JP6680502B2 - Moving body - Google Patents

Description

  The present invention relates to a moving body, and more particularly to a moving body including an object recognition unit.

  BACKGROUND ART Conventionally, various types of mobile bodies for carrying people and goods are in circulation. Further, an autonomous traveling type moving body has also been proposed as a moving body, which allows it to be used not only for transportation purposes but also for surrounding monitoring (security). Then, regardless of the purpose of transportation, the purpose of monitoring, or whether or not the vehicle is of the autonomous traveling type, some moving bodies are equipped with an object recognition unit. The object recognition unit is provided to avoid collision with other moving objects and obstacles, and uses an active sensor for obstacle detection and a camera that captures the surroundings such as the front to detect obstacles from captured images. There is a thing that recognizes. Examples of active sensors include radio wave radars using reflected waves and laser range finders.

  Patent Document 1 discloses that when remotely controlling a semi-autonomous vehicle that can move autonomously, it is possible to safely perform an obstacle avoiding operation and a course change at an intersection without temporarily stopping or significantly decelerating. An unmanned vehicle for the purpose is disclosed. In this unmanned moving body, the setting of the obstacle avoiding route and the presence / absence determination of the obstacle on the obstacle avoiding route are repeated, and when the obstacle avoiding route cannot be set, the unmanned moving body is stopped.

  Patent Document 2 discloses a collision safety device intended to allow a moving body to smoothly travel even when traveling in an area where an obstacle exists. This collision safety device is a device mounted on a moving body having a collision absorbing structure at a site where there is a possibility of collision with an obstacle, and detects an obstacle around the moving body, and the obstacle is detected. In this case, deceleration control is performed. Then, the control value of the speed at which the deceleration control is performed is set at least according to the impact absorption performance of the collision site of the moving body with the obstacle.

  In Patent Document 3, when an obstacle is detected and the driving force is suppressed by driving assistance for avoiding a collision, the driving force suppressed in a form that reflects the driver's intention while ensuring safety. A driving support device for the purpose of returning is disclosed. This driving assistance device detects an obstacle outside the vehicle, and an obstacle detection unit that obtains obstacle information including the distance to the obstacle, and based on the obstacle information obtained by the obstacle detection unit, A collision avoidance control unit that performs a driving force suppression control that suppresses the driving force by intervention, and a driving force restoration control unit that performs a driving force restoration control that restores the driving force suppressed by the driving force suppression control, The driving force restoration control unit has a plurality of restoration modes for restoring the driving force by the driving force restoration control. Depending on the possibility that an obstacle exists around the host vehicle, the driving force restoration control unit selects one of the restoration modes. The driving force is restored by one restoration mode selected from the above.

JP, 2011-150512, A JP, 2011-936, A JP, 2015-47954, A

  By the way, especially for autonomous vehicles, the safety of not colliding at a speed that causes harm to people by reducing the speed as much as possible, and the functionality of preventing the phenomenon of being stuck in a stopped state as much as possible. , Is required to be compatible. It is to be noted that there is a demand for a technology that enables both functionality and safety to be achieved without the involvement of a person even in a moving body driven by a person.

  However, it is difficult for conventional techniques including the techniques described in Patent Documents 1 to 3 to achieve both such safety and functionality without taking an avoidance route or involving a person.

  The present invention has been made in view of the above situation, and an object thereof is to achieve both safety and functionality in a mobile body including an object recognition unit.

In order to solve the above problems, a first technical means of the present invention includes a drive unit, a control unit that controls drive by the drive unit, and an object recognition unit that performs recognition processing for recognizing an object. A moving body having a stop determination unit that determines whether or not the moving body needs to be stopped to avoid a collision with a recognized object, based on the result of the recognition process, and the recognition process. Based on the result of, whether or not it is necessary to decelerate the moving body in order to avoid a collision with a recognized object, a deceleration determination unit that determines in a processing time shorter than the stop determination unit, and control unit, to execute the determination in the deceleration determination unit, if the is determined that deceleration determination unit it is necessary to decelerate, the have line instructs deceleration to the drive unit, then in the stop determination unit It is necessary to execute the judgment and stop at the stop judgment unit. If it is determined that the driving unit is instructed to stop, and if it is determined that the stop determination unit does not need to stop, the deceleration state of the moving body is maintained. .

  A second technical means of the present invention is the first technical means, wherein the stop determination unit makes a determination based on data obtained by performing noise removal processing on the data of the recognition processing, and the deceleration determination unit, The determination is performed based on the data obtained as a result of the recognition processing or the data obtained by performing noise removal processing that is simpler than the noise removal processing on the data obtained as a result of the recognition processing.

  According to a third technical means of the present invention, in the first technical means, the stop determination unit makes a determination based on data obtained by correcting the ground as a result of the recognition processing, and the deceleration determination unit. Is characterized in that the determination is performed based on the data that is not corrected.

  According to a fourth technical means of the present invention, in the first technical means, the deceleration determination unit makes a determination based on data obtained by performing correction to exclude the ground from the data of the recognition processing result, and the stop determination unit. Is characterized in that the ground is excluded from the data obtained as a result of the recognition processing, the gradient amount of the ground is estimated, and the height of the ground is corrected to make a determination.

