JP6192958B2 - Mobile environment map generation control device, mobile body, and mobile environment map generation method - Google Patents

Description

  The present invention relates to a technique for recognizing an obstacle around a moving body and generating an environment map.

  In order for autonomous mobile bodies such as unmanned vehicles and robots to autonomously move, it is necessary to recognize road surface conditions and environments around the autonomous mobile body and avoid obstacles in the traveling direction. In recent years, in addition to autonomous mobile objects, remote control or vehicles or robots that the driver rides on their own can also detect the surrounding environment of the vehicle, etc. Driving systems have been developed that display warnings to avoid collisions with objects or automatically control the brakes.

  For example, a laser range finder (hereinafter referred to as LRF) is provided in such various moving bodies as a sensor for detecting the surrounding environment of the moving body. The moving body detects the road surface condition, obstacles, and the like by scanning the surroundings using the LRF while moving and performing distance measurement, and generates a surrounding environment map based on the detection data. And the said mobile body determines a path | route based on the said environmental map.

  Obstacles include stationary obstacles such as buildings and moving obstacles such as pedestrians. Conventionally, areas once detected as obstacles are areas that cannot be run until the same area is scanned again. It continued to be recognized as. For this reason, even if the obstacle is a moving obstacle and actually moves from that area, it has been recognized as an area where it cannot travel unless the same area is scanned again. As described above, an afterimage of the moving obstacle remains on the environment map, which causes a problem that the movable range of the moving body is narrowed more than necessary.

  On the other hand, a method has been developed to determine whether or not an obstacle is a moving obstacle from the distance measurement data of the LRF, and to erase the data of the area where the moving obstacle exists every predetermined time to generate an environmental map. (See Patent Document 1).

JP 2011-150473 A

  However, the installation conditions of the LRF are generally determined according to the maximum speed of the moving body, and it is necessary to recognize the environment farther away as the maximum speed increases. When the LRF scans the road surface from the front part of the vehicle as a moving body as in Patent Document 1, the surrounding environment can be detected without any problem if it moves at a relatively low speed of less than 30 km / h. If the speed is higher than that, the resolution is insufficient and it is difficult to generate an accurate environmental map. In response to this, for example, if a sensor having a large number of scan lines is mounted or the number of sensors is greatly increased, the lack of resolution can be compensated, but a high-performance sensor is expensive, and the sensor If the number is increased, the number of parts is also increased accordingly. Further, since the amount of data becomes enormous, it becomes necessary to install a computer having high processing performance, and there is a problem that the cost is greatly increased.

  The present invention has been made to solve such a problem. The object of the present invention is to discriminate between moving obstacles and stationary obstacles even when moving at a relatively high speed while suppressing an increase in cost. Another object of the present invention is to provide a mobile environment map generation control device, a mobile body, and a mobile environment map generation method capable of generating an accurate environmental map without unnecessarily narrowing a movable region.

In order to achieve the above-described object, the invention according to claim 1 is a mobile environment map generation control device for generating an environment map used for movement control of a mobile body, and performs distance measurement from the mobile body toward a road surface. A first distance measuring unit that performs measurement, a second distance measuring unit that measures a distance from the moving body in a traveling direction, a first distance measurement information acquired by the first distance measuring unit, and a second distance measuring unit; Whether the obstacle detected by the first distance measurement means and the second distance measurement means is a stationary obstacle or a moving obstacle based on the transition of the second distance measurement information acquired by the distance measurement means. The obstacle identified by the obstacle identifying means and the area identified as having a stationary obstacle by the obstacle identifying means continue to be recognized as a non-movable area, and the area identified as having the moving obstacle is Accreditation after certifying that the area is not movable for a specified period of time Of by returning to the original certification, and a environmental map generation unit that generates the environmental map, the obstacle identifying means, on the environmental map an obstacle detected by the second distance measuring means Whether the obstacle detected by the first distance measuring means and the second distance measuring means is a stationary obstacle by performing obstacle identification processing including the grid m adjacent to the grid i which is the position, or a moving obstacle It is characterized that you identify with either ones.

  According to a second aspect of the present invention, the obstacle identification unit detects the obstacle detected by the first distance measuring unit and the second distance measuring unit. When the obstacle is in the same area on the environmental map, the obstacle in the area is moved and obstructed for a predetermined time set shorter than the predetermined period from the detection of the obstacle by the second distance measuring means. An area that is identified as an object and cannot travel because there is a stationary obstacle after the lapse of the predetermined period when there is an obstacle in the area for the predetermined time or more from the detection of the obstacle by the second distance measuring means. It is characterized by certification.

In the environment map generation control apparatus for a mobile body according to claim 3, in claim 1 or 2, before Symbol first distance measuring means, a laser measuring a distance by irradiating a laser toward the road surface from the mobile It is a range finder, and the second distance measuring means is a laser range finder that performs distance measurement by irradiating a laser in a traveling direction from the moving body.

