajectory judgment result of the self-vehicle can be effectively avoided. This can not only increase the probability of normal passage of narrow roads, avoid the autonomous driving vehicle staying in place and occupying the road, but also effectively reduce unnecessary resource consumption.

[0053] Furthermore, the high-precision trajectory and the low-precision trajectory are recorded as target trajectories, and the target trajectory includes multiple predicted positions of the autonomous driving vehicle; whether the target trajectory has a collision risk is judged in the following manner: obtaining perception environment data used to generate the target trajectory, and the perception environment data includes the position of each obstacle within a preset area around the autonomous driving vehicle; judging whether the distance between the position of each obstacle and the nearest predicted position on the target trajectory is greater than a preset safety distance threshold; if the distance between the position of all obstacles involved in the judgment and the nearest predicted position on the target trajectory is greater than the safety distance threshold, it is confirmed that the target trajectory has no collision risk.

[0054] The safety distance threshold can be determined based on the width value of the autonomous driving vehicle. The larger the width value of the autonomous driving vehicle, the larger the safety distance threshold. For example, the result of multiplying the width value of the autonomous driving vehicle by a preset ratio can be used as the safety distance threshold. Without limitation, the preset ratio can be selected from [0.5, 1.2]. Furthermore, the safety distance threshold is different depending on the working mode of the previous round; the safety distance threshold corresponding to the non-narrow road working mode of the previous round, the safety distance threshold corresponding to the narrow road detection working mode of the previous round, and the safety distance threshold corresponding to the narrow road passage working mode of the previous round are in an increasing relationship.

[0055] It's understandable that the smaller the safety distance threshold, the more likely the target trajectory is to be judged as having a collision risk. In other words, the probability of a target trajectory being judged as having a collision risk when the previous operating mode was narrow road traffic is lower than when the previous operating mode was non-narrow road traffic. Thus, compared to setting identical safety distance thresholds for different operating modes, this implementation helps increase the probability of narrow road traffic.

[0056] Reference Figure 2 , Figure 2 This is a flow chart of a narrow road planning method when the last working mode of the autonomous driving vehicle is narrow road detection. The narrow road planning method may include the following steps S21 to S211.

[0057] In step S21, it is determined that the working mode of the autonomous driving vehicle in the previous round was narrow road detection.

[0058] In step S22 , an attempt is made to generate a low-precision trajectory and a high-precision trajectory.

[0059] In step S23 , the generation result of the low-precision trajectory is first determined.

[0060] If the low-precision trajectory is not generated successfully, or the low-precision trajectory is generated successfully but there is a collision risk, then continue to execute the following step S24; if the low-precision trajectory is generated successfully and there is no collision risk, then jump to execute the following step S210.

[0061] In step S24, the generation result of the high-precision trajectory is determined.

[0062] If the high-precision trajectory is successfully generated and there is a collision risk, the following step S25 is executed; if the high-precision trajectory is not successfully generated, the process jumps to step S29; if the high-precision trajectory is successfully generated and there is no collision risk, the process jumps to step S211.

[0063] In step S25, a first indication signal is sent to the remote terminal.

[0064] The first indication signal is used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the perception environment data used to generate the high-precision trajectory.

[0065] In step S26, receipt of the first feedback signal from the remote terminal is confirmed.

[0066] The first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0067] In step S27, the working mode is switched from narrow road detection to narrow road passage as the working mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory.

[0068] In step S28, receipt of the second feedback signal from the remote terminal is confirmed.

[0069] The second feedback signal is used to indicate that the high-precision trajectory is not safe to pass.

[0070] In step S29, a target parking area is determined, and the autonomous driving vehicle is controlled to travel to the target parking area and park.

[0071] It is understandable that, based on the self-vehicle's confirmation that the high-precision trajectory has a collision risk, if the remote cloud still feedbacks that the said high-precision trajectory is not safe to pass (equivalent to the judgment results of both wheels indicating that the narrow road ahead is impassable), this situation means that the road ahead is very narrow or is likely to be completely blocked. Therefore, the autonomous driving vehicle can be controlled to stop in an appropriate area to avoid staying there and occupying the road.

[0072] Furthermore, determining the target parking area includes: sending a second indication signal to the remote terminal, wherein the second indication signal is used to indicate that the target parking area is determined based on the current position of the autonomous driving vehicle and the perception environment data used to generate the high-precision trajectory and fed back to the autonomous driving vehicle.