  A fifth technical means of the present invention is the first technical means, wherein the deceleration determination unit performs the recognition processing based on data obtained by performing thinning processing on the result data of the recognition processing or by reducing the recognition resolution. The stop determination unit makes a determination based on the executed data, and the stop determination unit makes a determination based on the data obtained as a result of the recognition processing without performing the thinning processing or the processing for lowering the recognition resolution. It is a thing.

  A sixth technical means of the present invention is the technical means according to any one of the first to fifth technical means, wherein a storage unit that stores map information including the planned travel route of the mobile body and a position indicating the position of the mobile body. A position information acquisition unit that acquires information, the control unit, based on the current position acquired by the position information acquisition unit, the drive unit to perform autonomous traveling along the planned traveling route. It is characterized by controlling.

  According to the present invention, in a moving body including an object recognition unit, first, the necessity of deceleration is determined based on the result of object recognition, deceleration is performed if necessary, and then the necessity of stop is determined based on the result of object recognition. Therefore, both safety and functionality can be achieved.

It is a block diagram which shows one structural example of the mobile body which concerns on the 1st Embodiment of this invention. It is an external view which shows an example of the mobile body of FIG. 1A. It is a flow diagram for explaining an example of a process in the moving body of FIG. 1A. It is a flow figure for explaining an example of processing in a mobile concerning a 2nd embodiment of the present invention. It is a flow figure for explaining an example of processing in a mobile concerning a 3rd embodiment of the present invention. It is a schematic diagram for demonstrating an example of the ground detection process in the process example of FIG. It is a schematic diagram for demonstrating an example of the ground detection process in the process example of FIG. It is a schematic diagram for demonstrating an example of the process in the mobile which concerns on the 4th Embodiment of this invention. It is a schematic diagram for demonstrating the other example of the process in the mobile which concerns on the 4th Embodiment of this invention. It is a flow figure for explaining an example of processing in a mobile concerning the present invention.

  The moving body according to the present invention is a moving body that is moved within a facility such as a factory or public facility, or within a site such as those facilities or a parking lot, or a moving body such as an automobile or a motorcycle that runs on a public road. In particular, some moving bodies that are automatically moved within a site or facility have an autonomous traveling control mechanism. A mobile body, which is basically driven by a driver such as an automobile, can also carry out autonomous traveling or autonomous traveling as a driving assistance for the driver by incorporating the autonomous traveling type control. Moreover, the moving body according to the present invention can be used not only for the purpose of carrying people or objects but also for monitoring the surroundings while moving, and in that case, the moving body can also be called a monitoring robot. Hereinafter, various embodiments of the present invention will be described with reference to the drawings.

(First embodiment)
A first embodiment of the present invention will be described with reference to FIGS. 1A and 2. First, a configuration example of the moving body according to the present embodiment will be described with reference to the block diagram of FIG. 1A and the external view of FIG. 1B.

  The moving body 1 is a machine provided with a moving mechanism for moving, and can also be called a moving device. In the example of FIGS. 1A and 1B, the moving mechanism includes a drive control unit 11 and a drive unit 12 including wheels 12 a controlled by the drive control unit 11. The drive unit 12 includes, for example, an engine and / or a motor for rotationally driving a plurality of wheels 12a (not shown). Of course, the track 12a (caterpillar (registered trademark)) or the like may be driven instead of the wheel 12a as illustrated.

  In addition, the moving body 1 is provided with a battery (rechargeable battery). The rechargeable battery is a portion that supplies electric power to each functional element of the vehicle, and includes, for example, a traveling function, an object recognition function (a road surface determination function may be provided in addition to an object distance detection function), a position information acquisition function, It is a part that supplies electric power to a part that realizes functions such as a communication function. As the rechargeable battery, for example, a lithium ion battery, a nickel hydrogen battery, a Ni-Cd battery, a lead battery, a fuel battery, or an air battery is used.

  Furthermore, the moving body 1 includes an object recognition unit 13, a storage unit 14, and a position information acquisition unit 15, and further includes a deceleration determination unit 10a, a stop determination unit 10b, and a speed instruction unit 10c. In FIG. 1A, an example in which the main control unit 10 for controlling the moving body 1 is provided with the respective units 10a to 10c is given, but the present invention is not limited to this. The storage unit 14 and the position information acquisition unit 15 are not essential components. The speed instruction unit 10c and the drive control unit 11 are an example of a control unit that controls the drive of the drive unit 12.

  The main control unit 10 controls the drive control unit 11 and reads / writes data to / from the storage unit 14, but is also configured to perform control such as recognition by the object recognition unit 13 and acquisition by the position information acquisition unit 15. You can also For example, the main control unit 10 includes a CPU (Central Processing Unit) or an MPU (Micro Processing Unit), a RAM (Random Access Memory) as a work area, and a control device such as a storage device, and a part or all of the control device is included. It can also be mounted as an integrated circuit / IC chip set. The storage device stores a control program (including a program for executing processing described later in each of the units 10a to 10c), various setting contents, and the like. As this storage device, various devices such as a HDD (Hard Disk Drive) and an SSD (Solid State Drive) can be applied.