A moving body according to a fourth aspect includes the environmental map generation control device according to any one of the first to third aspects, generates a movement route based on the generated environmental map, and moves along the movement route. It is characterized by comprising a movement control means for performing.
According to a fifth aspect of the present invention, in the environmental map generation method for a mobile body that generates an environmental map used for movement control of the mobile body, the first distance measurement information is obtained by performing distance measurement from the mobile body toward the road surface. Acquiring the second distance measurement information by measuring the distance from the moving body in the traveling direction, and changing the first state of the area on the environment map based on the transition of the certified state of the area. The obstacle identification step for identifying whether the obstacle detected from the distance measurement information and the second distance measurement information is a stationary obstacle or a moving obstacle, and the obstacle identification by the obstacle identification step The area identified as existing continues to be recognized as a non-movable area, and the area identified as having the moving obstacle is determined as a non-movable area for a predetermined period of time, and then before the approval. Return to the original certification, the environment Comprising a environmental map generation step of generating a figure, the obstacle identification step, also grid m adjacent to the grid i is the position on the environmental map an obstacle detected by the second distance measurement information It characterized that you identify which performs obstacle identification process, which is the moving obstacle or obstacles detected by the first distance measurement information and the second distance measurement information is a stationary obstacle, including It is said.

The environmental map generation method for a moving body according to claim 6 is the method according to claim 5 , wherein the obstacle identifying step is detected from the obstacle detected from the first distance measurement information and the second distance measurement information. When the obstacle is in the same area on the environment map, the obstacle in the area is moved for a predetermined time set shorter than the predetermined period from the detection of the obstacle by the second distance measurement information. An area that is identified as an obstacle and cannot travel due to the presence of a stationary obstacle after the lapse of the predetermined period when there is an obstacle in the area for the predetermined time or more from the detection of the obstacle by the second distance measurement information It is characterized by accrediting.

  According to the environmental map generation control apparatus, the mobile object, and the mobile environment map generation method according to the present invention using the above means, in addition to the first distance measuring means for detecting the road surface state, the movement In order to identify the obstacle, the second distance measuring means is provided. Since the second distance measuring means measures the distance in the moving direction of the moving body, the obstacle can be reliably detected even at a relatively high speed.

  Then, moving obstacles are identified based on the transition of the certified state of the area on the environmental map, and the area identified as the moving obstacle is recognized as an area that cannot be moved for a predetermined time, and then the original before the certification. Returning to the certification of, will be to generate an environmental map. Accordingly, it is possible to suppress the afterimage of the moving obstacle from remaining on the environment map without detecting the same region again, and it is possible to prevent the movable range of the moving body from being narrowed.

  This makes it possible to identify and move moving obstacles and stationary obstacles even when moving at relatively high speeds while suppressing an increase in cost, without providing a large number of ranging means or requiring high performance. It is possible to generate an accurate environmental map that does not unnecessarily narrow a region.

It is a schematic block diagram of the unmanned vehicle provided with the environmental map production | generation control apparatus which concerns on one Embodiment of this invention. It is a block diagram which shows the control structure of a computer unit. It is a flowchart which shows the 1st LRF data process performed in a LRF data processing part. It is a flowchart which shows the 2nd LRF data process performed in a LRF data processing part. It is a flowchart which shows the environmental map production | generation process performed in an environmental map production | generation part. 6 is a flowchart showing a continuation of FIG. It is a flowchart which shows the reflection process of the 1st data D1j performed by step S34 of FIG. It is a flowchart which shows the reflection process of the 2nd data D2j performed by step S38 of FIG. It is a flowchart which shows the position shift compensation process performed by step S71 of FIG. It is explanatory drawing about a position shift compensation process. It is a flowchart which shows the movement obstruction time-out process performed in step S41 of FIG.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic configuration diagram of an unmanned vehicle including an environmental map generation control device according to an embodiment of the present invention, and FIG. 2 is a block diagram showing a control configuration of a computer unit mounted on the unmanned vehicle. Hereinafter, the configuration of the moving body will be described with reference to FIG.
As shown in FIG. 1, an unmanned ground vehicle (UGV) 1 is an autonomous moving body that travels in accordance with an instruction from a remote control device (not shown).

The unmanned vehicle 1 is equipped with a GPS (Global Positioning System) receiver 2 for measuring its own position and a gyro sensor 4 for measuring its own attitude. The unmanned vehicle 1 is provided with a steering 6 for steering the vehicle and an accelerator 8 for operating the driving force of the vehicle, and actuators 6a and 8a capable of adjusting the operation amount are provided respectively.
Further, in the unmanned vehicle 1, a first LRF 10 (first distance measuring means) is fixed to the tip portion of the roof portion. The first LRF 10 is an LRF that mainly detects a road surface state by irradiating a laser from a roof end to a road surface in front of the vehicle to perform distance measurement. The first LRF 10 is a four-line LRF provided with four lasers that scan in the vehicle width direction in four ranges arranged in the traveling direction on the road surface in front of the vehicle. In FIG. 1, only the laser irradiation range for one line is shown.

