JPH10228316A - Autonomous traveling tobot with dead lock prevention device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to quickly escape by changing an operation mode when a robot can no escape from the corner part or the like of an area. SOLUTION: At the time of detecting that the robot 1 enters into the corner part of an area and can not escape in a normal operation mode, the robot 1 is retreated by a distance Db. Then the traveling direction is inverted 180 deg. to escape. Judgement for executing the escape operation is executed in accordance with the execution of hunting operation for repeating directional conversion based on revolution right and left.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an autonomous mobile robot, and more particularly, to an autonomous mobile robot having a function of preventing a deadlock at a corner of a running area.

[0002]

2. Description of the Related Art Mobile robots, such as cleaning robots, lawn mowing robots, plastering robots, and agricultural spraying robots, which perform a predetermined operation by automatically traveling over a given area are known. For example, Japanese Unexamined Patent Publication No.
The cleaning robot described in Japanese Patent Publication No. 6 orbits around a room before starting cleaning, detects the size, shape, and obstacles of the room and maps a traveling area, that is, a cleaning area. Then, based on the coordinates obtained by this mapping operation, the entire room is cleaned by performing spiral traveling in which the radius of the zigzag traveling or the round traveling is reduced for each revolution. This robot is configured to detect a wall surface by a contact sensor and an ultrasonic sensor to determine a course and to detect an end of a lap by a distance meter. Similarly, a robot that travels completely on the floor is disclosed in Japanese Patent Laid-Open No. 5-257.
No. 533, also disclosed.

[0003]

SUMMARY OF THE INVENTION Conventional robots include:
There are the following problems. As described above, the conventional robot has a large number of sensors, and determines the operation to be performed after grasping all the conditions of the traveling area based on the outputs of the sensors. That is, the information detected by the sensor is always taken into the central control device, and the control to be performed next to make the robot travel with high accuracy is determined based on this information. Then, a drive system actuator such as a motor is controlled based on this determination.

As described above, if all the outputs of a large number of sensors are taken into the central control unit for processing, the control system becomes extremely complicated and the processing speed becomes slow. Since the processing takes a long time, there is a problem in that, for example, after detecting an obstacle such as a wall surface, taking a measure to avoid the obstacle is delayed. Further, there is a problem that time is required for initial setting such as setting of thresholds for mapping, teaching, and various processes, and that the initial setting requires skill.

[0005] Therefore, the conventional robot is not only complicated and large, but also expensive, so that its use is limited. For example, at present, it is practically used only in very limited applications such as a transfer robot in a factory and a large cleaning robot.

On the other hand, the use of the robot is not limited to the case where it is necessary to travel the entire target area with high precision, and if the robot can travel the target area almost completely, the traveling direction and the traveling course are changed to the planned course. How accurate it may not be a big deal. For example, in a cleaning robot, depending on the application, it is not necessary to travel with high accuracy so that there is no uncleaned portion over the entire target area.
It is often sufficient to have some uncleaned parts.
Similarly, a lawn mowing robot or the like also needs to keep the mowing traces neat, but if it is a simple lawn mowing robot, it may be sufficient if the mowing is done with a certain degree of quality.

[0007] An object of the present invention is to provide an autonomous traveling robot that can travel in a given area with a simpler configuration based on the recognition of the current state.

[0008]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems and to achieve the object, the present invention provides an autonomous mobile robot whose traveling direction is determined by the rotation directions and rotation speeds of right and left wheels. In the control device, a right obstacle sensor provided on the front right side of the main body, a left obstacle sensor provided on the front left side of the main body, a distance to an obstacle detected by the right obstacle sensor, and the left side Means for respectively determining travel parameters corresponding to the distance to the obstacle detected by the obstacle sensor; and a priority of avoiding the obstacle among the determined at least two travel parameters. Means for determining the rotation direction and rotation speed of the right wheel and the left wheel based on the travel parameters selected based on the Hunting discriminating means for recognizing the occurrence of hunting, and when the occurrence of hunting is recognized by the hunting discriminating means, the right wheel and the left wheel are simultaneously reversely rotated and retracted, and then the right wheel and the left wheel are moved. A first characteristic feature is that there is provided an escape means for performing a reversing operation by a super-spinning operation in which rotation is mutually reversely performed.

Further, according to the present invention, the rotation speeds of the right wheel and the left wheel included in the travel parameters correspond to the distance to the obstacle, and the rotation speed is high when the distance is large and high when the distance is small. Is characterized in that it is set in advance so that the speed is low.

Further, according to the present invention, the hunting discriminating means alternately performs a left-right turning operation in which the right wheel and the left wheel reversely rotate, which is an operation when the distance to the obstacle is extremely small. The third feature is that the hunting is recognized by repeating the scheduled number of times.