[0073] In step S210, the working mode is switched from narrow road detection to non-narrow road detection as the working mode of the current wheel, and automatic driving is performed based on the successfully generated low-precision trajectory.

[0074] It can be understood that when the working mode of the previous round is narrow road detection, if the low-precision trajectory is successfully generated and there is no collision risk, it can be considered that the road ahead is most likely a non-narrow road. Therefore, the working mode can be switched to non-narrow road, and automatic driving can be performed based on the successfully generated low-precision trajectory to ensure driving speed and save resource overhead.

[0075] Furthermore, the autonomous driving vehicle is an autonomous driving sweeper, and the autonomous driving sweeper includes a retractable cleaning device. Before performing autonomous driving based on the successfully generated low-precision trajectory, the method also includes: extending the cleaning device.

[0076] In step S211, the working mode is switched from narrow road detection to narrow road passage as the working mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory.

[0077] It is understandable that, based on the self-vehicle's confirmation that the high-precision trajectory has a collision risk, if the remote feedback result shows that the high-precision trajectory is actually safe to pass (equivalent to the remote feedback result of the second round indicating that the narrow road ahead is safe to pass), this situation means that the self-vehicle's judgment result in the first round may be wrong. Therefore, the working mode can be switched from narrow road detection to narrow road passing as the working mode of the current round, and automatic driving can be performed based on the successfully generated high-precision trajectory.

[0078] Reference Figure 3 , Figure 3 This is a flow chart of a narrow road planning method when the last working mode of the autonomous driving vehicle is a non-narrow road. The narrow road planning method may include the following steps S31 to S35.

[0079] In step S31, it is determined that the working mode of the automatic driving vehicle in the previous round was non-narrow road.

[0080] In step S32 , an attempt is made to generate a low-precision trajectory.

[0081] In step S33 , the generation result of the low-precision trajectory is determined.

[0082] If the low-precision trajectory is successfully generated and there is no collision risk, step S34 is executed; if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a collision risk, step S35 is executed.

[0083] In step S34, the non-narrow road working mode of the previous round is continued as the working mode of the current round, and automatic driving is performed based on the successfully generated low-precision trajectory.

[0084] In step S35, the working mode is switched from non-narrow road to narrow road detection as the working mode of the current wheel.

[0085] Reference Figure 4 , Figure 4 This is a flow chart of a narrow road planning method when the last working mode of the autonomous driving vehicle is narrow road traffic. The narrow road planning method may include the following steps S41 to S47.

[0086] In step S41, it is determined that the working mode of the automatic driving vehicle in the previous round was narrow road passage.

[0087] In step S42 , an attempt is made to generate a low-precision trajectory and a high-precision trajectory.

[0088] In step S43 , the generation result of the low-precision trajectory is first determined.

[0089] If the low-precision trajectory is not generated successfully, or the low-precision trajectory is generated successfully but there is a collision risk, then the process continues with step S44 ; if the low-precision trajectory is generated successfully and there is no collision risk, then the process jumps to step S46 .

[0090] In step S44, the generation result of the high-precision trajectory is judged again.

[0091] If the high-precision trajectory is not generated successfully, or the high-precision trajectory is generated successfully but there is a collision risk, then the process continues with step S45 ; if the high-precision trajectory is generated successfully and there is no collision risk, then the process jumps to step S47 .

[0092] In step S45, the working mode is switched from narrow road passage to narrow road detection as the working mode of the current wheel.

[0093] In step S46, the working mode is switched from narrow road passage to narrow road detection as the working mode of the current wheel, and automatic driving is performed based on the successfully generated low-precision trajectory.

[0094] In step S47, the narrow road traffic working mode of the previous round is continued as the working mode of the current round, and automatic driving is performed based on the successfully generated high-precision trajectory.

[0095] Reference Figure 5 , Figure 5 This is a flowchart of another narrow road planning method in an embodiment of the present invention. The another narrow road planning method can be applied to a remote terminal. The remote terminal can be selected from, but not limited to, a computer, mobile phone, tablet computer, cloud server, cloud platform, etc. that can communicate remote data with an autonomous driving vehicle.

[0096] The alternative narrow road planning method may include steps S51 to S54.

[0097] In step S51, a first instruction signal is received.

[0098] In step S52 , the successfully generated high-precision trajectory and the perception environment data used to generate the high-precision trajectory are obtained from the autonomous driving vehicle.