  Further, the unmanned vehicle 1 has a second LRF 12 (second distance measuring means) fixed to the front portion of the vehicle body. The second LRF 12 is an LRF that mainly detects a moving obstacle by performing horizontal ranging by irradiating a laser in the traveling direction from the front of the vehicle body. The second LRF 12 is a four-line LRF provided with four lasers that scan in the vehicle width direction in four ranges aligned in the vertical direction with respect to the space ahead of the vehicle. For the second LRF 12 as well, only the laser irradiation range for one line is shown in FIG.

The unmanned vehicle 1 is equipped with a computer unit 20 for controlling autonomous traveling. The computer unit 20 receives an instruction from a remote control device (not shown), and the actuator is based on information from various sensors. 6a and 8a are operated to drive the unmanned vehicle 1.
FIG. 2 is a block diagram showing a control configuration in the computer unit 20, and the configuration of the computer unit 20 will be described with reference to FIG.

As shown in FIG. 2, in the computer unit 20, various sensors and various control units are connected via a communication line such as a LAN. The computer unit 20 roughly includes an environment recognition / self-position measuring unit 22 and a vehicle control unit 24.
The environment recognition / self-position measurement unit 22 includes a self-position evaluation unit 26, an LRF data processing unit 28, and an environment map generation unit 30.

The self-position evaluation unit 26 has a function of evaluating the self-position based on self-position information from the GPS receiver 2 and self-posture information from the gyro sensor 4.
The LRF data processing unit 28 has a function of processing distance measurement information from the first LRF 10 and the second LRF 12.
The information processed by the self-position assessment unit 26 and the LRF data processing unit 28 is sent to the environment map generation unit 30. The environment map generation unit 30 is based on the acquired information, and the environment map around the unmanned vehicle 1 It has the function to generate.

On the other hand, the vehicle control unit 24 includes a route generation unit 32 and a vehicle operation unit 34 (movement control means).
The route generation unit 32 acquires the self-location information evaluated by the self-position evaluation unit 26 of the environment recognition / self-position measurement unit 22 and the environment map information generated by the environment map generation unit 30, and a remote control device (not shown) A function for generating a route according to the instruction is provided.
Further, the vehicle operation unit 34 has a function of calculating a steering amount and a driving force necessary for traveling on the route generated by the route generation unit 32. The vehicle operation unit 34 controls the corresponding actuators 6a and 8a to perform the steering operation and the accelerator operation according to the calculated steering amount and driving force.

  Here, the LRF data processing unit 28 of the environment recognition / self-position measuring unit 22 detects obstacles by the first LRF 10 and the second LRF 12 respectively, and the environment map generation unit 30 determines whether the obstacle is a stationary obstacle. While identifying whether it is a moving obstacle (obstacle identification means, obstacle identification step), the area that can travel and the area that cannot travel are recognized to generate an environment map (environment map generation means, Environmental map generation step). Hereinafter, obstacle detection detection control and environment map generation control executed by the LRF data processing unit 28 and the environment map generation unit 30 of the environment recognition / self-position measurement unit 22 will be described.

(First LRF data processing)
First, the obstacle detection control executed in the LRF data processing unit 28 of the environment recognition / self-position measuring unit 22 will be described.
Referring now to FIG. 3, there is shown a flowchart regarding the flow of the first LRF data processing executed in the LRF data processing unit, and the first LRF data processing will be described below along the same flowchart.

  As shown in FIG. 3, first, the LRF data processing unit 28, as step S <b> 10, the attitude of the unmanned vehicle 1 in which information at a predetermined number (N) of measurement points measured by the first LRF 10 is detected by the gyro sensor 4. Based on the corners, the coordinates are converted to coordinate information that matches the generation of the environmental map. For example, the information of the measurement point measured by the first LRF 10 includes a distance and an angle, and the coordinate of the measurement point information is converted into a Cartesian coordinate system including three orthogonal axes of x, y, and z.

From the following step S11, the LRF data processing unit 28 sequentially repeats the following control for each of the N measurement points j from 1 to N after coordinate conversion.
In step S12, the LRF data processing unit 28 calculates a difference in height (z component) between adjacent measurement points, and determines whether the difference is larger than a predetermined difference. Then, it is determined whether there is an obstacle. If the determination result is false (No), that is, if there is no difference from the adjacent points or the difference is small, the process proceeds to step S13.

In step S13, the LRF data processing unit 28 determines that the measurement point j detects a road surface rather than an obstacle, and the obstacle detection result value Aj indicates that the vehicle can travel without an obstacle ( Aj = travel possible).
On the other hand, if the determination result in step S12 is true (Yes), that is, if the difference from the adjacent point has changed significantly, the process proceeds to step S14.

