JPH04153709A - Guidance controller for moving robot - Google Patents

Abstract

PURPOSE:To enable the robot to run by itself in not only adjusted environment, but also environment which is not adjusted to some extent by selecting optimum run control among dead reckoning navigation, dead reckoning accumulative error corrective running, and obstacle avoiding running according to the circumferential environment of the robot. CONSTITUTION:A dead reckoning navigation part 70, an obstacle avoiding run part 80, and a position correction part 30 are three run modules of the device. A navigator 60 selects the most suitable run system for a run on a route selected by a route planner 55 among the three run systems, i.e. dead reckoning navigation, accumulative error corrective running by dead reckoning navigation, and obstacle avoiding running, and fuzzy reasoning is used to decide the intention. Namely, the navigator 60 knows the environmental state of an obstacle, etc., from the output of a range finder 20, calculates the effectiveness values of the three run systems by applying the fuzzy reasoning, and select the run system having the largest effectiveness. Consequently, the robot is enabled to run completely by itself even in unadjusted environment.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a guidance control device for mobile robots used for transportation, building security, plant maintenance, etc., and is particularly suitable for both organized environments and unknown environments. The present invention relates to a guidance control device for a mobile robot that allows the mobile robot to travel using a traveling method.

[Conventional technology]

Conventional travel guidance technologies for mobile robots include: ■ A guide line or guiding body is placed on the travel path and the robot travels along a preset route. ■ Dead Reckoning (Dead Recko)
n! ng), which allows the robot to run autonomously.

[Problem to be solved by the invention]

In the prior art (2) above, the robot can only travel on a fixed course, and there is a problem in that changing the course requires a lot of time and effort.

The problem with the prior art (2) above is that the robot cannot run in an environment with many obstacles, and there are many restrictions on the environment in which the robot can be introduced.

This invention was made in view of the above circumstances, and an object of the present invention is to provide a guidance control device for a mobile robot that allows the robot to run independently not only in an organized environment but also in an unorganized environment to some extent. do.

[Means to solve the problem]

In this invention, a plurality of guidance marks are distributed and arranged on a traveling road surface of a mobile robot, and the mobile robot has a map memory in which the positions of the plurality of guidance marks and a travel destination are stored, information stored in the map memory, and a route selection means for selecting a route for the robot to travel based on instructed traveling destination information; a first measuring means for measuring the traveling length and direction of the robot; and a first measuring means for measuring the traveling length and direction of the robot. first travel control means that sequentially estimates the robot's self-position and causes the robot to travel by dead reckoning along the route selected by the route selection means;
a second measuring means for measuring the guide mark;
a second travel control means that performs travel control to correct the cumulative error of dead reckoning based on the detection output of the measurement means; a third measurement means that measures obstacles around the robot; a third travel control means that performs obstacle avoidance travel control based on the measurement output of the measurement means; and the first to third travel control means based on the outputs of the first to third measurement means. and navigator means for selecting one of them.

[Effect]

According to this configuration, the optimal travel control among dead reckoning navigation, dead reckoning cumulative error correction travel, and obstacle avoidance travel is selected depending on the environment around the robot.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows the hardware configuration of a control system for a mobile robot according to an embodiment of the present invention.

First, as a prerequisite, in this embodiment, the following three driving methods are switched according to environmental conditions to perform driving.

■Dead Reckoning ■Driving to compensate for accumulated errors in dead reckoning (position correction driving) ■Obstacle avoidance driving In Figure 1, the mobile robot drives a pair of drive wheels attached to the left and right sides of the vehicle body in different directions. The vehicle is configured to be driven by motors IL and IR, and the average value of the rotational speeds of the left and right wheels determines the vehicle speed, and the steering is determined based on the rotational speed difference between the left and right wheels.

The travel control unit 10 is a core module of the robot, and controls the vehicle speed and steering of the robot by driving motors IL and IR, which are movement actuators of the robot, through processing to be described later.

The range finder 20 is used to detect obstacles around the robot, and uses modulated infrared rays (laser light or LED light) that are narrowed down to a narrow beam to rotate and scan 360 degrees. The distance d between the robot and the obstacle is determined from the phase difference between the transmitted signal and the scattered light from the obstacle that is emitted and received. We are looking for it by detecting it. Therefore, the range finder 20 outputs a (d, θ) signal, which is a combination of the laser projection direction θ and the distance d, as a detection signal. Details regarding this range finder 20 can be found in the patent application No. 6 filed by the applicant on August 10, 1988.
3-19931.5, etc.

The position correction unit 30 performs position correction calculations using the detection output of the position correction sensor 2 that detects guidance marks (landmarks) for correcting accumulated errors in dead reckoning navigation.

Various methods have been proposed for correcting accumulated errors in dead reckoning, but in this case, for example, metal tape in the shape of two characters is distributed as appropriate on the floor where the robot runs, and these are mounted on the robot. Detected by an eddy current metal sensor,
A method is adopted to detect the robot's lateral deviation based on the detection output.