  The object recognition unit 13 is a unit that performs recognition processing for recognizing various objects such as a plurality of fixed objects and other moving objects. Basically, the object recognition unit 13 may recognize the position of the object with respect to the moving body 1 (distance and direction from the moving body 1 to the object). Needless to say, the object recognition unit 13 has a detectable range, but a configuration that allows detection at all angles may be adopted, or a measurement space region is defined within the detectable range and within the measurement space region. You may recognize an object.

  As the object recognition unit 13, a camera or the like for shooting a still image or a moving image can be used. Also, disparity information can be obtained by providing two or more cameras. It can be said that the camera is an example of a passive type sensor for detecting an obstacle. It should be noted that the camera for taking a still image can cope with the movement of the moving body 1 by shortening the taking interval. As the object recognition unit 13, an active sensor for detecting an obstacle (hereinafter, active sensor) for detecting the position of the obstacle with respect to the moving body 1 can be used. The speed of the object can be calculated by the object recognition unit 13 in consideration of the moving speed and the traveling direction of the moving body 1.

  The active sensor is a sensor for detecting them in advance in order to avoid collision with other moving bodies or obstacles, and actively emits light, infrared rays, other electromagnetic waves, ultrasonic waves, etc. It is a sensor that receives the reflected wave of the transmitted wave and detects the position of the obstacle. As the active sensor, various types of sensors such as LIDAR (Light Detection and Ranging or Laser Imaging Detection and Ranging), a laser range finder, a radio wave radar (millimeter wave radar, etc.), an ultrasonic sensor, etc. can be applied. In addition, the active sensor is an electronic device that calculates a detection angle by using a mechanical scan method in which a motor is moved left and right as necessary, or by using a plurality of channels for reception and using a phase difference generated between the reception channels. The scanning method may be adopted to detect the obstacle. The laser range finder is a distance-measuring sensor that employs a time-of-flight (TOF) distance measurement method. By providing a scanning axis with one axis or two axes, two-dimensional plane measurement and three-dimensional measurement are possible. Various measurements are possible. It can also be said that LIDAR is a kind of laser range finder. As described above, when the active sensor is used as the object recognition unit 13, laser, infrared ray, visible light, ultrasonic wave, electromagnetic wave, or the like can be used as the light emitted for sensing. However, it is preferable to use a laser because of high weather resistance and high distance measurement accuracy.

  Further, the object recognition unit 13 is preferably provided in front of the main body 16 of the moving body 1 as illustrated in FIG. 1B, but may be provided in another portion as long as it can recognize an object in at least the traveling direction. However, the object recognition unit 13 may be provided with its detection surface (transmission / reception unit, antenna, etc.) at a position where its sensitivity is improved. Further, by providing a plurality of object recognition units 13 at different positions and / or by providing a plurality of types of object recognition units 13, the position of the object can be recognized more accurately.

  The moving body 1 has a performance (weather resistance) suitable for the traveling speed and traveling range assumed by the moving body 1 (for example, a range depending on whether it is a public road, how large the site or indoors is). (Including sex) and the type of object recognition unit 13 may be appropriately selected and installed. Of course, since the performance of the object recognition unit 13 changes depending on the cost, the object recognition unit 13 may be mounted in consideration of the cost.

  The storage unit 14 is a unit that stores map information (map data) 14a including the planned travel route of the mobile body 1. The map information 14a also includes object information indicating the positions of a plurality of fixed objects. The map information 14a also includes information indicating the position of a passable area (such as a road or a non-parking area in a parking lot). When the mobile 1 is a surveillance robot, the map information 14a can be said to be information indicating an environment map. Here, the fixed objects may include not only real estate such as buildings but also other structures such as railroad tracks, trees and street trees, walls and columns. It can be said that the fixed object is a concept that is generally called a feature, and that excludes a region (road or the like) that actually exists and is passable.

  The position information acquisition unit 15 uses GPS (Global Positioning System) or the like to acquire position information indicating the position of the mobile body 1. For example, the position information acquisition unit 15 includes an antenna that receives radio waves from GPS satellites, an analysis unit that analyzes the received GPS signal, and obtains position information (latitude, longitude). The antenna may be arranged at an appropriate position on the main body 16 of the mobile body 1 like the position of the position information acquisition unit 15 illustrated in FIG. 1B. However, the antenna is not limited to the illustrated arrangement, and the antenna may be provided at a position where the sensitivity is improved. Note that the position information acquisition unit 15 is provided with a function of DGPS (Differential GPS) or a function of RTK-GPS (Real Time Kinematic GPS), and corrects the position from the positional relationship with the wireless communication unit and the wireless communication base station. By providing a function, the accuracy can be improved. Further, as the position information, when the moving body is run only in a very limited range such as on the premises or indoors, it is possible to simply adopt Cartesian coordinates or polar coordinates.