In step S14, the LRF data processing unit 28 determines that the measurement point j has detected some obstacle, and the obstacle detection result value Aj indicates that there is an obstacle (Aj = obstacle). Enter.
After inputting the obstacle detection result value Aj in step S13 or S14, the LRF data processing unit 28 proceeds to S15. In step S15, the LRF data processing unit 28 returns to step S11 when the process has not been completed for N measurement points j, and performs the processes of steps S12, S13, and S14 for the next measurement point j. Do. On the other hand, when the above process is completed up to N, the process proceeds to the next step S16.

In step S16, the LRF data processing unit 28 registers the obstacle detection result value Aj at each measurement point j and the time stamp (ts) at that time as first data D1j (first distance measurement information).
In step S17, the LRF data processing unit 28 transmits the first data D1j to the environmental map generation unit 30, and returns the first LRF data processing.

  In this way, in the first LRF data processing, it is identified whether the road surface that can be traveled or the obstacle is detected based on the height component of the measurement point j measured by the first LRF 10. Then, the result is transmitted to the environment map generation unit 30.

(Second LRF data processing)
Next, referring to FIG. 4, there is shown a flowchart of the flow of the second LRF data processing executed in the LRF data processing unit 28, and the second LRF data processing will be described along the same flowchart.
As shown in FIG. 4, first, in step S20, the LRF data processing unit 28 detects information at a predetermined number (M) of measurement points measured by the second LRF 12 by the gyro sensor 4 as in step S10. Coordinate conversion to coordinate information adapted to the generation of the environmental map based on the attitude angle of the unmanned vehicle 1.

From the following step S21, the LRF data processing unit 28 sequentially repeats the following control for each of the M measurement points j from 1 to M after coordinate conversion.
In step S22, the LRF data processing unit 28 determines whether or not there is an obstacle by determining whether or not the height (z component) of the measurement point j is greater than a predetermined reference height. If the determination result is true (Yes), that is, if the height of the measurement point j is higher than a certain height, the process proceeds to step S23.

In step S23, the LRF data processing unit 28 determines that the measurement point j detects an obstacle, and inputs a value (Aj = obstacle) indicating the obstacle as the obstacle detection result value Aj. To do.
On the other hand, when the determination result of step S22 is false (No), that is, when the height position of the measurement point j is equal to or lower than the reference height, the obstacle detection result value Aj is not set as it is, and step S24 is performed. Proceed to The initial value of Aj is no obstacle.

In step S24, if up to M measurement points j have not been processed, the LRF data processing unit 28 returns to step S21 and performs the processes of steps S22 and S23 for the next measurement point j. On the other hand, when the above process is completed up to M, the process proceeds to the next step S25.
In step S25, the LRF data processing unit 28 registers the obstacle detection result value Aj and the time stamp (ts) at each measurement point j as second data D2j (second distance measurement information).

In step S26, the LRF data processing unit 28 transmits the second data D2j to the environment map generation unit 30, and returns the second LRF data processing.
As described above, in the second LRF data processing, when the measurement point j higher than the reference height is detected by the second LRF 12, it is regarded as an obstacle, and the result is transmitted to the environment map generation unit 30.

(Environmental map generation process)
Next, environment map generation control executed by the environment map generation unit 30 of the environment recognition / self-position measurement unit 22 will be described.
Here, referring to FIGS. 5 to 11, FIGS. 5 and 6 are flowcharts for the flow of the environment map generation process executed in the environment map generation unit 30, and FIGS. The flowchart about each control performed within a process and its explanatory drawing are shown, and it demonstrates along each flowchart below.

First, the overall flow of the environment map generation process will be described with reference to FIGS.
In step S30 illustrated in FIG. 5, the environment map generation unit 30 receives the first data D1j and the second data D2j from the LRF data processing unit 28.
In subsequent step S31, the environment map generation unit 30 acquires the self-position and azimuth data at the data acquisition time based on the time stamp of each data.

And from step S32, the environment map production | generation part 30 repeats the following control sequentially about N each 1st data D1j to 1-N.
In step S33, the environment map generation unit 30 calculates a position on the environment map (hereinafter referred to as grid i) corresponding to the first data D1j based on the information acquired in step S31.
In step S34, the environment map generation unit 30 performs a process of reflecting the first data D1j on the environment map. Although the specific content of the reflection process of the first data D1j will be described later with reference to FIG. 7, the environment map generation unit 30 in step S34 determines the travelable area and the obstacle area detected by the first LRF 10. Register to the environmental map.

In step S35, the environment map generating unit 30 returns to step S32 when the first data D1j has not been processed up to N, and performs the processes of steps S33 and S34 for the next first data D1j. . On the other hand, when the above process is completed up to N, the process proceeds to step S36 in FIG.
From step S36 in FIG. 6, the environment map generation unit 30 sequentially repeats the following control for each of the M second data D2j from 1 to M.