According to the first to third features, the parameters for traveling are independently determined on the basis of the distance to the obstacle detected by the sensors provided on the right and left sides in the traveling direction of the robot. In each case, the wheel speed or the like is determined based on the priority so as to be suitable for avoiding an obstacle. Further, when hunting occurs, the hunting can be temporarily reversed, and thereafter, the traveling direction can be reversed to terminate the hunting. In particular, according to the third feature, the occurrence of hunting is recognized based on the rotational state of the left and right wheels, in which the left and right wheels are repeatedly rotating in reverse.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings. FIG. 1 is a schematic diagram showing a schematic configuration of a mobile robot including a control device according to an embodiment of the present invention. In the figure, a robot 1 is advanced by wheels 3 and 4 with endless tracks arranged on both sides of a main body 2, respectively.
It is configured so that each operation of retreat, stop, and turn can be performed. A motor (not shown) is individually connected to each wheel. The turning includes a pivot turning in which one of the left and right wheels 3 and 4 is stopped and only one of the wheels is rotated, and a pivot turning in which both the wheels 3 and 4 are reversed with respect to each other. “Shinchi turning”) is included. A contact sensor (not shown) for detecting contact with an obstacle by sensing pressure is attached to a bumper 5 provided in front of the main body 2.

Further, the robot 1 is provided with eight ultrasonic sensors for detecting obstacles in a non-contact manner.
Sensors 6R, 6 provided in the forward direction of the robot 1
L, sensors 6MR and 6ML provided diagonally in the front, sensors 6DR and 6DL provided in the lower front, and sensors 6SR and 6SL provided on the sides. These sensors preferably comprise ultrasonic sensors, but other obstacle sensors such as optical sensors may be used. Sensor 6
R, 6SR, 6MR, and 6DR detect obstacles on the right side of the traveling direction, and sensors 6L, 6SL, 6ML, and 6DL.
Detects an obstacle on the left side of the traveling direction. And
The rotation of the right wheel 3 is controlled based on the detection signals of the right sensors 6R, 6SR, 6MR, and 6DR, and the rotation of the left wheel 4 is controlled based on the detection signals of the left sensors 6L, 6SL, 6ML, and 6DL. I do.

In the following description, the sensors 6R, 6L, 6M
The R, 6ML, 6DR, 6DL, 6SR, and 6SL are collectively referred to as an ultrasonic sensor 6. The sensor 6
DR, 6DL, 6MR, and 6ML are in ECU 1a, sensors 6SR and 6SL are in ECU 2a, and sensors 6R and 6L are E
Each is connected to the CU 3a. Each ECU is an electronic circuit device that manages a drive circuit of the sensor 6 and input / output of a signal of the sensor 6, and is connected to a control device 7 described later.

In addition to the obstacle detecting sensor, a sensor (not shown) for detecting the rotation speed of the wheels 3 and 4 is provided. Specifically, the rotation speed detection sensor can be realized by an encoder coupled to the wheel driving motor, and an output pulse of the encoder can be used as a rotation speed output.

Next, the hardware configuration of the control device according to the present embodiment will be described with reference to the block diagram of FIG. In the figure, a control device 7 includes a CPU 8 and an ultrasonic sensor 6.
Is connected to the digital input unit 9 via an ultrasonic sensor drive circuit 16 (including the ECU 1a, the ECU 2a, and the ECU 3a) for managing the input / output of the ultrasonic sensor 6. The digital input unit 9 includes the ultrasonic sensor 6 and the bumper 5.
Sensor 5A provided on the left and right wheels 3 and 4
Is connected to a rotation speed sensor (encoder) 10 of a motor for driving the motor. With this configuration, the ultrasonic sensor 6, the contact sensor 5A, the rotation speed sensor 10
Is input to the CPU 8.

On the other hand, the CPU 8 has a digital output unit 11
Via the right wheel electromagnetic brake 12, the left wheel electromagnetic brake 13, and the right wheel motor (hereinafter referred to as "right motor").
14 and a left wheel motor (hereinafter referred to as “left motor”) 15. The instructions based on the processing in the CPU 8 are transmitted through the digital output unit 11 to the electromagnetic brake 12 for the right wheel and the electromagnetic brake 1 for the left wheel, respectively.
3, input to the right motor 14, the left motor 15, and the like. What is supplied to the right motor 14 and the left motor 15 through the digital output unit 11 is a rotation direction instruction signal. The right motor 14 and the left motor 15 have D /
A rotation speed instruction is input from the CPU 8 through the A converter 17.

With the above configuration, the ultrasonic sensor 6 and the contact sensor 5A (hereinafter, collectively referred to as "sensor")
CPU 8 based on the information from the right motor 14
And the operation of the drive system such as the left motor 15 is determined. The robot performs the forward, backward, stop, and turn operations as described above, and the control functions therefor are realized individually by the functions of the CPU 8 as modules. The input process of information from each sensor and the operation determination process are always operating, but the control module for turning, stopping, and retreating is normally in a sleep state, and only the straight-ahead control is activated. . In addition, the turn other than the super pivot turn is included in the function of the straight traveling control module.

The operation judging section of the CPU 8 is configured to cause a predetermined operation to be performed conditionally in accordance with information from each sensor. FIG. 3 is a schematic diagram illustrating an outline of a processing system of the operation determining unit. As shown in the figure, the operation determining unit 18 is configured in a hierarchical manner corresponding to each sensor, and the operation determining unit 18 determines an action plan according to the state of the sensors 6 and 5A and outputs an execution request. I do. Based on the execution request, a drive system (actuator) 19 including the right wheel electromagnetic brake 12, the left wheel electromagnetic brake 13, the right motor 14 and the left motor 15 is controlled. In this way, the execution requests based on the action plan individually determined based on the information from each sensor are accumulated, and the operation of the entire robot is determined.