  Although the position information acquiring unit 15 is described by exemplifying the part that acquires the GPS positioning information, another satellite positioning system (local navigation satellite system) similar to GPS can be applied. Other satellite positioning systems include Global Navigation Satellite System (GNSS) other than GPS, such as GLONASS (Global Navigation Satellite System) in Russia, Galileo in EU, and Beidou in China. Examples thereof include a Quasi-Zenith Satellite System (QZSS) and an Indian Regional Navigational Satellite System (IRNSS).

  As the network for wireless communication in the above-described wireless communication unit or the like, the Internet, which is open to the public, may be used, or a wireless network of a dedicated line in which devices that can be connected are limited may be used. Various wireless LANs (Local Area Networks) (whether or not WiFi (registered trademark) authentication is used), ZigBee (registered trademark), Bluetooth (registered trademark) LE (Low Energy), and the like are used as wireless transmission methods on the wireless communication path. There is a method based on the standard, which may be used in consideration of the wireless reach and the transmission band, but a mobile phone network or the like may be used, for example. Further, the wireless communication network as described above can also be used for transmitting and receiving route information (information designating a traveling route) from a server device or the like for managing the traveling route.

  A supplement will be given regarding the map information 14a. For example, when the mobile body 1 is used for monitoring in a specific area such as a parking lot or a site, a planned travel route is determined. Therefore, the mobile unit 1 stores the planned travel route of the mobile unit 1 as a part of the map information 14a in the storage unit 14, and based on the current position (current position on the map) acquired by the position information acquisition unit 15, The control may be performed so as to perform autonomous traveling along the planned traveling route. Note that this control is instructed from the main control unit 10 to the drive control unit 11 and is performed by the drive control unit 11 with respect to the drive unit 12.

  Supplement on autonomous driving. The moving body 1 includes a magnetic sensor or the like for measuring the geomagnetism, detects the traveling direction of the moving body 1 by detecting the direction in which the magnetic sensor faces, and outputs the detection result to the main control unit 10. It can also be configured to be used to correct the traveling direction.

  Next, the deceleration determination unit 10a, the stop determination unit 10b, and the speed instruction unit 10c, which are the main features of this embodiment, will be described. The deceleration determination unit 10a and the stop determination unit 10b can be configured by different chips or different substrates, and such a configuration facilitates the design and change of the moving body 1.

  The deceleration determination unit 10a determines whether or not it is necessary to decelerate the moving body 1 in order to avoid a collision with the recognized object based on the result of the recognition process. On the other hand, the stop determination unit 10b determines whether or not the moving body 1 needs to be stopped in order to avoid a collision with the recognized object, based on the result of the recognition process. Basically, the data used by the deceleration determination unit 10a and the stop determination unit 10b is the common data recognized by the object recognition unit 13 or the data obtained by performing some processing on the common data and the data acquired at the same time. Become.

  Both the deceleration determination unit 10a and the stop determination unit 10b determine the risk of collision with an object based on the recognition result of the object recognition unit 13, for example, the size, speed, or position of the recognized object. However, the thresholds of the determination are different between the two, and the determination by the stop determination unit 10b uses a higher risk value as the threshold than the determination by the deceleration determination unit 10a.

  When the deceleration determination unit 10a determines that deceleration is necessary, the control unit controls to reduce the target speed of the moving body 1, while the stop determination unit 10b needs to stop the vehicle. If it is determined that the moving body 1 is stopped (the target speed is set to zero). Here, the control for lowering the target speed may be control for lowering the target speed by a predetermined speed or control for lowering the target speed to a lower speed than the current speed.

  Hereinafter, it will be described that the control for setting such a target speed is performed by the speed instruction unit 10c in the control unit. The speed instruction unit 10c may instruct the drive control unit 11 about the target speed, and the drive control unit 11 may control the drive unit 12 to reach the target speed instructed by the speed instruction unit 10c. For example, the target speed may be controlled to reach the target speed in a certain period, or the target speed may be controlled to reach the target speed with a constant acceleration.

  However, it is not possible to achieve both functionality and safety with the above-mentioned control alone. For example, by dividing the measurement space region of the object recognition unit 13 into a partial region for stop determination on the front side and a partial region for deceleration determination on the back side, the functionality is improved by not stopping as much as possible, but the deceleration determination unit If the same algorithm is used in 10a and the stop determination unit 10b, the safety will be inferior such that the speed is reduced as much as possible.

  Therefore, the deceleration determination unit 10a in the present embodiment makes the determination in a shorter processing time than the stop determination unit 10b. That is, the processing in the deceleration determination unit 10a and the stop determination unit 10b is defined such that the processing time required for the determination by the deceleration determination unit 10a is shorter than the processing time required for the determination by the stop determination unit 10b.