In step S37, the environment map generation unit 30 calculates a grid i on the environment map corresponding to the second data D2j based on the information acquired in step S31.
In step S38, the environment map generation unit 30 performs a process of reflecting the second data D2j on the environment map. The specific content of the reflection process of the second data D2j will be described later with reference to FIG. 8, but in step S38, the environment map generation unit 30 changes the authorized state of the grid corresponding to the second data D2j. Based on this, obstacles are registered. Also, it is identified whether the obstacle detected by the first LRF 10 and the second LRF 12 is a stationary obstacle or a moving obstacle.

In step S39, the environment map generation unit 30 returns to step S36 if the processing for up to M pieces of second data D2j has not been completed, and performs the processing of steps S37 and S38 for the next second data D2j. . On the other hand, when the above process is completed up to M, the process proceeds to step S40.
In step S <b> 40, the environment map generation unit 30 transmits the environment map information generated through the steps so far to the route generation unit 32.

  Thereafter, in step S41, the environment map generation unit 30 performs a moving obstacle time-out process, and returns the flowchart. The specific contents of the time-out process for the moving obstacle will be described later with reference to FIG. 10, but the environmental map generation unit 30 determines in advance the grid i that is recognized as having a moving obstacle in step S41. During the time-out period (a certain period), the vehicle is recognized as an inoperable region, and after the time-out period elapses, a process for returning to the original authorization state before the certification is performed.

In this way, the environment map generation unit 30 reflects the first data D1j and the second data D2j on the environment map, performs a timeout process for the moving obstacle, and generates an environment map.
Hereinafter, each process will be described in detail.

(First data D1j reflection process)
FIG. 7 shows a flowchart showing the process of reflecting the first data D1j executed in step S34 of FIG. 5, and will be described along the same flowchart.
First, in step S50 of FIG. 7, the environment map generation unit 30 determines whether or not a value indicating an obstacle is input as the obstacle detection result value Aj of the first data D1j. If the determination result is true (Yes), the process proceeds to the next step S51.

In step S51, the environment map generation unit 30 determines whether or not the state of the grid i corresponding to the first data D1j calculated in step S33 of FIG. 5 is a travelable state at this time. If the determination result is true (Yes), the process proceeds to the next step S52.
In step S52, the environment map generation unit 30 stores the state immediately before switching of the grid i. Since both the determination results of step S50 and step S51 are true (Yes), this indicates a state where an obstacle has been detected in the grid i where no obstacle has been detected so far. The state before the obstacle is detected is stored. Specifically, the time ts_i at this time is stored as the time stamp ts_i_ch immediately before switching (ts_i_ch = ts_i), and the fact that the vehicle is in the travelable state is stored as the state A_i_ch immediately before switching of the grid i (A_i_ch = travelable).

After step S52, or if it is already recognized in step S51 that there is an obstacle in the grid i and the determination result is false (No), the process proceeds to step S53.
In step S53, the environmental map generation unit 30 updates the state of the grid i. Specifically, the time stamp ts_i of the grid i is updated to the time ts_j at this time, and the state A_i of the grid i is updated to an obstacle present state (impossible to travel). Note that i in the time stamp is the time held in the grid, j is the time of the acquired LRF data, and the relationship is ts_i <ts_j.

On the other hand, if the obstacle detection result value Aj of the first data D1j is a value indicating that the vehicle can travel and the determination result is false (No) in step S50, the process proceeds to step S54. In step S54, the environment map generation unit 30 determines whether or not the state of the grid i is an obstacle at this time. If the determination result is false (No), the process proceeds to the next step S55.
In step S55, the environment map generation unit 30 updates the state of the grid i. Specifically, the time stamp ts_i of the grid i is updated to the time ts_j at this time, and the state A_i of the grid i is updated to a state where traveling is possible.

After step S53 or after step S55, or when the state of the grid i is an obstacle and the determination result is true (Yes) in step S54, the process ends, and step S35 in FIG. Return to.
As described above, in the reflection process of the first data D1j, when an obstacle is detected by the first LRF 10, the data immediately before the corresponding grid i is switched is stored, and the state of the grid i is changed to the obstacle. Update to a state where there is a thing, and make it an area where traveling is impossible.

(Second data D2j reflection process)
Next, FIG. 8 shows a flowchart showing the reflection process of the second data D2j executed in step S38 of FIG. 6, and will be described below with reference to the flowchart.
First, in step S60 of FIG. 8, the environment map generation unit 30 determines whether or not a value indicating an obstacle is input as the obstacle detection result value Aj of the second data D2j. If the determination result is false (No), the process ends, and the process returns to step S39 in FIG. On the other hand, if the determination result is true (Yes), the process proceeds to step S61.