These two marks are shown in Japanese Patent Application No. 61-091515 and Japanese Patent Application No. 1-175097 filed by the present applicant.

The work machine drive unit 40 appropriately drives a transfer machine such as a manipulator or a push-pull mounted according to the purpose of the robot at the destination.

These travel control section 10, range finder 20, position correction section 30, and work equipment drive section 40 are connected via a parallel communication bus, and these modules freely communicate with each other using LAN (local area network) communication technology. We can meet.

FIG. 2 shows the software structure related to travel control, and the core part of this software structure is mainly installed in the travel control section IO.

The map memory 50 stores map information of the environment in which the robot runs, and an example thereof will be explained with reference to FIG.

In this example of FIG. 3, there are two robot waiting points C1 and C2.
and 10 landmarks #1 to #10 (in this case Z
The map memory 50 stores the absolute coordinates of each landmark #1 to #10 and the information between these landmark materials 1 to #10. Connection information (information indicating which landmark each landmark is directly connected to) and absolute coordinates of the work place and waiting point are stored. In this case, the positions A1 to A4 of the workplace are represented by landmarks #1 to #4, and the positions of waiting points C1 and C2 are represented by landmarks #9. I try to show it in #10.

The route planner 55 determines the optimal route to the destination, based on sensor information such as the range finder 20 and position correction sensor 2, map information in the map memory 50, and travel commands specified by the operator. Determine the optimal route. Possible indicators of optimality include the shortest distance and shortest time, but here we use the shortest time, which takes into account not only distance but also losses due to turns and curves. Optimal algorithms include DP (dynamic programming) and A
*Huristic methods such as algorithms and Dji
Although there are methods such as kstra method, Djikstra method is considered to be the most suitable.

For example, in Figure 3, if a robot waiting at mark #9 is given a run command to go to the work area marked #4 and work, there are multiple routes from mark #9 to mark #4. However, the route planner 55 determines the optimal route using the shortest time measure based on the Djikstra method. For example, #9→#7-#3
→Route #4 is selected.

The dead reckoning unit 70 uses dead reckoning to drive the robot, and the only sensors used in this dead reckoning are encoders attached to each drive wheel of the robot. That is, the dead reckoning section 70 calculates the traveling length of each drive wheel from the detection output of this encoder, and uses this calculated value to calculate the moving distance and traveling direction of the robot. Then, the dead reckoning unit 70 uses this calculated value as an estimated value of the self-position, and further calculates the lateral deviation and angular deviation from the straight course connecting the two landmarks, and eliminates these deviations so that the vehicle can travel on the course correctly. Each traveling motor IL,
Controls the rotation speed of IR.

As the assumed course, not only a straight course but also a circular arc or curved course can be adopted, and in this case as well, the robot is caused to travel by feedback travel similar to that described above. Note that when controlling the vehicle by dead reckoning navigation, a direction sensor such as a gyro may be used in addition to the encoder described above.

The obstacle avoidance travel section 80 performs travel control based on the output of the range finder 2° so that the vehicle can head toward the goal without colliding with obstacles, and fuzzy reasoning is also used during this travel control. .

For example, the obstacle avoidance traveling unit 80 includes a passability determination unit that divides the travel area when selecting a route to avoid obstacles, and determines whether each divided area is passable or not, and a passability determination unit that determines whether or not each divided area is passable. It has a traversable area selection unit that selects one optimal traversable area when there are traversable areas, and uses fuzzy inference when selecting the optimal traversable area. This obstacle avoidance technique is known, for example, in Japanese Patent Application No. 1-22 filed by the present applicant.
No. 6568.

As described above, the position correction unit 30 performs position correction calculations using the detection output of the position correction sensor 2 that detects landmarks for correcting accumulated errors in dead reckoning navigation.

The dead reckoning navigation section 70, the obstacle avoidance traveling section 80, and the position correction section 30 are three traveling modules in this device, and one of these traveling modules is selected by the navigator 60.

The navigator 60 selects the three driving methods most suitable for driving on the route selected by the route planer 55.
Two driving methods, ■ Dead Reckoning (Dead Reckoning)
g) ■Dead Reckoning Accumulated Error Correction Driving (Position Correction Driving)■
The choice is between driving to avoid obstacles, and fuzzy reasoning is used to make this decision. That is, the navigator 60 learns the surrounding environmental conditions such as obstacles from the output of the range finder 20, derives the effectiveness of the above three driving methods using fuzzy reasoning based on this, and selects the driving method with the highest effectiveness. Select a method.

The method of selecting the driving method by the navigator 60 is basically as follows.

■Dead reckoning is usually selected.

■If a landmark is detected during dead reckoning, position correction driving is selected.

■If it discovers an obstacle that can be avoided during dead reckoning, it selects obstacle avoidance driving.