  These determination processes are briefly mentioned, for example, (a) a process of determining that deceleration / stop is necessary when the size of the recognized object is equal to or larger than a predetermined size, or (b) ) Processing for determining that deceleration / stop is necessary when the size of the recognized object increases with a change amount of a predetermined speed or more, or (c) a predetermined area near the center of the detectable range of the object recognition unit 13. There is a process of determining that deceleration / stop is necessary when an object is detected at a position within the detection range. Regarding the above (a), when the object recognition unit 13 that cannot obtain the distance is adopted, the predetermined size may be determined in consideration of the size of another assumed moving body. Of course, with respect to the above (a) to (c), the thresholds are made different in the determination of deceleration and the determination of stop as described above. Further, in the case where only the stop determination unit 10b refers to the past data, if it is determined that the object detected laterally is not moving, the stop determination unit 10b does not stop the object, and there is a very high possibility of collision when traveling as it is. You may make it perform complicated processing, such as stopping only at. Further, for example, the deceleration determination unit 10a may execute only one of the above (a) to (c), and the stop determination unit 10b may execute two or more of them to make the processing time different.

  Then, the control unit in the present embodiment first causes the deceleration determination unit 10a to perform the determination, and when it is determined that the deceleration determination unit 10a needs to reduce the speed, the speed instruction unit 10c causes the drive unit 12 to perform the deceleration. At the same time as giving an instruction, the stop determination unit 10b executes the determination. When it is determined by the deceleration determination unit 10a that there is no need to decelerate, control may be performed such that the current speed is maintained or the speed is set to match the planned travel route.

  If the stop determination unit 10b determines that the control unit needs to stop, the speed instruction unit 10c instructs the drive unit 12 to stop, and the stop determination unit 10b determines that it is not necessary to stop. In this case, the deceleration state of the moving body 1 is maintained. When maintaining the speed, the speed instruction unit 10c may instruct the same target speed, but the instruction itself may not be performed.

  An example of such processing will be described with reference to FIG. FIG. 2 is a flow chart for explaining an example of processing in the mobile unit 1.

  The control unit first causes the deceleration determination unit 10a to perform a determination as a safety confirmation process (step S1), and confirms whether or not the result indicates safety (whether or not deceleration is necessary) ( If YES in step S2), the speed instruction unit 10c outputs a high-speed traveling instruction to the drive control unit 11 (step S3), and the process ends. When the high-speed traveling instruction is output and the vehicle has already traveled at high speed, the speed is maintained.

  On the other hand, if NO in step S2, the speed instruction unit 10c outputs a deceleration (slow) traveling instruction to the drive control unit 11 (step S4), and then the control unit determines to the stop determination unit 10b as a stop confirmation process. Is executed (step S5), and it is confirmed whether or not the result indicates that the vehicle should stop (step S6). If YES, the speed instruction unit 10c drives the stop (zero speed) instruction. The data is output to the unit 11 (step S7), and the process ends. If NO in step S6, the process ends. By executing the above-described processing at predetermined intervals, for example, even when the vehicle is stopped or decelerated, high speed traveling (or deceleration traveling) can be performed during the next processing.

  As described above, in the present embodiment, the determination result of the safety confirmation process is immediately output, and after that, the stop confirmation process is performed only when the stop is required, so that the deceleration travel is started as soon as possible. Then, by shifting to deceleration traveling first, it is possible to suppress the braking distance after it is suddenly determined that the stop is necessary. For example, if the safety confirmation process takes 3 ms for the simple process and the stop confirmation process takes 10 ms for the complicated process, it takes 3 ms for the high-speed travel output, 3 ms for the deceleration output, and 13 ms for the stop output.

  As described above, according to the moving body of the present embodiment, the necessity of deceleration is first determined based on the object recognition result, the deceleration is performed if necessary, and then the necessity of stop is determined based on the object recognition result. Therefore, both safety and functionality can be achieved. Specifically, by facilitating deceleration, it is possible to reduce the risk of causing harm in the event of a collision, and by making it difficult to stop, the side surface of the moving body 1 when traveling on a narrow road, for example, It is possible to prevent the wall of the vehicle from being detected and stopping and being stuck on the spot. Furthermore, it is possible to suppress the amount of recovery work required by a person when stopped, and to reduce the security risk that the security robot as an example of the mobile unit 1 cannot perform the expected security. You can also Furthermore, if the moving body 1 is autonomous, it can be controlled so that it will return to a predetermined point if it does not stop, so it is possible to prevent the battery from going up while it is stopped, so that a person can go to that point. It is not necessary to replace the battery or move the moving body 1, and the amount of work can be suppressed.

  Moreover, although the example in which the moving body 1 is an autonomous traveling device has been described above, it may be compatible with driving by a person. In that case, the mobile unit 1 includes an operation unit that receives a driving operation by the driver, and a navigation unit that performs navigation on the map indicated by the map information 14a based on the current position acquired by the position information acquisition unit 15. Just be prepared. The operation unit includes a handle, an accelerator, a brake, and the like.

  The navigation unit displays the map indicated by the map information 14a and, for example, heads for the destination registered in advance by the driver or passenger (that is, the planned travel route to the destination registered in the map information 14a). You can guide by voice and route display. Therefore, the navigation unit includes an image display unit and / or a voice output unit. In addition, even when these operation parts and the like are provided, an autonomous traveling function may be provided. In that case, it is necessary to supplement the autonomous traveling with the driving of the driver or to supplement the driving deficiency of the driver with the autonomous traveling. Good.