  In step S61, the environment map generation unit 30 does not set the moving obstacle flag Fm indicating the identification result of the moving obstacle on the grid i corresponding to the second data D2j calculated in step S37 of FIG. 6 (Fm = False). Note that the moving obstacle flag is Fm = false which is not set in the initial state. If the determination result is true (Yes), that is, if the moving obstacle has not yet been identified in the grid i, the process proceeds to step S62.

In step S62, the environment map generation unit 30 sets the moving obstacle flag Fm of the grid i to true, and proceeds to the next step S63.
In step S63, the environment map generation unit 30 determines whether or not the state A_i of the grid i at this time point is an obstacle (A_i = obstacle). If the determination result is false (No), that is, if the state A_i of the grid i is in a travelable state or an unmeasured state (A_i = runnable / unmeasured), the process proceeds to step S64.

  In step S64, the environment map generation unit 30 stores the state immediately before switching of the grid i. That is, the determination result of the above steps S60 to S63 shows a state in which a moving obstacle has entered the grid i that was a travelable or unmeasured area. In step S64, the moving obstacle enters. Remember the previous state. Specifically, the time ts_i at this time is stored as the time stamp ts_i_ch immediately before switching (ts_i_ch = ts_i), and the fact that the vehicle is in a travelable state or an unmeasured state is stored as the state A_i_ch immediately before switching the grid i (A_i_ch = Can run / not measured).

After step S64, or when the moving obstacle flag Fm of the grid i is already true and the determination result is false (No) in step S61, the process proceeds to step S65.
In step S65, the environment map generation unit 30 updates the state of the grid i. Specifically, the time stamp ts_i of the grid i is updated to the time ts_j at this time, the state A_i of the grid i is updated to an obstacle (unrunnable) state, and the moving obstacle flag Fm is updated to true. To do.

On the other hand, when the state A_i of the grid i is an obstacle (A_i = obstacle) in the above step S63 and the determination result is true (Yes), the process proceeds to step S66.
In step S66, the environment map generation unit 30 acquires the grid i immediately before switching data stored at this time.
In the next step S67, the stored time stamp ts_i_ch immediately before switching and the difference time δt between the current time ts_i are calculated.

In step S68, the environment map generation unit 30 determines whether or not the difference time δt is longer than a predetermined time (for example, 0.1 s). If the predetermined time has not elapsed since the previous switching point (ts_i_ch), the determination result is false (No), and the process proceeds to step S69.
In step S69, since the environmental map generation unit 30 has not passed the predetermined time since the obstacle detection by the second LRF 12 since the previous grid i switching state data was stored, the obstacle is a moving obstacle. The data immediately before switching is retained as it is.

  On the other hand, if the determination result in step S68 is true (Yes), that is, if a predetermined time has elapsed since the last switching time of the grid i, the process proceeds to step S70. That is, in this case, since the obstacle is detected by the first LRF 10 and the second LRF 12 for a predetermined time or more, it is determined that the object is a stationary obstacle. Thereby, for example, when a moving obstacle and a stationary obstacle overlap in the same area, it is possible to continue to recognize that there is a stationary obstacle even after the moving obstacle has passed.

  In step S <b> 70, the environment map generation unit 30 updates the data immediately before the switching of the grid i. Specifically, the time stamp ts_i_ch immediately before switching is updated to the time ts_i at this time (ts_i_ch = ts_i), and the state A_i_ch immediately before switching of the grid i is set to an obstacle present state (A_i_ch = obstacle).

Note that the steps S66 to S70 substantially identify whether the obstacle is a moving obstacle or a stationary obstacle, and this process is hereinafter referred to as an obstacle identification process.
After passing through step S65, S69, or S70, after performing the positional deviation compensation process for compensating for the deviation between the obstacle detected by the first LRF 10 and the obstacle detected by the second LRF 12 in step S71, the second data D2j And the process returns to step S39 in FIG.

(Displacement compensation processing)
Here, the positional deviation compensation process executed in step S71 will be described in detail.
FIG. 9 shows a flowchart showing the misalignment compensation processing executed in step S71 of FIG. 8, and the following description will be made along the same flowchart.
In step S80, the environment map generation unit 30 sets a range of the grid m adjacent to the grid i corresponding to the second data D2j calculated in step S37 of FIG. The number L of the grids m may be, for example, four grids on the front, rear, left, and right sides of the grid i, or a range beyond that.

In step S81, the environment map generation unit 30 sequentially repeats the following control for each of the L grids m from 1 to L.
In step S82, the environmental map generation unit 30 determines whether or not the moving obstacle flag Fm of the grid m is false. If the determination result is true (Yes), the process proceeds to step S83.

In step S83, the environment map generation unit 30 determines whether the state A_m of the grid m is an obstacle. If the determination result is true (Yes), that is, if the obstacle is recognized in the grid m, the process proceeds to step S84.
In step S84, the environment map generation unit 30 sets the moving obstacle flag of the grid m to true.