In the present embodiment, the action plan determined by the state of each sensor is not immediately executed, but is prioritized based on a preset urgency, and the action plan with a high urgency is executed with priority. I did it. FIG. 4 is a block diagram illustrating a function of the operation determination performed in the present embodiment. In the figure, action plans AP1, AP2, ..., AP
When n is determined, the selection unit 20 sets the action plans AP1 to AP1.
An action plan that performs an operation with the highest urgency when avoiding a collision with a wall surface is selected from APn. In the present embodiment, the first priority is set as the operation having the highest urgency when the reverse control is activated. Next, the super pivot turning control is performed in the second
Priority was given, and after that, priority was given in the order of pivot turn, sharp turn, and gentle turn. The pivot turn, the sharp turn, and the gentle turn are included in the straight-ahead control, and are distinguished by the speed difference between the left and right wheels 3 and 4. Based on this prioritization, for example, in order to make a gentle turn to the right at a turning angle θ at a medium speed, the right wheel 3 is set to V1rp.
m and the left wheel at V2 rpm (V2>
The action plan V1) is executed. The prioritization of the action plan is based on the detection result of the ultrasonic sensor 6, and does not include the stop control at the time of detecting an obstacle by the contact sensor.

The retreat control based on the above-mentioned action plan is an operation of retreating for a preset time. In most cases, the process then shifts to a super pivot. In a straight running state, which is a normal running state, the same output is given to the left and right motors 14 and 15.
Based on the pulse signal from the 15 rotation speed sensors 10,
The correction may be performed so that both are the same.

The stop is performed by instructing the rotation speed to be zero and driving the right wheel electromagnetic brake 12 and the left wheel electromagnetic brake 13.
Further, in the stop control, if necessary, an operation to apply regenerative braking may be instructed.

Next, a description will be given of a traveling pattern of the robot obtained by combining the above-described operations. First, a random running which is a basic running pattern of the robot 1 will be described. In FIG. 5, a region A surrounded by a boundary or a wall B
When the robot 1 goes straight ahead and enters within a predetermined distance from the wall B, it stops temporarily, performs a turn-back operation in the order of retreat and turn, and then goes straight again to another wall B.
At this time, the angle of the turning operation in the vicinity of the wall surface B, that is, the turning angle α (see FIG. 5B) is randomly selected and set for each turning operation.

Furthermore, the present inventors have made it possible to avoid running the robot 1 repeatedly in the same place as much as possible, and to run as much as possible in the area as much as possible (hereinafter referred to as "work efficiency"). As a result of the simulation, it has been found that there is an optimum turning angle α in order to improve the angle. The optimal turning angle α is 135 °. Hereinafter, the running pattern in which the turning angle α is set to 135 ° is referred to as fine tuning random running.

The present inventors have found that, in addition to the fine tuning random running, the spiral running is further added, whereby the working efficiency can be further improved.
In other words, the running pattern is that a spiral is formed at the time when the turning at an angle of 135 ° by the fine tuning random running is repeated a predetermined number of times. The running pattern including the spiral is called spiral running.

Next, the running pattern of the spiral running will be described in detail. In FIG. 6, the robot 1 is placed in the area A. This area A is assumed to be a rectangular room surrounded by a wall B. The position where the robot 1 is first placed is arbitrary. FIG.
As shown in (a), the robot 1 starts swirling at the position where the robot 1 is placed. The spiral running is a running pattern in which the turning radius is gradually increased in the turning running, and as will be described in detail later, is controlled on the basis of an operation judgment different from a straight-ahead running, a corner turn, a retreat, and the like. Here, the speed of the left and right wheels 3 and 4, that is, the motor 1 is set so that there is no gap in the running locus.
The rotation speeds of the rotations 4 and 15 are calculated, and these rotation speeds are updated to determine the turning radius. In this way, when the vortex expands and it is recognized that the robot 1 has approached the wall B within a predetermined distance based on the distance to the wall B detected by the output of the ultrasonic sensor 6, the vortex running is performed. It stops and starts straight traveling to move to the next spiral traveling start position (FIG. 6B). Note that the shaded portion in the figure is the traveling locus of the robot 1.

The trigger of the operation of stopping the spiral running and moving to the start position of the next spiral running is as follows. When the robot 1 approaches the wall surface B and the distance detected by each ultrasonic sensor 6 becomes equal to or less than the predetermined distance, the robot 1 performs a turning operation. For example, when the distance to the wall B when the robot 1 detects the wall B is smaller than a predetermined value, the robot 1 stops at that position, retreats by a predetermined distance, and then makes a 135 ° super turning. And move straight away from the wall B. If the distance to the wall surface B when the wall surface B is detected is equal to or longer than the predetermined value, the vehicle turns around at a small angle to avoid the wall surface B.

In this way, the vehicle turns straight back on the wall B, and when approaching another wall B, the vehicle turns straight by retreating and turning or turning, or turning away from the wall B again. When the turning operation is performed on the wall surface B a predetermined number of times N in this way, the vehicle stops at a place where it has traveled straight ahead for the scheduled time T (corresponding to the distance D) so as to move away from the wall surface on which the last turning operation has been performed. A similar spiral is performed (FIG. 6C). Hereinafter, these operations are repeated. In the following description, the distance D that travels straight away from the wall surface where the last turning operation was performed will be described as a representative example of the time T. Can be arbitrarily selected.