Note that if the obstacle is too large to avoid or if the user gets lost in a dead end, the selected travel module sends a message to the navigator 60 that the vehicle cannot travel. Upon receiving this, the navigator 60 requests the route planner 55 to create an alternative route. Upon receiving this request, the route planner 55 creates an alternative route and re-inputs the created alternative route into the navigator 60.

Further, when the navigator 60 is given a route from the route planer 55, it performs travel control by treating each landmark in the route as an intermediate goal for each travel section. For example, in Figure 3, if the route #9 → #7 → #3 → #4 is selected, the landmark #7 is the first route selected.
is treated as an intermediate goal, and when the landmark #7 is reached, travel control is executed to set landmark #3 as the next intermediate goal.

The operation of this embodiment will be explained with reference to the flowchart of FIG.

First, when a travel command indicating a target value is given by the operator, the route planner 55 determines a route using the method described above (steps 100, 110), and notifies the navigator 6o of this. The navigator 60 uses fuzzy reasoning to determine the most suitable travel module for traveling on the selected route using the dead reckoning unit 70. The driver selects the obstacle avoidance travel module 8o and the position correction module 3o, and activates the selected travel module (step 12o).

Note that if the navigator 60 determines that it is impossible to run the course based on the outputs of various sensors, etc., it requests the route planner 55 to replan the course (step 130).

Does the activated travel module perform travel control using the travel method of the travel module? (Step 1)
40) If a change occurs in the surrounding environment and it is determined from the outputs of various sensors that it is impossible to travel with the travel method of the travel module in question, a request is made to the navigator 60 to reselect the travel module (step 150). .

Such control is repeatedly executed until the destination is reached.

That is, in this travel control, the control basically flows from the upper hierarchy to the lower hierarchy, and the route planer 55 generates a route to the destination, and travel control is executed by sequentially deploying this to the lower hierarchy. Ru. One task is completed when the robot finishes its work at a predetermined destination, and the robot waits for the next command.

There are two ways in which information flows from a lower layer to an upper layer. One case is when the plan determined by the higher level cannot be executed, and in this case, the lower level can request the higher level to formulate a plan. The other case is when a change occurs in the surrounding environment, in which case the sensor information serves as a trigger and the navigator 60 is activated first. In this case, most of the problems are handled by replacing the travel module. An example of such a case is as described above: When a landmark is detected during dead reckoning navigation, position correction driving is selected.

- If the vehicle discovers an obstacle that can be avoided during dead reckoning, it selects obstacle avoidance driving.

However, if an obstacle is too large to be avoided, or if the vehicle gets lost in a dead end, it returns to the route planer 55 and re-plans the route.

Note that the present invention is not limited to what is shown in the above embodiments, and can be modified as appropriate. Landmarks can be optical marks, such as marks with retroreflectivity, that are illuminated by a light source mounted on a robot. Alternatively, the reflected light may be imaged with a TV camera or the like. Further, a triangulation method may be adopted as the range finder.

〔Effect of the invention〕

As explained above, according to the present invention, the optimal travel control among dead reckoning, dead reckoning cumulative error correction travel, and obstacle avoidance travel is selected according to the environment around the robot. The robot can run completely autonomously without the need for human assistance even in environments where mobile robots are not used, and as a result, the range of applications and activities of mobile robots can be expanded.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of the /%-ware configuration according to the embodiment of the present invention, FIG. 2 is a diagram showing an example of the software configuration according to the embodiment of the present invention, and FIG. 3 is a diagram explaining the stored contents of the map memory. , FIG. 4 is a flowchart showing the operation of the embodiment. IR, IL... Travel motor, 2... Position correction sensor, 10... Travel control section, 20... Range finder, 30... Position correction section, 40... Working machine drive section, 5
0...Map memory, 55...Root planar, 60
...Navigator, 70...Dead Reckoning Department, 80...
・Obstacle avoidance traveling section Fig. 1 Fig. 2 A2 Fig. 3 Fig. 4

Claims (2)

[Claims] (1) A plurality of guidance marks are distributed and arranged on the traveling road surface of the mobile robot, and a map memory in which the positions of the plurality of guidance marks and the travel destination are stored, and the information stored in the map memory and the designated travel purpose are provided. a route selection means for selecting a route for the robot to travel based on ground information; a first measuring means for measuring the travel length and direction of the robot; and sequential estimation of the self-position based on the measured travel length and direction. and causing the robot to travel by dead reckoning along the route selected by the route selection means.
a traveling control means, a second measuring means for measuring the guide mark, and a second measuring means for measuring the guide mark;
a second travel control means that performs travel control to correct the cumulative error of dead reckoning based on the detection output of the measurement means; a third measurement means that measures obstacles around the robot; and a third measurement means that measures obstacles around the robot. a third travel control means that performs obstacle avoidance travel control based on the measured output of the means;
A guidance control device for a mobile robot, the mobile robot comprising: a navigator means for selecting one of the first to third travel control means based on the output of the measuring means; (2) The guidance control device for a mobile robot according to claim 1, wherein the navigator means causes the route selection means to create an alternative route when traveling by the first to third travel control means is impossible. .