In subsequent step S85, the environment map generation unit 30 performs the obstacle identification process including steps S66 to S70 in FIG. 8 by replacing the grid i with the grid m. After identifying the moving obstacle or the stationary obstacle by the obstacle identifying process, the process proceeds to step S86.
On the other hand, if the moving obstacle flag Fm of the grid m is not false in step S82 and the determination result is false (No), or the determination result is that the state A_m of the grid m is not an obstacle in step S83. Is false (No), the process of steps S84 and S85 is not performed, and the process proceeds to step S86. That is, when the moving obstacle flag Fm of the grid m is true, or when the state A_m of the grid m is travelable or not measured, the process proceeds to step S86 as it is.

In step S86, if the environment map generation unit 30 has not finished processing up to L grids m, the process returns to step S81 and performs the processes in steps S82 to S84 for the next grid m. On the other hand, when the above processing is completed up to L, the misregistration compensation processing ends.
For example, as shown in the explanatory diagram of FIG. 10, the positional deviation compensation processing is a case where a pedestrian crosses in front of the unmanned vehicle 1. The first LRF 10 detects the shoulder portion of the pedestrian, and the second LRF 12 If a foot part of a pedestrian is detected, the grid ia corresponding to the first data D1j may not match the grid ib corresponding to the second data D2j. Even in such a case, by performing the obstacle identification process for the grid m adjacent to the grid i corresponding to the second data D2j by the above-described misalignment compensation process, the grid ia can be reliably identified as a moving obstacle. can do.

(Moving obstacle timeout process)
Finally, the moving obstacle timeout process executed in step S41 of FIG. 6 will be described.
FIG. 11 shows a flowchart showing the moving obstacle time-out process, which will be described below.
First, in step S90, the environment map generation unit 30 sets a predetermined range as a timeout processing range on the environment map. The timeout processing range is set as a range in which the unmanned vehicle 1 can generate a route, and is set to about 50 m, for example.

Next, in step S91, the environment map generation unit 30 determines whether or not the moving obstacle flag Fm of the grid i in the timeout processing range is true. If the determination result is false (No), that is, if the moving obstacle flag Fm is false, the time-out process is not necessary, and the process ends.
On the other hand, if the determination result in step S91 is true (Yes), that is, if it is identified that there is a moving obstacle, the process proceeds to step S92.

In step S92, the environment map generation unit 30 calculates δt2 that is the difference between the current time t_now and the time stamp ts_i registered in the grid i.
In step S93, it is determined whether or not δt2 is greater than a predetermined timeout time (for example, 0.2 s). The timeout time is longer than the predetermined time in step S68 of FIG. If the determination result is false (No), that is, if the time-out period has not elapsed since Fm = true in grid i, the process ends as it is. On the other hand, if the determination result is true (Yes), the process proceeds to step S94.

In step S94, the environment map generation unit 30 acquires state data immediately before switching corresponding to the grid i, updates the state of the grid i as the state immediately before switching, and ends the moving obstacle timeout process. Specifically, the time stamp ts_i of the grid i is set to the current time t_now, the state A_i is set to A_i_ch, and the moving obstacle flag Fm is set to false.
That is, when the predetermined time-out period elapses after the moving obstacle flag Fm = true for the grid i, the state of the grid i is returned to the original state. If the grid i was originally an area where the vehicle can travel, the vehicle can travel again. If there is a stationary obstacle, the region where the vehicle cannot travel as an obstacle is maintained. As a result, it is possible to prevent the range in which the moving obstacle passes from remaining unmovable.

After performing the time-out process as described above, the environment map generation process shown in FIGS. 5 and 6 is returned, and the above process is repeated to update the environment map as the unmanned vehicle 1 travels. Go.
As described above, the unmanned vehicle 1 according to the present embodiment includes the second LRF 12 for identifying the moving obstacle in addition to the first LRF 10 for detecting the road surface state. Since the second LRF 12 measures the distance in the traveling direction of the unmanned vehicle 1, an obstacle can be reliably detected even at a relatively high speed.

  In particular, as shown in the obstacle identification process, whether the obstacle is a stationary obstacle or a moving obstacle is identified based on the certified state of the grid on the environmental map. In the grid i identified as the moving obstacle, the environment map is generated by recognizing the area incapable of traveling for a predetermined timeout time and then returning to the original state A_i before the recognition. Accordingly, it is possible to suppress the afterimage of the moving obstacle from remaining on the environment map without detecting the same region again, and it is possible to prevent the movable range of the moving body from being narrowed.

  Accordingly, since only two LRFs of the first LRF 10 and the second LRF 12 are used without providing a large number of LRFs or requiring high performance, an increase in cost can be suppressed. Even when the unmanned vehicle 1 moves at a relatively high speed, it is possible to identify a moving obstacle and a stationary obstacle and generate an accurate environmental map that does not unnecessarily narrow the movable area.