[0099] In step S53, whether the high-precision trajectory is safe to travel is determined based on the acquired high-precision trajectory and the perceived environment data.

[0100] Furthermore, the high-precision trajectory includes multiple predicted positions of the autonomous driving vehicle, and the perception environment data used to generate the high-precision trajectory includes the position and contour information of each obstacle within a preset area around the autonomous driving vehicle; in a specific embodiment, the step S53 may specifically include: projecting the high-precision trajectory and the position and contour information of each obstacle within the preset area around the autonomous driving vehicle to the map coordinate system of the driving scene to which the autonomous driving vehicle belongs, and visually displaying them; receiving a judgment result from the outside, and the judgment result is used to indicate whether the high-precision trajectory is safe to pass.

[0101] Furthermore, receiving the judgment result from the outside may include: manually judging the visually displayed result and receiving the manual judgment result from the outside (for example, receiving the judgment result by button input, voice input, text input, etc.). Thus, this embodiment combines the two judgment schemes of automatic judgment by the autonomous driving vehicle (i.e., the aforementioned judgment based on the distance between the position of the obstacle and the nearest predicted position on the trajectory) and judgment by the remote terminal (i.e., data visualization and manual judgment). This can avoid the problem of narrow roads that are originally safe to pass being misjudged as having a collision risk due to errors in the autonomous vehicle's results, thereby effectively increasing the probability of passing narrow roads.

[0102] In step S54, if it is confirmed that the high-precision trajectory is safe to pass, a first feedback signal is sent to the autonomous driving vehicle, where the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0103] Furthermore, if it is determined that the high-precision trajectory is not safe to pass, a second feedback signal is sent to the autonomous driving vehicle, where the second feedback signal is used to indicate that the high-precision trajectory is not safe to pass.

[0104] Furthermore, after sending a second feedback signal to the autonomous driving vehicle, the method also includes: receiving a second indication signal; obtaining the current position of the autonomous driving vehicle and the perception environment data used to generate the high-precision trajectory from the autonomous driving vehicle; determining the target parking area based on the obtained current position of the autonomous driving vehicle and the perception environment data, and feeding back the result to the autonomous driving vehicle.

[0105] Furthermore, in this application, different narrow road detection results can be graded and corresponding decision status indicators can be set. The specific grading and status indicator setting schemes are as follows:

[0106] In the case where the working mode of the previous round is narrow road detection, if the low-precision trajectory is successfully generated and there is no collision risk, the method further includes: confirming that the narrow road detection result is not a narrow road, and setting the decision state flag to LEVEL_0.

[0107] When the operating mode of the previous round is narrow road detection, if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated but there is a collision risk, and the high-precision trajectory is not successfully generated, then before determining the target parking area, the method further includes: confirming that the narrow road detection result is a level 1 narrow road and setting a decision status flag to LEVEL_1; wherein the step of determining the target parking area is triggered in response to detecting that the decision status flag is LEVEL_1.

[0108] In a case where the working mode of the previous round is narrow road detection, if the low-precision trajectory is not successfully generated or the low-precision trajectory is successfully generated with a collision risk, and the high-precision trajectory is successfully generated with a collision risk, then before sending the first indication signal to the remote terminal, the method further includes: confirming that the narrow road detection result is a secondary narrow road, and setting the decision state identifier to LEVEL_2; wherein the step of sending the first indication signal to the remote terminal is triggered in response to detecting that the decision state identifier is LEVEL_2.

[0109] In the case where the working mode of the previous round is narrow road detection, if the low-precision trajectory is not successfully generated or the low-precision trajectory is successfully generated with a collision risk, and the high-precision trajectory is successfully generated without a collision risk, the method further includes: confirming that the narrow road detection result is a passable narrow road, and setting the decision state flag to LEVEL_3.

[0110] The following table 1 shows the corresponding relationship between different narrow road detection results and narrow road classification and decision status identification:

[0111] Table 1

[0112]

[0113] Reference Figure 6 , Figure 6 It is a schematic diagram of the driving status of an autonomous driving vehicle in a typical application scenario of an embodiment of the present invention.