FIG. 7 is a graph showing the working time of the robot 1 and the degree of progress of the work in each of the running modes when the present inventors conducted an experiment. In FIG. 7A, the vertical axis represents the ratio of the area of the area covered by the robot 1 traveling in the given area, and the horizontal axis represents the elapsed time from the start of traveling. As the conditions in this case, the traveling area is 4.2 m.
× 4.2 m square, the plane shape of the robot 1 is represented by a circle having a diameter of 20 cm, and the traveling speed of the robot 1 is 1
It was set at 3 cm / sec.

The traveling in a coordinate system is a system in which the vehicle travels along a course set in advance so as to cover the work area. According to the traveling system, the traveling is linearly proportional to the passage of time. The ratio of the area covered by the above increases. On the other hand, in other traveling methods including the spiral traveling, a gentle elongation is exhibited, so that it is difficult to aim for complete coverage of the area. Then, as an example, when the efficiency is compared with the time required to travel over 80% of the area, the spiral traveling is the shortest time (about 1800 seconds) among the three traveling methods except the coordinate system traveling. It can be seen that 80% of the area is covered. Also, the area size was doubled (4.2 m × 8.
4m) An example is shown in FIG. 7 (b). In this case, almost the same tendency was obtained.

Incidentally, it is necessary to set both the number N and the time T to an appropriate number. In other words, if the number N is small, the traveling speed is too close to the previous swirl traveling range, so that the vehicle travels in the same range and the work efficiency is not good. Is not good. If the time T is too short or too long, the spiral running starts near the wall surface, and the wall surface is immediately recognized, which is not efficient.

FIG. 8 shows the result of finding the most efficient time T by simulation. In FIG. 8, the vertical axis shows the work efficiency, and the horizontal axis shows the running time T (seconds) after the turning of the super-spot. The work efficiency is represented by what percentage of the entire travel area traveled in an average over one second. Examples of the fine tuning random running and the random running are values calculated based on the running time (see FIG. 7) required to cover 80% of the entire running area. However, the size of the traveling area is 4.2 m × 4.2 m in the example of FIG.
Is 4.2 m × 8.4 m, and the plane area of the robot 1 is represented by a circle having a diameter of 20 cm, and the traveling speed of the robot is 13 cm / sec, as in the previous experiment.

As shown in the figure, when the time T is changed, two points (peaks) with high working efficiency appear. Therefore, the time T is set by selecting a point at which high efficiency can be obtained in a shorter time. For comparison, the work efficiency in the case of random running and fine tuning random running are also shown.

FIG. 9 shows the result of finding the most efficient number N by simulation. In FIG. 9, the vertical axis shows the work efficiency, and the horizontal axis shows the number of times N. The size of the driving area,
The planar shape of the robot and the traveling speed of the robot are the same as those in the example of FIG. As shown in this figure, when the number N is changed, several points (peaks) with good working efficiency appear. However, the best number N is obtained when the number of times is set to 5 in the case where the region is narrow or wide. is there. Here, the work efficiency in the case of the fine tuning random running and the random running are also shown for comparison.

In the above-described simulation for finding the most efficient time T and number N, the time T and number N do not affect each other, that is, there is no change in the tendency. Therefore, it is only necessary to first select an appropriate value for the time T or the number of times N to determine one of the best values, and then determine the best value of another parameter. In the present embodiment, first, the time T is determined, and then the number N is determined.

Next, the operation of the control device 7 will be described with reference to a flowchart. First, the input processing of the ultrasonic sensor 6 will be described. In FIG. 10, in step S100, a processing request from the ultrasonic sensor 6 (ultrasonic sensor drive circuit 16) is waited. If there is a processing request, the process proceeds to step S110, and it is determined whether or not the processing request is an ultrasonic transmission processing request, that is, whether or not the ultrasonic sensor 6 has transmitted the ultrasonic wave to the CPU 8. If the request is an ultrasonic transmission processing request, the process proceeds to step S120, where a predetermined counter value is set and a down count is started. This counter value corresponds to the maximum time in which reception (reflected wave) is expected to be obtained in view of the reception capability of the ultrasonic sensor 6.

In step S130, it is determined whether or not there is a request for ultrasonic reception processing, that is, whether or not the ultrasonic sensor 6 has received a reflected sound. If there is no reception processing request, step S140
Then, it is determined whether or not a timeout has occurred, that is, whether or not the counter value is "0", and a reception processing request is waited until a timeout occurs. On the other hand, if it is determined in step S130 that a reception processing request has been made, the process proceeds to step S150, and the counter value at that time is obtained and stored. When the counter value is obtained, the counter is cleared in step S170. If there is no reception processing request until the timeout occurs, the counter value “0” is stored as the current detection value in step S160, and the counter is cleared in step S170. In step S180, it is preferable to perform a majority decision process for improving the reliability of the stored counter value. The details of this majority decision process will be described later with reference to FIG. In the following description, the counter value stored in the ultrasonic sensor input processing will be referred to as a distance counter value.