This is the end of the description of the embodiment according to the present invention, but the embodiment is not limited to the above embodiment.
In the above embodiment, the moving body is an unmanned vehicle, but the moving body is not limited to the vehicle, and may be, for example, a robot that performs autonomous walking.
In addition, the moving body is not limited to an autonomous moving body, and the present invention can be applied to a vehicle, a robot, or the like that is remotely operated or driven by a driver. For example, in a car on which a driver is boarded, the environment map generated by the present invention is used to recognize the surrounding environment, and driving assistance according to the surrounding environment, warning display for collision avoidance, automatic Or brake control.

  In the above-described embodiment, the 4-line LRF is used as each of the first LRF 10 and the second LRF 12, but the LRF is not limited to this.

1 Unmanned vehicle (moving body)
2 GPS receiver 4 Gyro sensor 6 Steering 8 Accelerator 6a, 8a Actuator 10 1st LRF (first ranging means)
12 2nd LRF (2nd ranging means)
20 Computer unit 22 Environment recognition / self-position measuring unit 24 Vehicle control unit (movement control means)
26 Self-location assessment unit 28 LRF data processing unit 30 Environmental map generation unit (obstacle identification means, environmental map generation means)
32 Route generation unit 34 Vehicle operation unit

Claims (6)

In an environment map generation control device for a mobile object that generates an environment map used for movement control of the mobile object,
First distance measuring means for measuring a distance from the moving body toward the road surface;
A second distance measuring means for measuring a distance from the moving body in a traveling direction;
Based on the transition of the certified state of the area on the environmental map, it is identified whether the obstacle detected by the first distance measuring means and the second distance measuring means is a stationary obstacle or a moving obstacle. Obstacle identification means;
The area identified as having a stationary obstacle by the obstacle identifying means continues to be recognized as a non-movable area, and the area identified as having a moving obstacle is immovable for a predetermined period of time. An environmental map generating means for generating the environmental map after returning to the original certification prior to the certification ,
The obstacle identification means performs obstacle identification processing including the grid m adjacent to the grid i which is the position on the environment map where the obstacle detected by the second distance measuring means is located, and distance measurement means and said second moving body of the environment map generation control unit obstacle detected by the distance measuring means, characterized that you identify whether the moving obstacle or a stationary obstacle. The obstacle identifying means, when the obstacle detected by the first distance measuring means and the obstacle detected by the second distance measuring means are in the same area on the environment map,
During the predetermined time set shorter than the predetermined period from the detection of the obstacle by the second distance measuring means, the obstacle in the area is identified as a moving obstacle,
When there is an obstacle in the area for more than the predetermined time from the detection of the obstacle by the second distance measuring means, it is determined that the vehicle cannot travel because there is a stationary obstacle after the fixed period has elapsed. The mobile environment map generation control device according to claim 1. The first distance measuring means is a laser range finder that performs distance measurement by irradiating a laser toward the road surface from the moving body,
The second distance measuring unit 130, the environment map according to claim 1 or 2 moving body, wherein the toward the traveling direction from said moving body is a laser range finder measuring a distance by irradiating a laser Generation control device. The environment map generation control device according to any one of claims 1 to 3 ,
A moving body comprising movement control means for generating a movement route based on the generated environment map and moving along the movement route. In a mobile environment map generation method for generating an environment map to be used for movement control of a mobile object,
Obtaining first ranging information by performing ranging from the moving body toward the road surface;
Obtaining second ranging information by performing ranging from the moving body in the traveling direction;
Whether the obstacle detected from the first distance measurement information and the second distance measurement information is a stationary obstacle or a moving obstacle based on the transition of the recognized state of the area on the environmental map. An obstacle identification step to identify;
The area identified as having a stationary obstacle by the obstacle identifying step continues to be recognized as a non-movable area, and the area identified as having the moving obstacle cannot move for a predetermined period of time. An environmental map generation step of generating the environmental map after recognizing a new area and returning to the original certification before the certification;
Equipped with a,
The obstacle identifying step performs an obstacle identification process including a grid m adjacent to a grid i that is a position on the environmental map where the obstacle detected by the second distance measurement information is located, distance measurement information and environmental map generation method of a mobile obstacle detected by the second distance measurement information is characterized that you identify whether the moving obstacle or a stationary obstacle. In the obstacle identification step, when the obstacle detected from the first distance measurement information and the obstacle detected from the second distance measurement information are in the same area on the environment map,
During the predetermined time set shorter than the predetermined period from the detection of the obstacle by the second ranging information, the obstacle in the area is identified as a moving obstacle,
When there is an obstacle in the area for more than the predetermined time from the detection of the obstacle by the second distance measurement information, it is determined that the vehicle cannot travel because there is a stationary obstacle after the fixed period has elapsed. The method for generating an environmental map of a mobile object according to claim 5 .

Images