[0114] Specifically, in this example, the autonomous vehicle is an autonomous road sweeper, which includes a retractable cleaning device. During normal travel, the autonomous road sweeper first enters the non-narrow road working mode. When approaching a narrow road ahead, it enters the narrow road detection mode because the low-precision trajectory is blocked by an obstacle ahead. The narrow road detection mode generates a high-precision trajectory and performs a safety check. After ensuring safety, it enters the narrow road passage mode. The narrow road passage mode retracts the autonomous vehicle's cleaning device (e.g., a sweeping brush) and slowly passes through the narrow road using a high-precision trajectory. After passing the narrow road, the low-precision trajectory is generated normally, and the narrow road passage mode switches to the narrow road detection mode and then to the non-narrow road mode, extending the sweeping brush to resume normal cleaning, and driving using a low-precision trajectory.

[0115] Reference Figure 7 , Figure 7 Schematic diagram of a narrow road planning device according to an embodiment of the present invention. The narrow road planning device may include:

[0116] An operating mode determination module 71 is configured to determine the operating mode of the autonomous driving vehicle in the previous round, wherein the operating mode includes at least narrow road detection and narrow road passage;

[0117] A trajectory generation module 72 is configured to attempt to generate a low-precision trajectory and a high-precision trajectory when the working mode of the previous round is narrow road detection;

[0118] The trajectory judgment module 73 is used to first judge the generation result of the low-precision trajectory. If the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated but there is a collision risk, then judge the generation result of the high-precision trajectory;

[0119] a remote signal sending module 74 configured to send a first indication signal to a remote terminal if the high-precision trajectory is successfully generated and there is a collision risk, wherein the first indication signal is used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the sensed environment data used to generate the high-precision trajectory;

[0120] The operating mode switching module 75 switches the operating mode from narrow road detection to narrow road passage as the operating mode of the current wheel upon receiving the first feedback signal from the remote terminal, and performs automatic driving based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0121] For the principle, specific implementation and beneficial effects of the narrow road planning device, please refer to the previous article and Figures 1 to 4 The description of the narrow road planning method shown above will not be repeated here.

[0122] Reference Figure 8 , Figure 8 This is a schematic diagram of the structure of another narrow road planning device according to an embodiment of the present invention. The narrow road planning device may include:

[0123] A remote signal receiving module 81 is configured to receive a first indication signal;

[0124] a data acquisition module 82 for acquiring, from the autonomous driving vehicle, a successfully generated high-precision trajectory and the perceived environment data used to generate the high-precision trajectory;

[0125] a narrow road passing judgment module 83, configured to judge whether the high-precision trajectory is safe to pass based on the acquired high-precision trajectory and the perceived environment data;

[0126] The feedback module 84 is configured to send a first feedback signal to the autonomous driving vehicle if it is determined that the high-precision trajectory is safe to pass. The first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

[0127] For the principle, specific implementation and beneficial effects of the narrow road planning device, please refer to the previous article and Figure 5 The description of the narrow road planning method shown above will not be repeated here.

[0128] The embodiment of the present invention further provides a storage medium, such as a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor. Figures 1 to 4 The steps of the narrow road planning method shown, or, performing the above Figure 5 The computer-readable storage medium may include a non-volatile memory or a non-transitory memory, and may also include an optical disk, a mechanical hard disk, a solid-state hard disk, etc.

[0129] Specifically, in the embodiment of the present invention, the processor may be a central processing unit (CPU), or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0130] It should also be understood that the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0131] An embodiment of the present invention further provides an autonomous driving vehicle, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the above-mentioned Figures 1 to 4 The steps of the narrow road planning method are shown. Without limitation, the autonomous driving vehicle can be an autonomous driving sanitation vehicle such as an autonomous driving sweeper or an autonomous driving sprinkler.

[0132] The embodiment of the present invention further provides a remote terminal, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the above-mentioned Figure 5 The steps of the narrow road planning method are shown.

[0133] The embodiment of the present invention further provides a computer program product, comprising a computer program, which executes the above-mentioned Figures 1 to 4 The steps of the narrow road planning method shown, or executing the above Figure 5The steps of the narrow road planning method are shown.

[0134] It should be understood that the term "and / or" as used herein simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " as used herein indicates that the related objects are in an "or" relationship.

[0135] The term "plurality" used in the embodiments of the present application refers to two or more.

[0136] The first, second, etc. descriptions appearing in the embodiments of this application are only for illustration and distinction of the description objects. There is no order, nor does it indicate any special limitation on the number of devices in the embodiments of this application, and cannot constitute any limitation on the embodiments of this application.