FIG. 11 is a flowchart of the input processing of the contact sensor 5A.
If the detection signal has been input from step S90, in step S90,
Instruction for stopping the robot 1, that is, the left and right motors 14, 15
And outputs the rotation speed “zero” to the right and the electromagnetic brake 12 for the right wheel and the electromagnetic brake 13 for the left wheel. The input processing of the contact sensor 5A is, for example, an interrupt processing every 10 msec.

Next, the control operation based on the output signal of each sensor will be described. In the general flow of FIG.
In step S1, a vortex processing start instruction is issued. As a result, the spiral processing is started. The details of the spiral processing will be described later with reference to FIG. In step S2, the result of the ultrasonic sensor input processing, that is, the distance counter value is requested. In step S3, a contact sensor input process is performed.
In step S4, it is determined whether or not a spiral is occurring. Initially,
This determination is affirmative since the spiral has been started in response to the spiral processing start instruction. If it is during the spiral and the determination is affirmative, the process proceeds to step S5, and a spiral continuous determination process is performed to determine whether or not to continue the operation. This spiral continuation determination processing is based on the distance counter value obtained as a result of the ultrasonic sensor input processing, and when it is determined that the robot 1 has approached within a predetermined distance from the wall B, the spiral is stopped and the next spiral is started. This is a process for determining whether to move to a position. In the present embodiment, the determination is made based on whether the vehicle approaches within 30 cm from the wall surface. In step S6, it is determined based on the processing in step S5 whether or not to stop the spiral. If it is not determined that the spiral is stopped, the process returns to step S2.

If it is determined that the spiral is to be stopped, the flow advances to step S8 to determine the next operation. The next action when the vortex is stopped is unconditional retraction. This is because, since the vehicle is approaching the wall surface B and hits the wall surface B in a normal turning operation, it is necessary to take a procedure of retreating, turning a super-spot, and moving forward.

If the operation subsequent to step S8 is "retreat", the flow advances to step S9 to determine whether the vehicle is currently retreating. Immediately after stopping the swirl, it does not retreat,
This determination is negative, and the process proceeds to step S10, stops instantaneously (20 to 30 ms), and then proceeds to step S11. If the determination in step S9 is positive, step S1
The process skips 0 and proceeds to step S11. Step S1
At 1, a request is made to start the retreat processing. The retreat processing will be described later.
If the reverse process is started in response to the reverse process request in step S11, it is not "whirlpool running", so the determination in step S4 is negative and the process proceeds to step S7. If the vehicle is not running in a spiral, the determination in step S8 is made based on the result of the determination processing in step S7.

In step S7, processing for determining the next operation to be performed based on the distance counter value for each ultrasonic sensor 6 (next operation determination processing) is performed. In the next operation determination process, the next operation to be performed is determined based on the travel parameters obtained from the travel parameter determination table of FIG. FIG.
In, parameters are passed to the control to be started in accordance with the detection values of the sensors corresponding to the distance to the wall surface. The parameters are wheel speed, turning level, and turning direction. The speed of each of the right and left wheels of the robot 1 is set at four stages: an extremely low speed (0.5 km / h), a low speed (1.0 km / h), a medium speed (2 km / h), and a high speed (3 km / h). , Gentle turning (angle 30 ° “3”), sharp turning (angle 60 ° “2”), pivot turning (turning with one wheel stopped “1”), super turning (left and right wheel reversing “0”) ) 4
The stage has been set. The turning direction is set to right and left. When the ultrasonic sensor 6 (6SR, 6R, 6DR, 6MR) provided on the right half of the robot 1 detects a wall surface, the left turning and the left half are performed. When a wall surface is detected by the ultrasonic sensor 6 (6SL, 6L, 6DL, 6ML) provided in the above, a right turn determination is made. here,
Two parameters representing the speed of the wheels of the robot 1 are provided. One is the speed of the wheel on the side provided with each sensor, and the other is the speed of the wheel on the side opposite to the side provided with each sensor. For example, the right sensor 6R
, Two speeds to be output to the motor on the self side (right side) and the motor on the other side (left side) are determined. However, since a large number of parameters are obtained corresponding to the detection values of all the sensors 6, which parameter is finally adopted follows a predetermined rule. In principle, the rules are determined so that the one with a lower speed takes precedence. However, in exceptional cases, when “stop” is passed to both the left and right motors 14 and 15, the “extremely low speed” Shall be determined.

For example, if the distance determined from the distance counter value of the right front sensor 6R is 0.5 m to 1 m, the right motor speed is "low" and the left motor speed is "stop".
Becomes The control name to be activated is turning, and the turning level is “parallel turn: 1” and the turning direction is “left”. On the other hand, at this time, if the distance determined from the distance counter value of the left front sensor 6L is 1 m to 1.5 m, the left motor speed is "medium speed" and the right motor speed is "extremely low speed". The control name to be activated is turning,
The parameters of the turning level are “rapid turning: 2” and the turning direction is “right”. If these parameters are collated and a lower speed is adopted, the right motor is determined to be “extremely low” and the left motor is determined to be “stop”.
Will make a left turn at a very low speed.