(Second embodiment)
A second embodiment of the present invention will be described with reference to FIG. FIG. 3 is a flow chart for explaining an example of processing in the moving body according to the present embodiment. In the present embodiment, the description of the overlapping parts with the first embodiment is basically omitted, but various application examples described in the first embodiment can be applied.

  In the present embodiment, the stop determination unit 10b performs determination based on data obtained by performing noise removal processing on data (hereinafter referred to as raw data) obtained as a result of the recognition processing by the object recognition unit 13, and the deceleration determination unit 10a The determination is performed based on the raw data or data obtained by subjecting the raw data to noise removal processing that is simpler than the noise removal processing. Note that the noise removal processing may actually be performed by the object recognition unit 13, in which case the deceleration determination unit 10a uses the intermediate data to make the determination, and the stop determination unit 10b uses the final data. You may make a decision by It should be noted that the intermediate data here refers to the data of the result of detection of an object by the object recognition unit 13, and, for example, when a distance measuring device such as LIDAR is adopted as the object recognition unit 13, distance measurement data and a camera are adopted. When it does, it refers to shooting data.

  An example of such processing will be described with reference to FIG. First, as a safety confirmation process, the control unit causes the deceleration determination unit 10a to perform an obstacle determination process for determining an object (obstacle) from raw data (step S11). Does the result indicate that there is an obstacle? Whether or not (deceleration is necessary) is confirmed (step S12), and if NO, the speed instruction unit 10c outputs a high-speed traveling instruction to the drive control unit 11 (step S13) as in step S3, and the processing is performed. To finish.

  On the other hand, if YES in step S12, the speed instruction unit 10c outputs a deceleration (slow) traveling instruction to the drive control unit 11 as in step S4 (step S14). Next, as the stop confirmation process, the control unit instructs the stop determination unit 10b to perform the noise removal process on the raw data, causes the noise removal process to be executed (step S15), and uses the data after the noise removal process for obstacles. It is determined whether or not there is (step S16). Then, the control unit confirms whether or not the result indicates that there is an obstacle (that is, that the vehicle should stop) (step S17), and if YES, the speed instruction unit 10c determines that the speed instruction unit 10c is the same as step S7. A stop (zero speed) instruction is output to the drive control unit 11 (step S18), and the process ends. If NO in step S17, the process ends. By executing the above-described processing at predetermined intervals, for example, it is possible to perform high-speed traveling (or deceleration traveling) during the next processing even in a stopped or decelerated state.

  In the present embodiment, as described above, first, the noise component included in the raw data such as the distance information (for example, raindrops during rain or snow during snowfall) is not removed (or the removal level is reduced to almost remove). If the obstacle determination process is performed by setting the threshold for determining that there is an object loosely, and performing the obstacle determination process, the speed can be reduced in a suspicious situation. As a result, it is possible to reduce the risk of missing an actually existing object due to erroneous noise removal or the like in the obstacle determination process, and to decelerate to a speed that does not cause harm even if a collision occurs. Then, the stop determination unit 10b performs noise removal processing as compared with the deceleration determination unit 10a or performs noise removal processing with an increased noise removal level to determine whether or not there is an obstacle to be surely stopped. It is possible to prevent accidental stop due to noise.

  Note that raindrops, snow, and the like can be detected by using an object recognizing unit 13 having a strong directivity such as 3D-LIDAR, and as a method of removing it as noise, it is possible to detect the object as described in (a) above. There are various things such as filtering by size.

(Third Embodiment)
A third embodiment of the present invention will be described with reference to FIGS. 4 to 5B as well. FIG. 4 is a flow chart for explaining an example of processing in the moving body according to the present embodiment, and FIGS. 5A and 5B are schematic diagrams for explaining an example of ground detection processing in the processing example of FIG. is there. Note that, in the present embodiment, the description of the overlapping portion with the first embodiment is basically omitted, but various application examples described in the first embodiment can be applied, for example.

  In the present embodiment, the stop determination unit 10b makes a determination based on data obtained by correcting the ground recognition data (raw data) obtained by the object recognition unit 13, and the deceleration determination unit 10a uses the raw data. Is determined based on the data that is not corrected. Note that the above-described correction may be actually performed by the object recognition unit 13 as in the noise removal process. In that case, the deceleration determination unit 10a uses the intermediate data to perform determination, and the stop determination unit 10b performs determination. The determination may be made using the final data. It should be noted that the stop determination unit 10b (and the deceleration determination unit 10a) may be subjected in advance to the noise removal processing as described in the second embodiment.

  An example of such processing will be described with reference to FIG. First, the control unit executes the processes of steps S11 to S14 of FIG. 3 (steps S21 to S24). Next to step S24), as the stop confirmation process, the control unit instructs the stop determination unit 10b to perform ground exclusion processing such as detecting (estimating) the ground from the raw data and correcting the height of the object as necessary. , Execute it (step S25). Then, the said control part performs the process similar to step S16-S18 with respect to the data after a ground exclusion process (step S26-S28).