[0137] It should be noted that the serial numbers of the steps in this embodiment do not limit the execution order of the steps.

[0138] Although the present invention is disclosed above, and the present invention is not limited thereto, any person skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope defined by the claims.

Claims

1. A narrow road planning method, characterized in that: include: Determining the working mode of the autonomous driving vehicle in the previous round, wherein the working mode at least includes narrow road detection and narrow road passage; When the working mode of the previous round is narrow road detection, try to generate a low-precision trajectory and a high-precision trajectory; First, determine the generation result of the low-precision trajectory. If the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a risk of collision, then determine the generation result of the high-precision trajectory. If the high-precision trajectory is successfully generated and there is a risk of collision, a first indication signal is sent to the remote terminal, where the first indication signal is used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the sensed environment data used to generate the high-precision trajectory; If a first feedback signal is received from the remote terminal, the working mode is switched from narrow road detection to narrow road passage as the working mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

2. The method according to claim 1, characterized in that In the case where the working mode of the previous round is narrow road detection, the method further includes: If the high-precision trajectory is not successfully generated, or a second feedback signal is received from the remote terminal, a target parking area is determined, and the autonomous driving vehicle is controlled to drive to the target parking area and park, wherein the second feedback signal is used to indicate that the high-precision trajectory is not safe to pass; If the high-precision trajectory is successfully generated and there is no collision risk, the working mode is switched from narrow road detection to narrow road passage as the working mode of the current wheel, and automatic driving is performed based on the successfully generated high-precision trajectory.

3. The method according to claim 2, characterized in that: Determining the target parking area includes: A second indication signal is sent to the remote terminal, where the second indication signal is used to indicate that the target parking area is determined based on the current position of the autonomous driving vehicle and the perception environment data used to generate the high-precision trajectory and fed back to the autonomous driving vehicle.

4. The method according to claim 1, characterized in that: The working mode also includes non-narrow road; In the case where the working mode of the previous round is narrow road detection, the method further includes: If the low-precision trajectory is successfully generated and there is no collision risk, the working mode is switched from narrow road detection to non-narrow road as the working mode of the current wheel, and automatic driving is performed based on the successfully generated low-precision trajectory.

5. The method according to claim 1, characterized in that The working mode also includes non-narrow road; In the case where the working mode of the previous round is non-narrow road, the method further includes: Try to generate low-precision trajectories; If the low-precision trajectory is successfully generated and there is no collision risk, the non-narrow road working mode of the previous round will continue to be the working mode of the current round, and automatic driving will be performed based on the successfully generated low-precision trajectory; If the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a risk of collision, the working mode is switched from non-narrow road to narrow road detection as the working mode of the current wheel.

6. The method according to claim 1, characterized in that In the case where the working mode of the previous round is narrow road traffic, the method further includes: Try to generate low-precision and high-precision trajectories; First, determine the generation result of the low-precision trajectory. If the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a risk of collision, then determine the generation result of the high-precision trajectory. If the high-precision trajectory is not successfully generated, or the high-precision trajectory is successfully generated and there is a risk of collision, the working mode is switched from narrow road passage to narrow road detection as the working mode of the current wheel.

7. The method according to claim 6, characterized in that In the case where the working mode of the previous round is narrow road traffic, the method further includes: If the low-precision trajectory is successfully generated and there is no collision risk, the working mode is switched from narrow road passage to narrow road detection as the working mode of the current wheel, and automatic driving is performed based on the successfully generated low-precision trajectory; If the high-precision trajectory is successfully generated and there is no collision risk, the narrow road passage working mode of the previous round will continue to be used as the working mode of the current round, and automatic driving will be performed based on the successfully generated high-precision trajectory.

8. The method according to any one of claims 1 to 7, characterized in that: The high-precision trajectory and the low-precision trajectory are recorded as a target trajectory, and the target trajectory includes multiple predicted positions of the autonomous driving vehicle; The following method is used to determine whether the target trajectory has a collision risk: Acquire perceived environment data for generating the target trajectory, the perceived environment data including the location of each obstacle within a preset area around the autonomous driving vehicle; Determine whether the distance between the position of each obstacle and the nearest predicted position on the target trajectory is greater than a preset safety distance threshold; If the distances between the positions of all obstacles involved in the judgment and the closest predicted positions on the target trajectory are greater than the safety distance threshold, it is confirmed that there is no collision risk on the target trajectory.