As another example, if the distance determined from the distance counter value of the sensor 6R is 1.5 m to 2 m, the right motor speed is "medium speed" and the left motor speed is "low speed". Becomes At this time, if the distance determined from the distance counter value of the sensor 6L is also 1.5 m to 2 m, the left motor speed is “medium speed” and the right motor speed is “low speed”. If these parameters are collated and a lower speed is adopted, the right motor is determined to be “low speed” and the left motor is determined to be “low speed”.
Will go straight at low speed. For example, when the robot 1 is moving straight against the wall, the left and right sensors 6L, 6L
The detected value of R becomes the same, the speed gradually decreases, approaches the wall surface, and finally, the vehicle moves backward and makes a turn with a super-revolution to move away from the wall surface.

As a result of the next operation determination processing in step S7, if it is determined in step S8 that the turning of the pivot point is performed, the process proceeds to step S12. In step S12, it is determined whether or not the current pivot turn is being performed. Initially, the determination is negative, and the process proceeds to step S13 to stop instantaneously (20 to 30 ms). In step S14, an instruction to start the super turning operation is issued. In the cycle after the start instruction of the pivot turn processing is issued in step S14, the determination in step S12 is affirmative, and the process proceeds to step S15 to perform the escape mode processing. In consideration of the fact that if the robot 1 is stuck in the corner of the area, a state in which the robot 1 cannot escape from the corner in the normal retreat and the turning of the super-spot is generated, the escape mode processing is performed during the turning of the super-spot. Escape mode processing
It will be described later with reference to FIG.

After turning at a predetermined angle (for example, 135 °) by turning the super-powered base, if it is determined in step S8 that the vehicle is going forward in the next cycle, the process proceeds to step S16. In step S16, it is determined whether or not the vehicle is currently moving forward.
At first, the process proceeds to step S17 and is instantaneous (20 to 30 ms).
After stopping, the process proceeds to step S19 via the hunting prevention process in step S18, and an instruction to start the forward process is issued. The robot 1 moves forward in accordance with the start instruction of the forward processing, and when approaching a predetermined distance from the wall surface, an operation (running) to be performed next is determined according to the distance counter value.
The hunting prevention processing is performed, for example, when a plurality of obstacles are detected, a frequent change of direction that occurs immediately after an operation to avoid one obstacle is detected and another obstacle is detected and the avoidance operation is performed again. (Hunting). Since the hunting prevention processing is not a main part of the present invention, a detailed description is omitted. Note that Japanese Patent Application No.
No. (A961333).

Next, the operation of each control (software) module based on the output signal of each sensor will be described. First, the spiral processing of the robot 1 will be described. The swirling operation is shown in FIG.
As described with regard to, the operation is started at the start of the work and when the time T has elapsed since the last reflection operation, the reflection involving the turning of the pivot point among the operations of detecting and reflecting the wall surface is repeated N times. At the start of the operation, step S1 (FIG. 1)
The spiral processing is started by the processing start instruction by the operator in 2), and during the operation, the spiral processing is performed by the processing start instruction issued from the super-real processing module after executing the super-real turning N times (see step S33 in FIG. 15). Is started. Here, a case where the processing is started by a processing start instruction from the super-spot processing module will be described.

In FIG. 14, in step S20, the process waits until a spiral processing start instruction is received from the super base processing module. In other words, after performing the super-revolution turn a predetermined number of times N, it waits for an instruction issued from the super-reception processing module. If there is an instruction to start the vortex process, the timer is started by setting the timer to, for example, 26 seconds as the straight-ahead time until the start of the vortex after the time T, that is, the last turn of the corner, in step S21. It should be noted that the spiral processing start instruction is output almost simultaneously with the start of the last, that is, the start of the N-th super-revolution turn processing (see step S33 in FIG. 15).
Here, the time T includes a retreat time for a super-spot turn and a super-spot turn time (see step S35 in FIG. 15). In step S22, it is determined whether or not the time T has elapsed. If the time T has elapsed, the process proceeds to step S2.
Proceed to 4. Until the time T elapses, the process proceeds to step S23 to determine the presence or absence of the processing stop instruction. If the processing stop instruction is received before the time T elapses, the process proceeds to step S20.
Return to

In step S24, it is determined whether or not the spiral has ended. If this determination is affirmative, the process returns to step S20, but otherwise proceeds to step S25. Step S
At 25, the speeds of the left and right wheels 3, 4 are calculated and set to determine the size of the spiral. According to the speeds of the left and right wheels 3, 4 set here, the left and right motors 14,
The instruction of the rotation speed is given to 15, and the spiral running is executed.

In step S26, a time t until the speeds of the wheels 3 and 4 are updated so that the spiral smoothly expands is calculated, the time is set as a timer, and the timer is started. In step S27, it is determined whether or not the time t has elapsed. If the time t has elapsed, the process proceeds to step S2.
Proceed to 4. In step S28, the presence or absence of a stop instruction is monitored. If there is no stop instruction, steps S27 and S28 are repeated until the time t elapses. If the distance to the wall surface or the obstacle is equal to or less than the predetermined distance or if there is a stop instruction based on the detection signal of the contact switch 5A based on the distance counter value, the determination as to whether the spiral has ended is affirmative. .

Next, the super turning operation will be described. FIG.
In step S30, the process waits for a processing start instruction. In step S31, the number of times of the super-site turning (hereinafter, simply referred to as “the number of super-sites”) n is incremented (+1). In step S31, it is determined whether or not the number of times of superposition n has reached the predetermined number of times of reflection "N". Since the initial value of the number of superpositions n is set to “0”, the number of superpositions n is “1” at the time of the first judgment, the judgment in step S32 is negative, and the steps S33 and The process skips S34 and jumps to step S35.