  An example of the ground exclusion process will be described with reference to FIGS. 5A and 5B, taking an example of detection in 3D-LIDAR as the object recognition unit 13. For example, as illustrated in FIG. 5A when the actual obstacle D exists, when the depression G exists in the ground G or when the inclination of the ground G changes, the ground G is an object in the 3D-LIDAR. Will be detected as. The detection result on the surface where the obstacle D exists is illustrated in FIG. 5B. In FIG. 5B, it can be seen that a uniform object has been detected in a region where the angle β is 10 degrees or more.

  Whether or not it is the ground is determined by referring to the height component of the data and whether or not there is a uniform height in the horizontal or diagonal direction. It may be performed depending on whether or not an object is detected. Preferably, the moving body 1 is provided with an acceleration sensor or the like, and it is determined that the obstacle is the ground while the downward acceleration is detected and when there is such a uniform lower obstacle. However, the correction may be performed so as to exclude the lower part. Of course, the ground exclusion process is not limited to this example, and various examples can be given, and various sensors may be mounted on the moving body 1 as necessary.

  In the present embodiment, as described above, without first removing the ground from the raw data such as the distance information, and by setting the threshold for determining that there is an object loosely, by performing the obstacle determination process, It can slow down in suspicious situations. As a result, it is possible to reduce the risk of missing an actually existing object due to erroneous ground detection processing in the obstacle determination processing, and to decelerate to a speed that does not cause harm even if a collision occurs. Then, the stop determination unit 10b performs ground detection processing as compared with the deceleration determination unit 10a to determine whether or not there is an obstacle to be surely stopped, so that it is possible to prevent accidental stop due to unnecessary ground data. can do.

  Further, as another processing example regarding the ground, the deceleration determination unit 10a may also perform the determination based on data obtained by correcting the raw data as a result of the recognition processing by the object recognition unit 13 so as to exclude the ground.

  In that case, the stop determination unit 10b may exclude the ground from the raw data, further estimate the gradient amount of the ground, and make a determination based on the data obtained by correcting the height of the ground. By estimating the gradient amount of the ground (obtaining the estimated road surface straight line), it is possible to prevent a situation in which the inclination of the ground is erroneously calculated due to an obstacle existing on the slope. If the height of the ground is corrected by the estimated amount of gradient, the existing obstacle can be detected.

  In this processing example, as described above, first, the ground is removed from the raw data such as distance information, and the threshold value for determining that there is an object is also set loosely to perform the obstacle determination process, so that a small step The speed can be reduced if unevenness is present or in a suspicious situation. This reduces the risk of missing a real object even in an environment where the height of the stone, such as a stone rolling forward, is unstable due to incorrect ground correction processing in the obstacle determination processing that involves the estimation of the amount of slope, and Even if the speed is not harmed, the speed can be reduced. Then, the stop determination unit 10b performs more accurate ground detection processing than the deceleration determination unit 10a to determine whether there is an obstacle that should be surely stopped. Therefore, it is possible to avoid a situation where an object on a slope does not stop. be able to. However, in situations such as when the gradient amount is too large or the height does not change uniformly, the stop determination unit 10b may not be able to calculate the gradient amount (inclination amount) correctly. It is determined that 10b should not be stopped and the deceleration state is maintained.

(Fourth Embodiment)
A fourth embodiment of the present invention will be described with reference to FIGS. 6A and 6B together. 6A and 6B are schematic diagrams for explaining an example of processing in the moving body according to the present embodiment. Note that, in the present embodiment, the description of the overlapping portion with the first embodiment is basically omitted, but various application examples described in the first embodiment can be applied, for example.

  In the present embodiment, the deceleration determination unit 10a makes a determination based on data obtained by performing thinning processing on the data (raw data) obtained as a result of the recognition processing, or on the basis of data obtained by executing the recognition processing with a lower recognition resolution. To do. On the other hand, the stop determination unit 10b makes a determination based on the data of the result of performing the recognition process without performing the thinning process or the process of lowering the recognition resolution.

  Note that the thinning processing and the processing for reducing the recognition resolution may be actually performed by the object recognition unit 13 as in the noise removal processing. In that case, the deceleration determination unit 10a uses the intermediate data to make the determination. The stop determination unit 10b may make the determination using the final data. Note that the stop determination unit 10b (and the deceleration determination unit 10a) further performs the noise removal processing as described in the second embodiment and / or the ground exclusion processing and the gradient amount estimation processing described in the third embodiment. May be.

  As an example of the case where 3D-LIDAR is adopted as the object recognition unit 13, as shown in FIG. 6A, for the measurement points in a cross section perpendicular to a certain traveling direction, for example, the data of the measurement points shown as thinned-out data 62. Is thinned out, and the rest is set as valid data 61. Taking the case where the 2D-LIDAR 13a is adopted as the object recognition unit 13, for example, as shown in FIG. 6B, with respect to the survey line in a horizontal section in a certain traveling direction, for example, the data of the survey line shown as the thinning data 64 is thinned out. , And the rest is valid data 63. Although the thinning-out processing has been described, lowering the recognition resolution and performing the recognition processing means that only the necessary data is extracted and the recognition processing is performed based on the data as in the thinning-out processing.