9. The method according to claim 8, characterized in that The working mode of the previous round is different, and the corresponding safety distance threshold is different; Among them, the safety distance threshold corresponding to the non-narrow road in the previous round of working mode, the safety distance threshold corresponding to the narrow road detection in the previous round of working mode, and the safety distance threshold corresponding to the narrow road passage in the previous round of working mode are in an increasing relationship.

10. A narrow road planning method, characterized in that: include: receiving a first indication signal; Acquire a successfully generated high-precision trajectory and perceived environment data used to generate the high-precision trajectory from the autonomous driving vehicle; Determining whether the high-precision trajectory is safe to travel according to the acquired high-precision trajectory and the perceived environment data; If it is confirmed that the high-precision trajectory is safe to pass, a first feedback signal is sent to the autonomous driving vehicle, where the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

11. The method according to claim 10, characterized in that The high-precision trajectory includes multiple predicted positions of the autonomous driving vehicle, and the perceived environment data used to generate the high-precision trajectory includes the position and contour information of each obstacle within a preset area around the autonomous driving vehicle; Judging whether the high-precision trajectory is safe to travel according to the acquired high-precision trajectory and the perceived environment data includes: Projecting the high-precision trajectory and the position and contour information of each obstacle within a preset area around the autonomous driving vehicle to a map coordinate system of the driving scene to which the autonomous driving vehicle belongs, and visually displaying them; A judgment result is received from the outside, where the judgment result is used to indicate whether the high-precision trajectory is safe to travel.

12. The method according to claim 10, characterized in that The method further comprises: If it is confirmed that the high-precision trajectory is not safe to pass, a second feedback signal is sent to the autonomous driving vehicle, where the second feedback signal is used to indicate that the high-precision trajectory is not safe to pass.

13. The method according to claim 12, characterized in that After sending the second feedback signal to the autonomous driving vehicle, the method further includes: receiving a second indication signal; Acquiring from the autonomous vehicle a current position of the autonomous vehicle and perceived environment data for generating the high-precision trajectory; Based on the acquired current position of the autonomous driving vehicle and the perceived environment data, a target parking area is determined and fed back to the autonomous driving vehicle.

14. A narrow road planning device, characterized in that: include: A working mode determination module, used to determine the working mode of the autonomous driving vehicle in the previous round, wherein the working mode includes at least narrow road detection and narrow road passage; A trajectory generation module, used for attempting to generate a low-precision trajectory and a high-precision trajectory when the working mode of the previous round is narrow road detection; A trajectory judgment module, used to first judge the generation result of the low-precision trajectory, and if the low-precision trajectory is not successfully generated, or the low-precision trajectory is successfully generated and there is a risk of collision, then judge the generation result of the high-precision trajectory; A remote signal sending module, configured to send a first indication signal to a remote terminal if the high-precision trajectory is successfully generated and there is a risk of collision, wherein the first indication signal is used to indicate whether the high-precision trajectory is safe to pass based on the successfully generated high-precision trajectory and the sensed environment data used to generate the high-precision trajectory; The working mode switching module switches the working mode from narrow road detection to narrow road passage as the working mode of the current wheel if a first feedback signal is received from the remote terminal, and performs automatic driving based on the successfully generated high-precision trajectory, wherein the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

15. A narrow road planning device, characterized in that: include: A remote signal receiving module, used for receiving a first indication signal; A data acquisition module, used to acquire the successfully generated high-precision trajectory and the perceived environment data used to generate the high-precision trajectory from the autonomous driving vehicle; A narrow road passing judgment module, used to judge whether the high-precision trajectory is safe to pass according to the acquired high-precision trajectory and the perceived environment data; A feedback module is used to send a first feedback signal to the autonomous driving vehicle if it is confirmed that the high-precision trajectory is safe to pass, and the first feedback signal is used to indicate that the high-precision trajectory is safe to pass.

16. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the narrow road planning method according to any one of claims 1 to 9 are executed, or the steps of the narrow road planning method according to any one of claims 10 to 13 are executed.

17. An autonomous driving vehicle, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor runs the computer program, the steps of the narrow road planning method according to any one of claims 1 to 9 are performed.

18. A remote terminal comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor runs the computer program, the steps of the narrow road planning method according to any one of claims 10 to 13 are performed.

19. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the narrow road planning method according to any one of claims 1 to 9 are executed, or the steps of the narrow road planning method according to any one of claims 10 to 13 are executed.

Images