In step S35, the turning time is calculated.
Since the turning angle is determined by the turning time, a time corresponding to the turning angle of 135 ° is calculated. After the turning time has been calculated, the process proceeds to step S36, and an instruction is issued to cause the right wheel 3 and the left wheel 4 to reverse each other. Here, which of the right wheel 3 and the left wheel 4 is rotated forward or backward is determined by the “turning direction” of the traveling parameter. In step S37, the process waits until the turning time elapses or a stop instruction is issued. In step S38, it is determined whether or not the pivot turn is completed. Upon completion of the pivot turn, the flow advances to step S39 to give a forward rotation instruction to the left and right wheels 3 and 4. That is, the vehicle is returned to the straight traveling, which is the basic traveling mode.

If the number of times of superposition n reaches the number of times of reflection N, the flow advances from step S32 to step S33 to instruct a spiral processing start. Then, in step S34, the number of times of super-simultaneous determination n used for determining whether or not to perform a super-simultaneous pivot is cleared. Subsequently, in steps S35 to S39, the processing of the pivot turn is completed, and the next processing start instruction of the pivot turn is awaited. In step S33, a spiral processing start instruction is issued. However, in the previously described spiral processing, the actual spiral is started after the elapse of the time T, and therefore, the spiral does not overlap with the super turning.

Next, the backward processing will be described. In FIG. 16, in step S50, the process waits for a process start instruction. In step S51, an instruction to reverse the right wheel 3 and the left wheel 4 is issued. In step S52, the process waits until the scheduled retreat time elapses or until there is a stop instruction. In step S53,
When the retreat is completed, the process proceeds to step S54, in which an instruction for forward rotation is given to the left and right wheels 3, 4.

Next, a majority decision (sensor signal selection) process in the ultrasonic sensor input process will be described. Since the output of the ultrasonic sensor 6 may become unstable depending on the environment, it is preferable to perform majority processing in order to improve the reliability of the acquired data (distance counter value). In the majority decision processing employed in the present embodiment, when new data is acquired, a majority decision is made between two pieces of past data and highly reliable data is selected. In FIG. 17, margins m are set above and below the previous value MID, and if the current value NEW is within this margin m, the current value NEW is not far from the previous value MID, so it is determined to be a normal detection signal. Then, the current value NEW is adopted as it is as the current distance counter value (see FIG. 17A).

However, if the current value NEW is outside the range of the margin m, it is determined whether or not the value OLD before the previous time is within the range of the margin m. If the value OLD before the last time is within the margin m of the previous value MID, the previous value MID
Is not far from the previous turn OLD, it is determined to be a normal detection signal, and the current value NEW is replaced with the previous value MID to obtain a distance counter value (see FIG. 17B).

Further, when the previous value OLD is not within the margin of the previous value MID, it is considered that the three data are far apart from each other, and the current value NEW is detected as the highly reliable data at this time. This is adopted as the distance counter value (see FIG. 17C).

Next, the details of the "escape mode process" will be described. FIG. 18 is a diagram illustrating a state where the robot 1 is stuck in a corner of the area and is stuck, that is, a deadlock state. In the same figure, the robot 1 has a right wall W at a position P1.
When R is detected, the vehicle turns left to avoid the wall WR. It is assumed that the distance from the position P1 to the right wall WR is so close that the robot 1 cannot avoid the wall WR unless the robot makes a corner turn. When the robot 1 detects the left wall WL at the position P2 after avoiding the wall WR, the robot 1 turns right (super turning) to avoid the wall WL. Then, in order to detect the right wall WR at the position P3, the vehicle turns left again (super-revolution) to avoid the wall WR. In this way, the vehicle proceeds to the corner of the area while gradually reducing the inversion interval, and eventually stops when both the left and right walls WR and WL are detected at a short distance.

In such a state, it is impossible to escape in the normal operation mode. Therefore, in this embodiment, when the robot 1 repeatedly repeats the left and right turn, a hunting operation at a corner occurs. It is determined that the user has performed the “escape operation”. FIG. 19 is a schematic diagram illustrating the movement path of the robot 1 in the “escape operation” from the deadlock. As shown in the figure, in the “escape operation”, the robot 1 retreats a distance Db longer than the retreat in the normal mode, and performs a 180 ° super-spinning turn (PT180) to escape from the deadlock state. Aim. Note that the distance Db is substantially the same even if it is time or the amount of rotation of the wheels 3 and 4.

The "escape mode process" will be described with reference to a flowchart. In FIG. 20, the direction of the pivot turn is determined in step S61, and depending on whether the direction of the pivot turn is right or left, step S62 or step S6 is performed, respectively.
Proceed to 3. In step S62 and step S63,
It is determined whether or not the direction of the pivot turn is the same as the previous pivot turn. If the direction of the turn is the same, it is determined that the hunting operation has not occurred, and the process exits the escape mode processing.