  In the present embodiment, as described above, first, the information amount of raw data such as distance information is reduced, and the threshold value for determining that there is an object is set loosely to perform the obstacle determination process, thereby making a suspicious situation. Then you can slow down. As a result, it is possible to reduce the risk of considering even a small object that is difficult to actually collide in the obstacle determination process, and to decelerate to a speed that does not cause harm even if the collision occurs. Then, the stop determination unit 10b determines whether or not there is an obstacle to be surely stopped from the data having a larger amount of information than the deceleration determination unit 10a, so that the stop can be surely performed when necessary.

(Other)
In each of the above-described embodiments, after the safety confirmation processing, the vehicle speed is immediately reduced when it is not safe, but in reality, other processing examples are possible. An example of such processing will be described with reference to FIG. 7. FIG. 7 is a flow chart for explaining an example of processing in the moving body related to the present invention.

  The control unit configured to correspond to this processing example first causes the deceleration determination unit 10a to perform the determination as the safety confirmation process (step S31) as in steps S1 to S3, and confirms that the result is safe. It is confirmed whether or not it is shown (step S32), and if YES, the speed instruction unit 10c outputs a high-speed traveling instruction to the drive control unit 11 (step S33), and the process ends.

  On the other hand, if NO in step S32, does the control unit cause the stop determination unit 10b to perform the determination as the stop confirmation process as in step S5 (step S34), and the result indicates that the stop should be performed. Whether or not it is confirmed (step S35), and if YES, the speed instruction unit 10c outputs a stop (speed zero) instruction to the drive control unit 11 (step S37), and the process ends. In the case of NO in step S35, in the control unit, the speed instruction unit 10c outputs a deceleration (slow) traveling instruction to the drive control unit 11 (step S36), and the process ends. By executing the above-described processing at predetermined intervals, for example, even when the vehicle is stopped or decelerated, high speed traveling (or deceleration traveling) can be performed during the next processing.

  In this processing example, if the safety confirmation processing takes 3 ms for the simple processing and the stop confirmation processing takes 10 ms for the complicated processing, the high-speed traveling output is 3 ms, the deceleration output is 13 ms, and the stop output is 13 ms. It can be seen that the time until deceleration output is delayed compared to the example in the above embodiment. However, this processing example has an advantage that the moving body 1 can be easily designed and changed when the deceleration determination unit 10a and the stop determination unit 10b are configured by different chips or different substrates.

DESCRIPTION OF SYMBOLS 1 ... Moving body, 10 ... Main control part, 10a ... Deceleration determination part, 10b ... Stop determination part, 10c ... Speed instruction part, 11 ... Drive control part, 12 ... Drive part, 12a ... Wheel, 13 ... Object recognition part, 14 ... Storage unit, 14a ... Map information, 15 ... Position information acquisition unit, 16 ... Main body, 61, 63 ... Effective data, 62, 64 ... Thinning data.

Claims (6)

A moving body comprising: a drive unit, a control unit that controls driving by the drive unit, and an object recognition unit that performs recognition processing for recognizing an object,
Based on the result of the recognition process, a stop determination unit that determines whether or not the moving body needs to be stopped in order to avoid a collision with the recognized object, and based on the result of the recognition process, it is recognized. A deceleration determination unit that determines whether or not the moving body needs to be decelerated in order to avoid a collision with an object in a processing time shorter than the stop determination unit,
Wherein the control unit to execute the determination in the deceleration determination unit, if the is determined that deceleration determination unit it is necessary to decelerate, the have line instructs deceleration to the drive unit, then, by the stop determination unit If it is determined that the stop determination unit needs to stop, the drive unit is instructed to stop, and if the stop determination unit does not need to stop, the movement is performed. A mobile body characterized by maintaining a decelerated state of the body. The stop determination unit makes a determination based on data obtained by performing noise removal processing on the result data of the recognition processing,
The deceleration determination unit makes a determination based on data obtained as a result of the recognition processing or data obtained by performing noise removal processing that is simpler than the noise removal processing on the data obtained as a result of the recognition processing. The described moving body. The stop determination unit makes a determination based on the corrected data that excludes the ground from the result data of the recognition processing,
The moving body according to claim 1, wherein the deceleration determination unit makes a determination based on data that is not corrected. The deceleration determination unit makes a determination based on the corrected data that excludes the ground from the data of the recognition process,
The stop determination unit excludes the ground from the data of the result of the recognition process, further estimates the gradient amount of the ground, and makes a determination based on the data obtained by correcting the height of the ground. The mobile object according to 1. The deceleration determination unit makes a determination based on data obtained by performing thinning processing on the result data of the recognition processing, or based on data obtained by performing the recognition processing with a lower recognition resolution,
The moving body according to claim 1, wherein the stop determination unit makes a determination based on data obtained by performing the recognition process without performing the thinning process or the process of reducing the recognition resolution. A storage unit that stores map information including the planned travel route of the mobile unit; and a position information acquisition unit that acquires position information indicating the position of the mobile unit,
The control unit controls the drive unit to perform autonomous traveling along the planned travel route based on the current position acquired by the position information acquisition unit. The moving body according to any one of items.

Images