On the other hand, if the direction of the pivot turn is different from the previous time, it is determined that a hunting operation may have occurred, and the process proceeds from steps S62 and S63 to step S64. In step S64, the value of a counter (escape counter) for determining whether or not to proceed to the escape operation, that is, the value of the number of times of the pivot turn is incremented. Note that the initial value of the escape counter value is “0”. In step S65,
The escape counter value is a preset value C (for example, “10”)
Is determined, and if the determination is affirmative, the process proceeds to step S66 to instruct an escape operation. The instruction of the escape operation is given by a slightly longer distance set in advance (for example, 1
m) A backward instruction and a 180 ° turn-in-turn instruction.
After the instruction for the escape operation, the process proceeds to step S67 to clear the escape counter value.

Next, a function for realizing the escape mode processing will be described. FIG. 21 is a main functional block diagram of the escape mode processing. The super-spot determination unit 21 includes a right motor control unit 22 and a left motor control unit
Based on the control states of 4 and 15, it is detected that the pivot turn is performed, and the turning direction is stored in the turning direction memory 24. The comparison unit 25 compares the current turning direction with the previous turning direction, and outputs a clear signal to the escape counter 26 if they are the same, and outputs an increment signal to the escape counter 26 if they are different. . When the escape counter 26 reaches a predetermined value, a count-up signal is supplied to the escape instruction unit 27. Escape instruction unit 27
In response to the count-up signal of the escape counter 26, the right motor control unit 22 and the left motor control unit 23 are instructed to perform the scheduled escape operation, that is, the retreat and the turning of the corner.

[0063]

As is apparent from the above description, according to the first to third aspects of the present invention, each of them is based on the detection results obtained by the sensors provided on the right and left sides in the traveling direction of the robot. The original traveling parameters are determined, and the final traveling parameters are determined according to the priority. Therefore, the respective control units corresponding to the left and right sensors only determine their own traveling parameters without considering the others, and the control is simplified. Especially,
When a deadlock state occurs at a corner or the like of a region, the state can be quickly escaped from the state.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a configuration of a robot according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a hardware configuration of a control device according to the embodiment of the present invention.

FIG. 3 is a schematic diagram illustrating an outline of a processing system of an operation determining unit.

FIG. 4 is a block diagram for explaining an action of selecting an action plan.

FIG. 5 is a schematic diagram showing a basic running pattern of the robot.

FIG. 6 is a schematic diagram showing a traveling pattern of spiral traveling.

FIG. 7 is a diagram of a simulation result showing a degree of progress of a work according to various traveling patterns.

FIG. 8 is a diagram of a simulation result showing a relationship between work efficiency and time T.

FIG. 9 is a diagram of a simulation result showing a relationship between work efficiency and the number of turns N.

FIG. 10 is a flowchart of an ultrasonic sensor input process.

FIG. 11 is a flowchart of a contact sensor input process.

FIG. 12 is a general flowchart showing determination of the operation of the spiral running.

FIG. 13 is a traveling parameter setting table showing the correspondence between sensor outputs and traveling parameters.

FIG. 14 is a flowchart showing processing of a spiral part.

FIG. 15 is a flowchart showing a process of turning a super-spot.

FIG. 16 is a flowchart showing retreat processing.

FIG. 17 is a schematic diagram of a sensor signal selection process.

FIG. 18 is a schematic diagram illustrating a deadlock state of the robot.

FIG. 19 is a schematic view showing an operation of escaping from a deadlock.

FIG. 20 is a flowchart showing escape mode processing.

FIG. 21 is a block diagram showing main functions of a control device for an escape mode process.

[Explanation of symbols]

1: Robot, 3: Right wheel, 4: Left wheel, 5: Bumper, 5A: Contact sensor, 6: Ultrasonic sensor,
7 ... control device 18 ... operation judging unit 20 ... selecting unit
21: super-spot determination unit 26: escape counter 27
… Escape instruction section

Claims (3)

[Claims] 1. An autonomous mobile robot in which a traveling direction is determined by a rotation direction and a rotation speed of each of a right wheel and a left wheel, a right obstacle sensor provided on a front right side of a main body, and a right obstacle sensor provided on a front left side of a main body. The left obstacle sensor, the distance to the obstacle detected by the right obstacle sensor,
And means for respectively determining travel parameters corresponding to the distance to the obstacle detected by the left obstacle sensor; and a schedule for avoiding the obstacle among the at least two determined travel parameters. Means for determining the rotation direction and rotation speed of the right wheel and the left wheel based on the traveling parameter selected based on the priority of hunting; hunting discriminating means for recognizing the occurrence of hunting that alternately changes the traveling direction left and right; and When the occurrence of hunting is recognized by the hunting discriminating means, the right wheel and the left wheel are simultaneously reverse-rotated and retracted, and then the reversing operation is performed by a super-spinning turn in which the right wheel and the left wheel are reverse-rotated with each other. An autonomous mobile robot with a deadlock prevention device, comprising: an escape means for performing the operation. 2. The rotation speed of the right wheel and the left wheel included in the traveling parameters corresponds to the distance to the obstacle, and the rotation speed is high when the distance is large and low when the distance is small. The autonomous mobile robot with a deadlock prevention device according to claim 1, wherein the autonomous mobile robot is set in advance. 3. The hunting discriminating means repeats a left-right turning operation of rotating the right wheel and the left wheel in opposite directions, which is an operation when the distance to the obstacle is extremely small, alternately left and right a predetermined number of times. The autonomous mobile robot with a deadlock prevention device according to claim 1, wherein the occurrence of the hunting is recognized.